The Over 45S Adult Sexuality and Intimacy Scale: a New Approach to Understanding and Measuring Sexuality in Mid and Later Life Ashley Macleod Bsc, Bpsysc

The Over 45s Adult Sexuality and Intimacy Scale: A new approach to understanding and measuring sexuality in mid and later life Ashley Macleod BSc, BPsySc. (Hons).

Health and Ageing Research Group Faculty of Health, Arts & Design Swinburne University of Technology

This thesis is submitted in total fulfilment of the requirements for the degree of Doctor of Philosophy 8 April 2020

Primary supervisor: Professor Marita McCabe Associate supervisor: Dr Simone Buzwell

Abstract Sexuality and ageing research predominantly focuses on sexual problems such as erectile dysfunction, vaginal dryness, and the impact of physical health problems on the sexuality of older adults. However, limited evidence suggests that these topics reflect the sexual priorities of older adults. This thesis aims to develop a novel approach to understanding and measuring sexuality for heterosexual adults in mid and later life that reflects the experience of sexuality and associated factors as described by this cohort. The studies outlined in this PhD thesis are presented in five manuscripts. Two published systematic reviews highlight gaps in the literature, justifying the development of a new approach to conceptualising the experience of sexuality among heterosexual adults in mid and later life and the need for a validated measurement scale that aligns with this new approach. The third published manuscript describes the development of this new approach, and the fourth manuscript (in preparation) describes the development and refinement of a new measurement scale - the Over 45s Adult Sexuality and Intimacy Scale (OASIS). Three of the nine OASIS subscales reflect areas related to the experience of sexuality (aspects of intimacy and pleasure, sexual expression and sexual urges), and six subscales reflect factors related to the experience of sexuality (partner compatibility, factors associated with intimacy and pleasure, determinants of sexual desire, determinants of sexual expression, barriers to intimacy and barriers to sexual expression). The final manuscript (in preparation) examines similarities and differences between cohorts based on sex (male compared to female) and age cohort membership (midlife compared to later life) using the OASIS. A few differences between cohorts were identified: sexual expressions and sexual urges were more relevant to the experience of sexuality for men than women; partner compatibility was more relevant to the experience of sexuality for women than men; determinants of sexual desire and sexual urges were more relevant to midlife individuals than those in later life. These findings highlight the similarity of sexual experiences and preferences for men and women as they age, and the necessity of using an approach to understanding sexuality in mid and later life that reflects the experience of sexuality for these cohorts. Implications of this work are explored, and an argument is made for using an approach to understanding sexuality in mid and later life that reflects the sexual priorities of older adults instead of approaches that emphasise sexual functioning and sexual activity frequency.

Acknowledgements

I would first like to acknowledge my supervisors, Professor Marita McCabe and Dr Simone Buzwell. Their support and guidance has made me a better scientist, and has helped immeasurably throughout my candidature. I will be eternally grateful for the friendship and encouragement they have given me during this time, and I do not regret for one minute the added challenges posed by following my primary supervisor Marita across different Universities. They were worth it to continue to work with such an amazing woman. I would also like to acknowledge the ongoing support of a former supervisor, Dr Lucy Busija, who continued to provide her invaluable statistical expertise beyond what was required of her. Beyond the academic world, this thesis would not have been possible without the unwavering support and encouragement given to me by my family and friends. I would most especially like to thank my husband. It’s a difficult job to be the partner of a PhD candidate, and he has borne the brunt of having to cope with my stress, angst, anxiety, sleepless nights, distraction, and all the other emotional burdens that come with doing a PhD. He has been by my side, believing in me and supporting me, especially in the moments when I couldn’t believe in myself. His patience and belief in my abilities have often been the push I needed to keep moving forward, and I could not ask for a better partner in life. And finally, I would like to acknowledge the unwitting support and unconditional love my little boy has given me in his time on earth thus far. Since his birth halfway through my PhD journey, he has kept me grounded in times of stress, given me endless cuddles and affection when I needed them (and sleepless nights when I didn’t), and has helped prevent me from completely disappearing into this thesis. His mere existence has given me endless opportunities to practise mindfulness, especially in the final stages of writing my thesis, and this has been invaluable.

iii

Declaration by candidate

This thesis by publication is an original body of work undertaken solely during the period of candidature. It does not contain any material that has previously been admitted (or accepted) for the award of any other degree or diploma. To the best of my knowledge, this thesis does not contain material that has been written or published by any other person, except where due reference is made within the text for direct quotes. Five of the seven chapters included in this thesis have been drafted as manuscripts for high-ranking peer-reviewed journals. I warrant that I have obtained, where necessary, permission from the copyright owners to use any third party copyright material reproduced in the thesis (such as artwork, images, unpublished documents), or to use any of my own published work (such as journal articles) in which the copyright is held by another party (such as publisher, co-author). In line with this, the relative contributions of each listed author are noted in the five attached Authorship Indication Forms. Because of the linear nature of the work undertaken in this thesis, each manuscript has, by necessity, required the acceptance of the preceding article before it could be submitted for publication with the appropriate journal. At the time of submission, three of the five manuscripts have been published. The fourth and fifth manuscripts are currently in preparation for submission to appropriate high ranking journals. The fifth manuscript cannot be submitted until the fourth manuscript has been accepted for publication, as it directly references the preceding manuscript.

Authorship declaration forms

vii

viii

xii

xiii

xiv

Table of Contents

Authorship declaration forms ...... v List of Tables ...... xx List of Figures ...... xxi List of Publications ...... xxii List of Presentations ...... xxiii List of Awards ...... xxiv List of Project Media ...... xxv Chapter 1. Introduction and rationale ...... 1 1.1 General Introduction ...... 2 1.2 A history of sexuality research ...... 3 1.2.1 Early sexuality and ageing research ...... 4 1.2.2 Research in the late 20th century ...... 4 1.2.3 Recent developments ...... 7 1.2.4 A history of sexuality and ageing research – Summary ...... 8 1.3 Methodological inconsistencies that impede sexuality and ageing research ...... 9 1.3.1 Inconsistent use of terminology and definitions ...... 9 1.3.2 Limitations of current later life sexuality measurement approaches ...... 14 1.3.3 Methodological issues – Summary ...... 17 1.4 Theoretical framework of the thesis: Lifespan developmental theory ...... 17 1.4.1 Models of development in adulthood ...... 18 1.4.2 Sexuality throughout the life course ...... 19 1.4.3 Sexuality and ageing ...... 20 1.4.4 Sexuality and ageing within the normative-crisis model ...... 21 1.5 Aim of the thesis ...... 21 1.6 Thesis overview ...... 22 Chapter 2. Defining sexuality in later life: A systematic review ...... 25 2.1 Chapter guide ...... 26 2.2 Abstract ...... 27 2.3 Introduction ...... 28

2.4 Methods ...... 29 2.5 Results ...... 32 2.5.1 Implicit definitions of sexuality ...... 33 2.5.2 Explicit definitions of sexuality ...... 34 2.6 Discussion ...... 38 2.6.1 Limitations ...... 40 2.7 Conclusions and future directions ...... 41 Chapter 3. How well do measurement scales reflect the actual experience of sexuality in midlife and beyond? ...... 42 3.1 Chapter guide ...... 43 3.2 Abstract ...... 44 3.3 Introduction ...... 45 3.4 Method ...... 46 3.5 Results ...... 49 3.5.1 Aspects of sexuality identified within the qualitative articles ...... 51 3.5.2 Aspects of sexuality measured using validated measurement scales ...... 60 3.6 Discussion ...... 63 3.6.1 Limitations ...... 65 3.7 Conclusion ...... 65 Chapter 4. Study 1: Mapping the perceived sexuality of heterosexual men and women in mid and later life: A mixed-methods study ...... 67 4.1 Chapter guide ...... 68 4.2 Abstract ...... 69 4.3 Introduction ...... 71 4.4 Method ...... 73 4.4.1 Group concept mapping ...... 74 4.4.2 Thematic analysis of perceived change to sexual experience with increasing age ...... 83 4.5 Results ...... 84 4.5.1 Group concept mapping ...... 84 4.5.2 Thematic analysis of perceived change to sexual experience with increasing age ...... 90

xvi

4.6 Discussion ...... 93 4.6.1 Limitations ...... 97 4.7 Conclusions and future research ...... 98 Chapter 5. Study 2: Development and psychometric evaluation of the Over 45s Adult Sexuality and Intimacy Scale (OASIS) ...... 100 5.1 Chapter guide ...... 101 5.2 Abstract ...... 102 5.3 Introduction ...... 103 5.4 Study one: Pre-testing of developed items ...... 105 5.4.1 Method ...... 106 5.4.2 Results ...... 108 5.4.3 Discussion ...... 108 5.5 Study two: Quantitative assessment of items and construction of subscales ...... 109 5.5.1 Method ...... 110 5.5.2 Results ...... 114 5.5.3 Discussion ...... 123 5.6 Study 3: Scale evaluation ...... 124 5.6.1 Method ...... 126 5.6.2 Results ...... 131 5.6.3 Discussion ...... 134 5.7 General discussion ...... 135 5.7.1 Limitations ...... 137 5.8 Conclusions ...... 138 Chapter 6. Study 3: Similarities and differences in the experience of sexuality for men and women in mid and later life ...... 139 6.1 Chapter guide ...... 140 6.2 Abstract ...... 141 6.3 Introduction ...... 142 6.4 Method ...... 145 6.4.1 Participants ...... 145 6.4.2 Materials ...... 145 6.4.3 Procedure ...... 147

xvii

6.4.4 Data analysis ...... 147 6.5 Results ...... 149 6.5.1 Subscales measuring aspects related to the experience of sexuality ...... 149 6.5.2 Subscales measuring factors associated with the experience of sexuality ...... 149 6.6 Discussion ...... 154 6.6.1 Understanding the experience of sexuality and ageing ...... 154 6.6.2 Factors associated with sexuality and ageing ...... 156 6.6.3 Limitations ...... 158 6.7 Conclusions and future directions ...... 158 Chapter 7. General discussion ...... 159 7.1 General discussion ...... 160 7.2 Chapter guide ...... 161 7.3 A novel approach to understanding sexuality for heterosexual men and women in mid and later life ...... 162 7.4 Development of the OASIS ...... 164 7.5 Areas of similarity and difference in the experience of sexuality and associated factors across the later life stages ...... 165 7.5.1 Areas of similarity and difference between men and women ..... 166 7.5.2 Areas of similarity and difference between individuals in mid and later life ...... 168 7.5.3 Summary of the experience of sexuality for community-dwelling heterosexual adults in mid and later life ...... 170 7.6 Theoretical alignment of results to the normative-crisis model .. 170 7.7 Implications ...... 171 7.8 Limitations ...... 173 7.9 Future research ...... 175 7.10 Conclusions ...... 176 References ...... 178 Appendix A: Systematic review search terms ...... 195 Appendix B: Study 1 ethics approvals and related documentation ...... 196 Appendix C: Study 2 ethics approvals and related documentation ...... 226

xviii

Appendix D: Chapter 5 supplementary materials ...... 261 Appendix E: Study 3 ethics approvals and related documentation ...... 281

xix

List of Tables Table 2.1 Themes discussed in articles that use an implicit definition of sexuality……………………………………………………………….....35 Table 2.2 Themes captured in explicit definitions of sexuality (by article) …..……37 Table 3.1 Inclusion and exclusion criteria……………………………..…………….….50 Table 3.2 Quality assessment of included qualitative articles………………..…….…52 Table 3.3 Quality assessment of included quantitative articles……………..……..….54 Table 3.4 Demographic details of qualitative articles and the aspects they discussed ………………………………………………………………...56 Table 3.5 Demographic details of quantitative articles, the validated measurement scales used, and the aspects they measure ………………….…………..61 Table 4.1 Demographic properties of study sample across phase one and two…….76 Table 4.2 Distribution of clustered statements and mean importance scores across identified clusters…………………………………………………….....85 Table 4.3 Sum totals of described changes to the experience of sexuality with age across gender and age cohorts…………………………………………….…..91 Table 5.1 Study 2 participant demographics ...... 110 Table 5.2 Dimensionality assessment of original and final versions of the OASIS 115 Table 5.3 DIF analysis of original and final versions of the OASIS subscale ...... 120 Table 5.4 Predicted correlations between OASIS subscales with matched and unmatched existing sexuality measurement subscales ...... 127 Table 5.5 Study 3 Participant demographics ...... 128 Table 5.6 OASIS Subscale reliability assessments using Cronbach’s α and ICC .. 132 Table 5.7 Concurrent validity assessments of OASIS subscales and matched later life sexuality measurement scales using Pearson’s r ...... 133 Table 6.1 Demographic profile of participants ...... 1500 Table 6.2 Significant univariate effects for sex and age cohort membership across OASIS subscale ...... 151 Table 6.3 Summary of predicted and obtained main effects identified for each OASIS subscale, based on age cohort membership, sex, and age cohort membership and sex combined ...... 1522 Table 6.4 Mean and standard deviation of scores for age cohort membership and sex groups for each OASIS subscale ...... 153

List of Figures Figure 2. 1 PRISMA flowchart for original and updated systematic review ...... 31 Figure 3. 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart of searches for articles examining aspects of sexuality at midlife and later life ...... 48 Figure 4. 1 Depiction of cluster proximity using a bivariate point map...... 88 Figure 5. 1 Person-item distribution of the final version of the OASIS subscales. . 119

xxi

List of Publications i. Macleod, A., & McCabe, M. P. (2019). How well do measurement scales reflect the actual experience of sexuality in mid-life and beyond? Sexual Medicine Reviews, 7(1), 29-45. doi: 10.1016/j.sxmr.2018.04.004 Received date: 9 January 2018. Accepted date: 19 April 2018. Available online: 26 July 2018.

ii. Macleod, A., & McCabe, M. P. (Accepted). Defining sexuality in later life: A systematic review. Australasian Journal on Ageing. Received date: 20 April 2019. Received in revised form: 26 August 2019. Accepted date: 18 September 2019. iii. Macleod, A., Busija, L., & McCabe, M. P. (2020). Mapping the perceived sexuality of heterosexual men and women in mid and later life: A mixed-methods study. Sexual Medicine, 8(1), 84-99. doi: http://dx.doi.org/10.1016/j.esxm.2019.10.001 Received date: 22 July 2019. Received in revised form: 1 October 2019. Accepted date: 3 October 2019, Available online: 6 November 2019.

xxii

List of Presentations Oral presentations (National)  Australian Association of Gerontology, Early Researchers in Ageing Conference – "Defining sexuality in an ageing population" (Dec 2015, Melbourne VIC)  Institute of Health and Ageing Symposium, Australian Catholic University - "From intercourse to intimacy - Reshaping what we know about sexuality in mid and later life" (Nov 2016, Melbourne VIC)  51st Australian Association of Gerontology Conference - "Defining the sexual experience of older adults: A conceptual model" (Nov 2018, Melbourne VIC)  52nd Australian Association of Gerontology Conference - "The OASIS: A new measurement scale to examine later life sexuality" (Nov 2019, Sydney NSW)

Oral presentations (International)  7th Multidisciplinary Sexual Dysfunction Conference – "Redefining later life sexuality" (May 2016, Auckland, NZ)

Poster presentations (International)  International Association of Geriatric Psychiatry 40th Annual Meeting – "Understanding sexuality in later life: Presenting a new conceptual model to define the sexual experience of older adults." (Mar 2018, Honolulu, USA)

xxiii

List of Awards  3 Minute Thesis, Australian Catholic University Finalist (2016)  3 Minute Thesis, Australian Catholic University Winner - Melbourne Division (2017)

xxiv

List of Project Media Radio interviews  ABC Radio Canberra – 18 Feb 2016 (LINK)  ABC Radio Ballarat – 25 Feb 2016  ABC Radio Perth – 15 June 2016  ABC Radio NSW – 16 June 2016  ABC Radio Canberra – 16 June 2016  ABC Radio Adelaide – 21 June 2016  ABC Radio National – 22 June 2016 (LINK)  ABC Radio NSW – 4 September 2018

Print articles  Starts at 60 Website (2016). What getting intimate at 60 really means (LINK)  The Ageing Agenda (2016). Do not disturb: Partnered baby boomers want couple rooms in aged care facilities (LINK)  Huffington Post (2016). One day you will be old and you may still want to have sex (LINK)  The Age (2016). Hugs not drugs an aphrodisiac for older Australians (LINK)  The Conversation (2017). Older people still have sex, but it's the intimacy and affection that matters more (LINK)  Starts at 60 Website (2018). How saucy Baby Boomers are busting the 'asexual oldie' stereotype (LINK)  Go 55's Website (2018). Are the baby boomers having a second sexual revolution? (LINK)  The Senior (2018). Swinburne University of Technology seeks seniors for OASIS sex study (LINK)

xxv

This thesis is dedicated to anyone whose sexuality has ever been ignored or invalidated.

xxvi

"He'd noticed that sex bore some resemblance to cookery: it fascinated people, they sometimes bought books full of complicated recipes and interesting pictures, and sometimes when they were really hungry they created vast banquets in their imagination – but at the end of the day they'd settle quite happily for egg and chips. If it was well done and maybe had a slice of tomato."

Terry Pratchett, The Fifth Elephant

xxvii

Chapter 1. Introduction and rationale

1.1 General Introduction More than half a century ago, Abraham Maslow identified sex as one of our basic physiological needs (Maslow, 1943). Since then, research has further established the role of sex and sexuality in the quality of life (QoL) and wellbeing of men and women throughout the lifespan (e.g., Baldissera, Bueno, & Hoga, 2012; Henry & McNab, 2003). Yet the myth of the 'asexual oldie' is firmly entrenched in today's society (Bouman & Kleinplatz, 2014; Gott, 2006); sex is seen as something for the young, and attractiveness and desirability are often equated with a youthful appearance. The misguided (and ageist) belief that older adults are no longer sexual beings has helped shape the perspectives of both society and the health professions, and has had a direct impact on the availability, type, and quality of sexual health care available to older adults (Aboderin, 2014). As the proportion of older adults in the population continues to rise (United Nations, Department of Economic and Social Affairs, Population Division, 2019), the emphasis on the concept of ‘healthy ageing’ has led to the pathologising of changes associated with the ageing process (Bradway & Beard, 2015; Gott, 2006). Currently, the majority of sexuality and ageing research relies on an approach that assumes that changes in physiological functioning associated with age are inherently negative (Gott, 2004; Phillipson, 2013). However, the assumption that the experience of sexuality declines as one gets older ignores evidence that older adults are able to enjoy their sexuality in both the presence or absence of any sexual dysfunction (Bouman & Kleinplatz, 2014; DeLamater, 2012; Hinchliff & Gott, 2004; Hurd Clarke, 2006; Ménard et al., 2014). Using an approach to ageing that focuses on medical interventions to improve age-related health problems is logical where age-associated changes may reduce an individual's life-expectancy or their psychological or physical functioning. However, the extension of this medicalised approach to other aspects of the ageing process bears further scrutiny. In particular, the pathologising of naturally-occurring changes in sexual functioning or sexual responses that occur with age reduces an individual's sexuality to its physiological components. This pathologisation, in turn, promotes the idea that an individual's ability (or inability) to engage in penile-vaginal intercourse is a complete indication of that individual's sexuality. Such an approach not only fails to recognise sexual activity beyond that of penile-vaginal intercourse, it also captures

only a fraction of the overall experience of sexuality (e.g., DeLamater, 2012; Herbenick et al., 2010). The studies in this thesis are designed as a counterpoint to the over- medicalisation of sexuality across mid and later life that is present within research, and throughout society more generally. The current thesis focuses only on community-dwelling heterosexual adults in mid and later life. By focusing exclusively on this cohort, several topics were not considered within the scope of this thesis. First, including only adults in mid and later life means that potential differences between older and younger cohorts were not investigated. Second, the focus on the sexuality of community-dwelling older adults means that the experience of sexuality for aged care residents is not explored (see Rahn, Bennett, Lykins, & Jones, 2018 for more information about this topic). Third, examining the experience of sexuality within the context of clinical diagnoses was not within the scope of this research, although no exclusion criteria were set that excluded individuals with diagnosed health conditions from participating in the studies. Lastly, LGBTQIA+ individuals were not included in the current research, despite the likelihood that the results produced in this thesis may, to some extent, be relevant to older LGBTQIA+ people. The current chapter provides an overview of the history of sexuality and ageing research (Section 1.2), before outlining methodological inconsistencies that impede the investigation of sexuality across the later life stages (Section 1.3). These sections are followed by an introduction to the theoretical approach used to frame the thesis (Section 1.4) and the aim of the thesis (Section 1.5). The chapter concludes with an outline of the remaining thesis chapters (Section 1.6). 1.2 A history of sexuality research Sexuality has been a focus of research for over a hundred years and has been highlighted as an integral part of the human experience (Deacon, Minichiello, & Plummer, 1995; Kaas, Rousseau, & Haworth Continuing Features Submission, 2008; McIntosh, 1981; Shearer & Shearer, 1977; Stimson, Wase, & Stimson, 1981). Since the original Kinsey studies on human sexuality in the late 1940s (Kinsey, Pomeroy, & Martin, 1948), sexuality research has consistently demonstrated that sexual interest and sexual activity is sustained well into old age (Catania & White, 1982; Kaplan, 1990; Shearer & Shearer, 1977; Weizman & Hart, 1987). However, only a fraction of participants in these ground-breaking studies were in mid or later life (Kinsey et al., 3

1948; Kinsey, Pomeroy, & Martin, 1953). In spite of the prevailing myth that older adults are no longer interested in sex or sexual expression (Papaharitou et al., 2008), the idea that sexuality is experienced by older people is not new, and research with a focus on sexuality and ageing has been published from as early as 1960 (Christenson & Gagnon, 1965; Newman & Nichols, 1960; Pfeiffer, Verwoerdt, & Wang, 1969). 1.2.1 Early sexuality and ageing research Before the 1960s, limited research focused on the sexuality of older adults. However, in a landmark study published in 1960, Newman and Nichols became some of the first researchers to examine an exclusively older population, sampling 250 community-dwelling adults aged between 60 and 93 years (Newman & Nichols, 1960). The results of this study identified a decline in sexual activity frequencies and sexual urges with increasing age. In this study, personal or spousal physical health was also a key factor in whether individuals engaged in sexual activities, with most healthy couples engaging in sexual activity well into later life. These results were later supported by other early sexuality studies (Christenson & Gagnon, 1965; Pfeiffer et al., 1969). Subsequent to the pioneering work of Newman and Nichols (1960), most early sexuality and ageing research referred to sexuality almost exclusively within the context of sexual behaviours and sexual urges (e.g., Bretschneider & McCoy, 1988; Catania & White, 1982; Holzapfel, 1994; Pfeiffer et al., 1969; Shearer & Shearer, 1977). Primary measures of sexuality during this period included the type and frequency of sexual activities, the orgasmic potential of older adults, and the perceived strength of older adults’ sexual urges. Participants were either volunteers or a subpopulation of a much larger study and were predominantly Caucasian. In line with the work conducted by Masters and Johnson (1966) with younger populations, research examining the sexual physiology and sexual functioning of older people highlighted the tendency for sexuality research to use a mostly biologically-driven, male-dominated perspective that emphasised penile-vaginal intercourse over other forms of sexual expression (Shearer & Shearer, 1977). 1.2.2 Research in the late 20th century In the 1980s, Masters and Johnson (1981) highlighted that the term ‘sexuality’ had yet to be clearly defined in the literature, and subsequently defined ‘sex’ and ‘sexuality’ as “terms employed to identify the sexual practices and the sexual identity of the individual man or woman” (Masters & Johnson, 1981, p. 385). However, these 4

definitions have been modified and redefined numerous times since they were first introduced by Masters and Johnson (1981). Beyond the definition published by Masters and Johnson (1981), many of the numerous definitions that have been used in the latter half of the 20th Century referred to sexuality as ‘complex’ and frequently but inconsistently grouped aspects of sexuality such as sexual behaviours and attitudes related to sexuality with other external but related factors such as relationships, self-esteem and loyalty (Johnson, 1996; Loehr, Verma, & Seguin, 1997; Russell, 1998). Sexuality studies published during this period generally framed sexuality as the domain of the young and highlighted this as a significant barrier to the study of sexuality in older adults (e.g., Holzapfel, 1994; McIntosh, 1981; Stimson et al., 1981; Travis, 1987). Despite issues related to the absence of clear definitions across sexuality and ageing research, the 1980s and 1990s included a more holistic examination of the different biological and psychosocial aspects of sexuality and ageing than previous decades. Significant efforts were made to highlight how negative attitudes towards sexuality and ageing impact the older person, and researchers began examining the effects of attitudes towards sexuality and ageing on the experience and expression of sexuality for older adults (McIntosh, 1981; Portonova, Young, & Newman, 1984; Winn & Newton, 1982). For example, a study by McIntosh (1981) demonstrated that non-permissive attitudes towards sex among older people were, in part, shaped by negative societal attitudes towards sexuality and ageing and that older women were less permissive of sexual expression by their peers than they were for their younger counterparts. In contrast, a study by Winn and Newton (1982) revealed that non- permissive attitudes towards the sexuality of older adults are not universal; in traditional societies, it was accepted that older men and women would remain sexually active well into later life. Broadly, the research from the late 20th Century focused on measuring sexual interest, sexual activity, sexual knowledge, sexual satisfaction, attitudes relating to sex and sexuality, and sexual functioning (e.g., Dello Buono et al., 1998; Johnson, 1996; Matthias, Lubben, Atchison, & Schweitzer, 1997; Portonova, et al., 1984). Several related factors were also incorporated into sexuality measures, such as mental and physical health, relationships, religiosity, and social attitudes (e.g., Bergström‐Walan & Nielsen, 1990; Bortz, Wallace & Wiley, 1999; Dello Buono et al., 1998; Matthias, et al., 1997; Nilsson, 1987; Steinke, 1986). Research also examined reasons for 5

decreases in sexual activity, with these decreases often identified as related to the absence of a sexual partner or the poor health of a spouse (e.g., Deacon et al., 1995; Nilsson, 1987). For example, McCracken (1988) likened the decline in sexual expressions for older populations as being similar to why they stop riding bicycles: poor physical health, fear of ridicule, or simply the absence of a bike. Studies by Brettschneider and McCoy (1988) and Masters and Johnson (1981) also indicated that higher levels of sexual activity in early adulthood were related to prolonged sexual functioning in later life. Consistently, results have indicated that continued sexual expression is possible for older populations (Bretschneider & McCoy, 1988; Deacon et al., 1995; Herron & Herron, 1999). Researchers and clinicians have been urged by sexuality and ageing experts to be conscious of their own biases towards the sexuality of older adults and the importance of incorporating sexual health and sex education discussions into clinical care, with calls for both health professionals and the wider public to receive education relating to sexuality and ageing (Allen, 1987; Bates- Jensen, 1989; Gurian, 1986; Herron & Herron, 1999; McIntosh, 1981; Myers, 1985; Newman & Nichols, 1960; Renshaw, 1983; Shearer & Shearer, 1977; Steinke & Bergen, 1986; Weizman & Hart, 1987). However, societal attitudes and stereotypes of the asexual older person have continued to restrict the dissemination of these findings more broadly throughout society and the health professions (Bouman & Kleinplatz, 2014; Kaas et al., 2008; McIntosh, 1981; Steinke & Bergen, 1986). Towards the end of the 20th Century, several articles summarising the research on sexuality and ageing began appearing. These articles were written to educate health professionals about sexuality and ageing, potential barriers to sexual expression, and the importance of incorporating sexual history taking and discussions about sex into regular clinical practices (Barber, 1996; Russell, 1998). In line with much of the literature, there was a notable biomedical focus that emphasised the physiological aspects of ageing (Drench & Losee, 1996; McCracken, 1988; Segraves & Segraves, 1995). Articles examining sexual behaviours and sexual interests also began looking at the reasons behind sexual decline or the cessation of sexual activities (Bergström‐ Walan & Nielsen, 1990; Crose & Drake, 1993; Kellett, 1991; Marsiglio & Donnelly, 1991). As a result, a consistent message in the sexuality and ageing literature appeared: while sexual frequency may decline over time, the reasons for this decline

are varied, and many older adults remain sexually active well into later life (Catania & White, 1982; Kaplan, 1990; Shearer & Shearer, 1977; Weizman & Hart, 1987). 1.2.3 Recent developments Since the turn of the century, there has been an increase in the volume of sexuality research published that focuses on older populations. A large portion of this research has continued to expand on ideas outlined in the 1980s and 1990s, and has focused on sexual functioning as it relates to the ageing process, relationships, sexual health, attitudes related to sexuality, sexual behaviours, and the education of health professionals (e.g., DeLamater, Hyde, & Fong, 2008; Gott & Hinchliff, 2003a; Guan, 2004; Hyde et al., 2010; Killinger, Boura, & Diokno, 2014; Langer-Most & Langer, 2010; Watters & Boyd, 2009; Willert & Semans, 2000). A notable exception to this is research by Ménard et al. (2014) which is some of the first research to frame older adults as 'experts' in sexual expression. In the research by Ménard et al. (2014), men and women aged 60 years and above who were in long term relationships and who defined themselves as having experienced ‘great sex’ were interviewed about their sexual experiences. In this study, the results highlighted the role of relationship dynamics, partner availability, openness to new experiences, and overcoming learned ageism. In addition, it demonstrated the importance of specific elements of relationship dynamics, including communication effectiveness, mutual empathy, and depth of relationship (Ménard et al., 2014). As the sexuality and ageing literature continued to expand, so too did the breadth of topics being measured. The qualitative research centred around types of sexual behaviours and sexual expression (e.g., Hurd Clarke, 2006; Moore, 2010; Sandberg, 2013), the personal experience of sexual desire (Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; Lagana & Maciel, 2010; Wood, Mansfield, & Koch, 2007), and attitudes towards sexuality (Hinchliff & Gott, 2008). The quantitative research focused on measures of sexual activity type and frequency (Galinsky, McClintock, & Waite, 2014; Lindau et al., 2007; Palacios-Cena et al., 2012; Wang, Lu, Chen, & Yu, 2008), sexual functioning (Beckman, Waern, Gustafson, & Skoog, 2008; Lindau et al., 2007; Trudel et al., 2014), sexual interest and desire (Galinsky et al., 2014; Ginsberg, Pomerantz, & Kramer-Feeley, 2005; Trudel et al., 2014; Umidi, Pini, Ferretti, Vergani, & Annoni, 2007), and sexual and relationship satisfaction (DeLamater et al., 2008; Malakouti, Salehi, Nojomi, Zandi, & Eftekhar, 2012; Müller, Nienaber, Reis, Kropp, & Meyer, 2014). 7

Despite research expanding to include a diverse range of topics, most research continued to focus primarily on the sexual functioning and sexual activity frequency of older adults (e.g., Arias-Castillo, Ceballos-Osorio, Ochoa, & Reyes-Ortiz, 2009; Beckman et al., 2008; Bortz et al., 1999; DeLamater et al., 2008; Momtaz, Hamid, Ibrahim, & Akahbar, 2014; Palacios-Cena et al., 2012; Trudel et al., 2014). In addition, few efforts have been made to incorporate the topics identified in the expanded research scope into the definitions of sexuality used for older adults. Most often, the term ‘sexuality’ is left undefined in the sexuality and ageing literature, and where definitions are present, these are often borrowed from sexuality research conducted with younger cohorts and demonstrate little consistency across studies (Bauer, McAuliffe, & Nay, 2007; Guan, 2004; Kessel, 2001). In addition to problems with definitions, some health practitioners still fail to be accepting of and incorporate discussions of sexuality into the regular health care of older adults, and little progress has been made in this area over the past 50 years (Bauer, Haesler, & Fetherstonhaugh, 2015; Gewirtz-Meydan, Levkovich, Mock, Gur, & Ayalon, 2019; Haesler, Bauer, & Fetherstonhaugh, 2016). As a result, the myth of the asexual older person has been maintained, despite evidence that adults continue to desire and pursue sexual expression well into old age and that the experience of sexuality is an integral part of wellbeing at any age. (Deacon et al., 1995; Kaas et al., 2008; McIntosh, 1981; Shearer & Shearer, 1977; Stimson et al., 1981). 1.2.4 A history of sexuality and ageing research – Summary An examination of the history of sexuality and ageing research identifies a consistent message: individuals see themselves as sexual beings throughout the lifespan and continue to engage in various forms of sexual activity well into later life. While there is evidence that the frequency of sexual behaviours can decrease with age, this decline is often related to personal or partner health issues, or the absence of a partner, rather than a loss of interest. Negative attitudes towards the sexuality of older populations have also contributed to the myth of the asexual older person, despite research indicating that older adults still consider themselves sexual beings, and still engage in sexual activities (Gewirtz-Meydan et al., 2019). Of particular concern across the sexuality and ageing literature are the inconsistencies between qualitative evidence showing that older adults are primarily focused on the quality of their sexual experiences, and quantitative research that focused on the frequency of sexual intercourse and the maintenance or improvement 8

of sexual functioning in later life stages (e.g., Hinchliff & Gott, 2004; Hurd Clarke, 2006). These inconsistencies highlight problems with how sexuality is being conceptualised across the later stages of life, and underscore gaps in how sexuality across the later life stages is being measured. These issues then present challenges for cultivating a clear message for describing the experience of sexuality for older men and women in a way that is consistent with the experience of sexuality for these individuals. The following section provides an overview of the key methodological issues present within the sexuality and ageing research that impede the systematic development of a clear and consistent message about how sexuality in the later stages of life is understood. 1.3 Methodological inconsistencies that impede sexuality and ageing research To date, no single approach to understanding sexuality within the context of ageing exists that incorporates both qualitative and quantitative research to reflect the experience of sexuality for adults in mid and later life accurately. Likewise, the narrow scope of existing later life sexuality measurement scales and the limited number of such scales available to researchers makes it difficult for researchers to investigate the sexual experiences of older cohorts. Past research has identified inconsistencies in how sexuality and other key terms are defined, the limited number of validated later life sexuality measurement scales available, and variability in age cut-off points used to identify ‘older’ adults (Bell, Reissing, Henry, & VanZuylen, 2017). These measurement issues have meant that researchers and clinicians must carefully examine the context within which results have been obtained to utilise the research effectively. The following sub-sections describe two significant hurdles to developing a comprehensive understanding of the experience of sexuality for adults in mid and later life. The inconsistent use of terminology and definitions related to sexuality in mid and later life will be discussed first, followed by a discussion of the limitations of current measurement approaches. 1.3.1 Inconsistent use of terminology and definitions Throughout the sexuality and ageing literature, it is common for researchers to rely on implicit definitions of sexuality and other related terms. However, when examining the definitions provided for these terms (when they are present at all), no universal definitions are apparent. Both researchers and participants have used a range 9

of interpretations for what sexuality and other related terms mean, raising questions about the validity of questionnaire items and data obtained in the absence of clear definitions of sexuality terms. The purpose of the following sub-sections is to provide an overview of the use of explicit and implicit definitions for sexuality in later life within the existing literature and discuss the lack of consistency in how the terms 'sexuality', 'sexual activity', and 'later life' are being defined. 1.3.1.1 Explicit definitions of sexuality. Several definitions of sexuality are used in sexuality and ageing research, and the use of multiple definitions within a single paper to describe sexuality is not uncommon (e.g., Baldissera et al., 2012; Benbow & Beeston, 2012; Johnson, 1996; Kessel, 2001). The range of existing definitions is extensive and can include broad and detailed descriptions (e.g., World Health Organisation, 2006) or narrow descriptions (e.g., Guan, 2004). Definitions frequently used to describe sexuality in later life assume an unchanging experience of sexuality throughout life, despite evidence that the experience of sexuality changes over time (e.g., Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; Sandberg, 2013). Many of these definitions describe a broad range of topics and reference numerous biological, psychological, social, and environmental factors associated with the experience of sexuality. However, in many cases, it is unclear how these definitions have been developed or what populations have been used to construct them. Despite the broad range of existing definitions of sexuality, few directly reference sexuality in the later years (e.g., Barber, 1996). Some authors have combined multiple definitions of sexuality to describe sexuality in later life more accurately. For example, Bauer et al. (2007) used three different definitions of sexuality to define sexuality in general and in later life. The apparent need for researchers to combine multiple definitions to adequately reflect their conceptualisation of sexuality indicates that existing definitions do not adequately describe the experience of sexuality for older adults. This inadequacy is worrisome, especially when considering that a majority of research studies rely on implicit definitions of sexuality. 1.3.1.2 Implicit definitions of sexuality. Many theoretical or summary articles discuss later life sexuality without providing a clear definition of sexuality. Instead, they rely on a reader's existing 10

understanding of sexuality to frame the discussion. For most later life sexuality studies that rely on implicit definitions of sexuality, it is necessary to examine the types of sexuality measures used to understand how sexuality is being conceptualised within the study. The problem with using an approach that relies on assumed knowledge is highlighted when the topics discussed in different studies are compared. For example, Lusti-Narasimhan and Beard (2013) examined sexual health, sexuality, sexual desire, sexual activity, sexual functioning, physical intimacy, sexual drive, sexual experiences, sexual behaviours, sexual practices, sexual needs, and sexual orientation. In contrast, Bancroft (2007) examined sexual activity, intimate relationships, sexual arousal, sexual desire, sexual interest, menopause, fertility, orgasm, sexual response, sexual function (including erectile function), relationship factors, sexual wellbeing, intimacy, and sexual pleasure. While many of the topics examined overlap, there are notable differences, including the discussion of fertility, orgasm, sexual pleasure, sexual orientation, and menopause. The issues caused by this lack of consistency in conceptualisations of sexuality are further compounded by a lack of clarity in how researchers define other sexuality terms, such as sexual activity or sexual behaviours. 1.3.1.3 Inconsistent definitions of sexual activity. A significant area of focus in later life sexuality research has been sexual activity. However, many studies frequently rely on implicit definitions of sexual activity to frame their research (e.g., Bortz & Wallace, 1999; DeLamater et al., 2008). When a definition of sexual activity is provided, these can vary from study to study. For example, in a paper by Beckman, Waern, Ostling, Sundh and Skoog (2014), sexual activity is defined only as sexual intercourse. This definition provides a highly heteronormative description of sexuality. In contrast, Hyde et al. (2010) defined sexual activity as "any mutually voluntary activity with another person that involves sexual contact, regardless of whether intercourse or orgasm occurs" (p 694). The more inclusive nature of this definition highlights the vast differences in how sexual activity is conceptualised in later life sexuality research. Variations in how older adults interpret and respond to questions about sexual behaviours further complicate the process of untangling inconsistencies across explicit definitions of sexual activity, and the use of subjective interpretations for implicit definitions of sexual activity. For example, older adults often interpret ‘sexual activity’ to mean penile-vaginal intercourse (Bevilacqua, Leite, Hildebrandt, & do 11

Carmo Jahn, 2013; Hurd Clarke, 2006; Yan, Wu, Ho, & Pearson, 2011). However, many older adults have also reported that they have broadened their definitions of sexual expression to include affectionate behaviours such as kissing, petting, and cuddling (Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; Gott & Hinchliff, 2003b; Hinchliff & Gott, 2004; Hurd Clarke, 2006; Ménard et al., 2014; Sandberg, 2013). These adjustments primarily occur as a result of changes in relationship dynamics (e.g., Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015), or through older adults redefining what constitutes ‘sexual’ behaviours as a result of decreased sexual function (e.g., Gott, 2006; Ménard et al., 2014). Given the variability in how many older adults define ‘sexual activity’, it is worrisome that studies that examine sexual activity do not always specify whether sexual activity was defined for study participants, or whether it was up to participants to define sexual activity for themselves. For example, studies that ask participants "are you sexually active?" could be interpreted as referring to the incidence or frequency of penile-vaginal intercourse or a much more diverse range of activities including masturbation or kissing. Without clearly defining what is meant by the term ‘sexual activity’, researchers cannot accurately determine whether a negative response to questions about sexual activity indicates the absence of all activities that could be described as sexual or only the absence of penile-vaginal intercourse. In this scenario, a subjective judgement call would be needed about what behaviours are defined as ‘sexual activity’. The necessity of this judgement call, in turn, raises other issues, including how to consolidate detailed information about a range of defined sexual activities with data that uses an implicit definition of sexuality. 1.3.1.4 Definition of sexuality used in the current thesis. There is a need for the more consistent use of terminology regarding sexuality across studies. Such descriptions of sexuality help to provide context on the scope of the research topic and help orient the reader to the perspectives used by the researcher. The current thesis uses the definition of sexuality published by the World Health Organisation (WHO) to define the scope of included topics in subsequent investigations of sexuality and ageing (World Health Organisation, 2006). The WHO defines sexuality as: "...a central aspect of being human throughout life encompasses sex, gender identities and roles, sexual orientation, eroticism, pleasure, intimacy and reproduction. Sexuality is experienced and expressed in thoughts, fantasies, 12

desires, beliefs, attitudes, values, behaviours, practices, roles and relationships. While sexuality can include all of these dimensions, not all of them are always experienced or expressed..." (World Health Organisation, 2006). Definitions provided by the WHO (including the definition of sexuality), are frequently used across different disciplines and incorporate scientific knowledge developed across many diverse cultures, ages, and geographical locations. This definition of sexuality was selected as a broadly relevant and readily available definition that reflects the experience of sexuality for individuals throughout the lifespan. 1.3.1.5 Inconsistent definitions for ‘midlife’ and ‘later life’. At present, no consistent age cut-off points are being used to define what ages are considered ‘older’ across sexuality and ageing research. As a result, research has included participants from as young as 50 years (e.g., Bortz & Wallace, 1999; Helgason et al., 1996; Knodel & Chayovan, 2001) to those over 100 years (e.g., Dello Buono et al., 1998; Padoani et al., 2000). Lifespan definitions consider the age of 50 as consistently midlife, with later life beginning at age 60 or 65 years (Lachman, 2004). Because of the large amount of diversity present in older populations and longer life spans, it is becoming increasingly common to encounter distinctions within older cohort groups such as ‘young-old’ and ‘oldest-old’ (Phillipson, 2013). The variability in age ranges that are used in sexuality and ageing research to describe 'older' adults means that research often combines groups of individuals that span several decades of life. In doing so, groups that include several distinct generations of life are treated as a single cohort. For example, in the study by Dello Buono et al. (1998) participant ages ranged from 65 years to 106 years, a span of 41 years. As a result, this sample includes several distinct cohorts that could be identified using age-group defined generations such as the ‘first-wave baby boom’ generation (i.e., those born between 1945 and 1954; Phillipson, 2013), or historical generations such as the ‘war babies’ (i.e., those born during World War II), or those born prior to World War II. Likewise, this sample may also include several generations of a single family such as grandparent, parent, and child (Phillipson, 2013). Under a lifespan developmental approach that splits adulthood into three stages (early, middle and later adulthood; Lehmiller, 2017), the variation in how later life is being defined raises questions about whether sexuality and ageing research 13

reflect the experiences of individuals in midlife, later life, or individuals across both mid and later life. This lack of consistency has implications for the generalisability of results. Given the likely lifestyle differences of, for example, a 50-year old working full time with children at home and that of a 75-year-old retiree, the extrapolation of data collected using a younger cohort is unlikely to be generalisable to older populations. The limited generalisability of these results will, in turn, have implications for how data are systematically consolidated, and defining adults who are more accurately described as being in midlife (i.e., those aged 45-64 years; United Nations, Statistical Office, 1982) as ‘older’ adults increases the heterogeneity of the sample. 1.3.1.5.1 Defining midlife and later life in the current thesis. The presence of significant variability throughout the literature in terms of what ages are captured within the umbrella term of ‘ageing’ meant that a decision was made to use the age cut-off recommendations published by the United Nations (UN) to define midlife and later life (United Nations, Statistical Office, 1982). These guidelines were selected because they are readily available and frequently used across many different disciplines and incorporate scientific knowledge developed across many diverse cultures, ages, and geographical locations (United Nations, Statistical Office, 1982). The UN guidelines define midlife as occurring between 45 and 64 years of age and identify later life as age 65 years and over (United Nations, Statistical Office, 1982). By including both individuals in midlife and later life in the thesis scope, the variability in terms of what age constitutes an ‘older’ adult in the existing literature is accounted for, and further clarification of differences as they relate to mid and later life stages can be made. This expansion of age ranges to include those within midlife will help to frame discussions of sexuality more effectively across the later life stages, and allow for a more nuanced approach to understanding sexuality throughout adulthood. However, developing a more nuanced understanding of sexuality across mid and later life is currently restricted by several issues related to how sexuality is being measured across the later life stages. These issues are outlined in the next subsection. 1.3.2 Limitations of current later life sexuality measurement approaches Many approaches to measuring sexuality for older adults have been used across studies. However, the diversity in approaches and interpretations used across 14

studies have generated a useful but mixed collection of results that are often difficult to extrapolate to broader populations, and that may have questionable validity. This has meant that researchers are often required to choose between existing measures of sexuality that do not adequately assess the area of inquiry or undertaking the time- consuming and complex task of developing their own measures. These limitations highlight the need for well-designed, psychometrically sound later life sexuality measures. Few scales have been developed to measure sexuality for older adults. Existing later life sexuality scales focus predominantly on assessing sexual activities, attitudes relating to sexual permissiveness, sexual satisfaction, and sexual interest (e.g., the Geriatric Sexuality Inventory [GSI], Kazer, Grossman, Kerins, Kris, & Tocchi, 2013; and the Senior Adult Sexuality Scale [SASS], E. Weinstein, 2013). However, sexuality and ageing researchers regularly develop their own survey item(s), select specific items from pre-existing scales designed using younger cohorts, or combine pre-existing sexuality scales designed using younger populations with independently developed items instead of using the available later life sexuality measurement scales (e.g., Matthias et al., 1997; Minichiello, Plummer, & Loxton, 2004; Momtaz et al., 2014; Penhollow, Young, & Denny, 2009). By using unvalidated survey items or by combining unvalidated items with existing scales, researchers are able to bypass the limited amount of tools available to them, but face additional challenges in terms of how well the results can be generalised to broader populations, or how reliable the results are compared with results obtained with validated measurement scales. The following sub-sections outline specific challenges created by the different approaches to measuring sexuality across mid and later life used in the existing literature. 1.3.2.1 Unvalidated and single-item scales. Several studies have included items that have been specifically developed to address the research question in the study (e.g., Beckman et al., 2008; Papaharitou et al., 2008). However, there is often a lack of information provided about how these new items were assessed for their validity and reliability, or whether the psychometric properties of the items or scales were assessed at all. For example, a study by Bach, Mortimer, VandeWeerd, and Corvin (2013) used the single item "are you sexually active?" to measure the sexual activity of older adults, but provided no data about the 15

item’s validity or reliability. Similarly, although Beckman et al. (2008) provided detail about the inter-rater reliability of their semi-structured interview items, no information was given about the validity of these items. The use of unvalidated or single-item measures in research is problematic for several reasons. The absence of validity assessments for survey items poses risks in terms of whether these items actually measure the construct they were designed to measure, whether items measure the same or different constructs, or whether the measure can be used to discriminate between two respondents. Likewise, the absence of reliability assessments for these measures means that there is no information available about the consistency of responses by a single individual to an item (or scale) at different time points, whether different assessors can produce the same scores for a single respondent using the same measures, or whether there is internal consistency across items. Unfortunately, given the narrow scope of most later life sexuality measurement scales, many researchers are required to, at the very least, supplement existing validated sexuality measures with unvalidated scales or items to investigate their research topic. 1.3.2.2 Combined approaches to measurement. Most commonly, researchers have combined selected items from existing sexuality scales, which may or may not be supplemented with additional items developed as needed. It is common for these customised sexuality scales to be reported without also reporting information about the origin of the scale items. In several instances, no reliability or validity information has been provided for the included items or the scales that the items were borrowed from. For example, Penhollow et al. (2009) developed a new questionnaire using items selected from pre- existing scales and new items. However, while information about the type of data collected with the pre-existing items was reported, no information was given about the original scales, and no reliability or validity information was reported for any of the items, new or borrowed. This absence of information about the psychometric properties of these measures raises questions about how reliable or valid these results are beyond the context of the study itself. Similarly, Wang et al. (2008) incorporated the question “how many instances of sexual activity have you experienced in the last year?” as part of a larger face-to- face interview. Participant responses were examined alongside data collected using the Sexuality Knowledge and Attitudes Scale (ASKAS; White, 1982) and a specially- 16

designed Basic Characteristics Form that assessed demographic data, social characteristics, life stresses, and the capacity for participants to engage in day-to-day activities (Wang et al., 2008). While the ASKAS is a well-established and psychometrically sound measure, both the interview question and the Basic Characteristics Form were explicitly developed for the study. Of these, only the Basic Characteristics Form was evaluated for its reliability and validity using expert opinion. Thus, interpreting the results of this study requires two subjective assessments about how sexual activity has been defined; first, for the definition of sexuality used by participants, and then again for how Wang et al. (2008) have chosen to define sexual activity. Such an approach creates a high level of ambiguity, thereby increasing the potential for the misinterpretation of participant responses. 1.3.3 Methodological issues – Summary The diversity in how sexuality and related terms are defined and the limited availability of validated later life measurement scales have shaped what is known about sexuality in the later stages of life to date. While the information on sexuality and ageing produced thus far is incredibly valuable, the lack of consistency across studies limits the ability of researchers to consolidate this information systematically. Many of the issues created by these inconsistencies could be addressed by using clear, consistent definitions for sexuality and related terms. The consistent use of key terms across studies would help remove ambiguity about what areas of sexuality are under investigation, provide greater clarity about to whom the results may be generalised, and would help ensure that participants are responding to sexuality measures consistently. This added consistency would make it easier to compare results across studies and would facilitate a more systematic approach to the study of sexuality and ageing. In establishing a clear understanding of what sexuality means to adults across later life stages, additional sexuality measurement scales could also be developed that measure a broader range of topics relevant to the experience of sexuality for older adults than current measures assess. The next section describes how a lifespan development approach can be used to facilitate the development of a more nuanced and versatile description of sexuality that better reflects changes to the experience of sexuality across adulthood. 1.4 Theoretical framework of the thesis: Lifespan developmental theory Lifespan developmental theory is concerned with identifying areas of constancy and areas of enduring and reversible change (interindividual change and 17

intra-individual variability, respectively) across the lifespan (Eddy, St. Pierre, & Alles, 1982; Fingerman, Berg, Smith, & Antonucci, 2010; Schulz & Heckhausen, 1996). Approaches for examining changes across the lifespan can focus on intra- individual changes across life stages (longitudinal research), or interindividual differences across life stages (cross-sectional research). Both longitudinal and cross- sectional research have strengths and weaknesses that influence how results are interpreted. For example, longitudinal research is a practical approach for studying intra-individual changes in a way that controls for potential cohort effects on the developmental processes being investigated. However, this approach is also more susceptible to time-of-measurement effects than cross-sectional research. Conversely, cross-sectional research limits the risk of confounding variables associated with collecting data over multiple time points but does not distinguish between age-graded changes and changes that occur as a result of cohort effects. The research within this thesis uses a cross-sectional approach to understanding interindividual differences in the experience of sexuality and the factors associated with sexuality that occur across different stages of adulthood. 1.4.1 Models of development in adulthood Early lifespan developmental theories focused predominantly on the physiological and cognitive developmental milestones during gestation and occurring across infancy, childhood, and adolescence (Eddy et al., 1982; Fingerman et al., 2010; Schulz & Heckhausen, 1996). Over time, lifespan developmental theories have evolved to include a more thorough investigation of the various stages of adulthood than those provided by the early theories (Fingerman et al., 2010; Hoffnung et al., 2016). One such theory is the normative-crisis model of lifespan development. This model is used in the thesis to frame the exploration of sexuality for heterosexual men and women in mid and later life, and is described in greater detail in the next subsection. 1.4.1.1 Normative-crisis model. Within the normative-crisis model, individuals mature through a series of distinct, sequential stages of development that reflect progressively more complex developmental milestones (Hoffnung et al., 2016). Chronological age-related averages are used to compute 'average' developmental milestones during different life stages, and changes across the lifespan are to some extent, attributed to age-related biological changes. 18

One of the most well-known normative-crisis models for lifespan development is Erikson's psychosocial development model (Erikson & Erikson, 1998). A lesser- known normative-crisis model that focuses on adult development, however, is Levinson’s (1986) seasons of adult life approach. In Levinson's seasons of adult life approach, adulthood occurs across three distinct phases or 'seasons' of adulthood: early adulthood, middle adulthood, and later adulthood (Hoffnung et al., 2016). In Levinson's model, changes in adulthood occur in a consistent pattern across the three 'seasons' of adult life. Within each season, individuals form a new 'life structure' that is informed by the significant relationships they have developed with others during that life stage, and each life stage presents new opportunities for personal growth and development (Hoffnung et al., 2016). Under Levinson's approach, changes in the experience of sexuality for adults in middle and later adulthood are theoretically supported by the expectation that an individual continues to experience developmental changes that shape their life throughout middle and later adulthood. 1.4.2 Sexuality throughout the life course Sexuality within the framework of lifespan developmental theory spans well beyond topics related to reproduction, hormones, and the sexual maturation process of puberty. Sexuality is a highly subjective experience and is influenced by a broad range of biological, psychological, sociocultural, and environmental factors (Lehmiller, 2017). As with other aspects of an individual's life, sexuality is not static, and an individual's needs, desires, attitudes and abilities can change over time (Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; Gott & Hinchliff, 2003b; Hinchliff & Gott, 2004; Hurd Clarke, 2006; Ménard et al., 2014; Sandberg, 2013). Early expressions of sexuality are directly observable in infancy and childhood, and erections have even been detected in-utero (Lehmiller, 2017). The vast number of changes that occur during puberty then go on to inform the types of sexual experiences that an individual engages in as an adult (Fingerman et al., 2010; Lehmiller, 2017). However, an individual’s sexual evolution does not cease once an individual reaches adulthood, and an individual's sexuality and their sexual experiences are continually influenced by their environment, their health, and their relationships (Fingerman et al., 2010; Lehmiller, 2017). These developmental processes continue throughout midlife and well into later life, even as fertility declines

and age-related changes to sexual functioning become increasingly prominent (Carpenter & DeLamater, 2012; DeLamater, 2012; Fingerman et al., 2010). 1.4.3 Sexuality and ageing As with adolescence, hormonal changes in middle and late adulthood can impact on an individual's sexuality and sexual expression. For women, a dramatic drop in sex hormone production occurs during menopause. This dramatic change occurs most often in a woman’s early 50s (Lehmiller, 2017), and can occasionally lead to some sexual problems, such as vaginal dryness (DeLamater, 2012). In contrast, decreased sex hormone production in men is typically a more gradual process (Lehmiller, 2017). However, a reduction in the production of sex hormones for both men and women does not equate to a complete cessation of sexual interest or sexual expression (Lehmiller, 2017), and the desire to engage in various forms of sexual expression is evident well into old age (Ginsberg et al., 2005; Hyde et al., 2010; Palacios-Cena et al., 2012; Sandberg, 2013). Beyond hormonal changes during middle and late adulthood, an individual’s physical and sexual health status (and that of their sexual partner) also influences the sexual experiences of older adults. Medical conditions, such as cardiovascular disease and diabetes among others, can negatively impact sexual function and sexual expression (DeLamater, 2012; Karraker & DeLamater, 2013). For men, sexual dysfunctions such as erectile difficulties appear to be associated with age (DeLamater, 2012). However, the presence of a medically-diagnosed condition does not always prevent older adults from being able to engage with or express their sexuality, and some older adults redefine what sexuality means to them to make allowances for the presence of their own medical or sexual problems or those of their sexual partner (Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; Gott & Hinchliff, 2003b; Hinchliff & Gott, 2004; Hurd Clarke, 2006; Sandberg, 2013). Several other significant changes also occur during middle and late adulthood. These shifts occur across numerous domains, including physiological changes associated with ageing, and interpersonal or lifestyle changes (Fingerman et al., 2010; Lachman, 2004; Schulz & Heckhausen, 1996). For example, changes in lifestyle related to adult children moving out of the family home, or the onset of retirement can increase the amount of opportunity for older adults to engage in sexual expression (DeLamater, 2012). However, approaches to understanding sexuality in the later life stages that emphasise the maintenance of ‘youthful’ sexual functioning, particularly in 20

relation to the maintenance of erection rigidity (Ginsberg et al., 2005; Langer, 2009; Renshaw, 1983), do not adequately reflect attitudinal changes such as the increased importance of intimacy and connection with a partner as individuals’ age, regardless of the degree of sexual function possible (e.g., Ménard et al., 2014; Moore, 2010; Sandberg, 2013). Thus, these approaches limit the way that the sexuality of older adults is understood, and prevents a thorough examination of the ways that developmental changes over the life course can influence an individual’s sexual perceptions and preferences for various forms of sexual expression. 1.4.4 Sexuality and ageing within the normative-crisis model Under Levinson’s ‘seasons of adult life’ normative-crisis model of lifespan development (1986), midlife and later life exist as distinct, successive developmental stages. This approach recognises that biological processes are partly responsible for changes that occur at different life stages and posits an ‘average’ picture of life within these developmental stages. In this way, the normative-crisis model can be used as a means of understanding the broad experience of a particular age cohort, based on a range of developmental markers. For example, menopause in women and decreasing testosterone levels in men are primary ageing events associated with middle adulthood. These events are broadly universal, although the exact timing of these events may differ between individuals (Hoffnung et al., 2016). Thus, under Levinson’s approach to lifespan development, the experience of sexuality for adults across the lifespan is a dynamic, continuously evolving process in which the sexual priorities of an individual are guided by the different life events they experience throughout the life course. This approach is capable of accommodating the diverse range of life experiences that occur across both mid and later life, as well as individual differences in the sexual experiences and sexual priorities of adults across the later stages of life. As such, this thesis uses Levinson’s model of lifespan development as the starting point from which the experience of sexuality for adults in mid and later life is explored. In using Levinson’s model to frame the work in this thesis, the ‘average’ experience of sexuality for middle and later adulthood can be developed. The specific aims of the thesis are outlined in the following section. 1.5 Aim of the thesis The current medicalised approach to understanding sexuality throughout the lifespan has several limitations in terms of its ability to adequately reflect the 21

experience of sexuality for older adults (e.g., DeLamater, 2012; Herbenick et al., 2010). As such, the field of sexuality and ageing research would benefit from a re- evaluation of how sexuality is conceptualised and measured across the later stages of life. Specifically, the development of a new approach to how the experience of sexuality for heterosexual older adults is understood and the development of a sexuality measurement tool that captures a broader range of topics than existing scales assess would provide much-needed opportunities for the expansion of current perspectives that emphasise sexual functioning and sexual activity frequency. To date, most sexuality and ageing research has not adequately incorporated the perspectives of adults in mid and later life into how sexuality is conceptualised for this cohort. However, to successfully incorporate the experience of sexuality for adults in mid and later life and the factors associated with these experiences into a new conceptualisation of sexuality requires more comprehensive research than has been conducted thus far. To that end, the overarching aim of this thesis is to develop a novel approach to understanding and measuring sexuality for community-dwelling heterosexual adults in mid and later life that more accurately reflects the experience of sexuality and associated factors as described by this cohort. This overarching aim can be broken down into three separate sub-aims. Specifically, this thesis aims to: 1. Develop a new approach to understanding sexuality in mid and later life that reflects the lived experience of sexuality for community- dwelling heterosexual adults in mid and later life; 2. Operationalise the experience of sexuality and the factors associated with these experiences identified in (1) to produce a valid and reliable measurement scale that quantifies the lived experience of sexuality for community-dwelling heterosexual adults in mid and later life; and 3. Utilise the new approach developed in (1) and the associated scale developed in (2) to examine sex and age differences for men and women in mid and later life. These aims are addressed in the remaining chapters. The following section provides an overview of the contribution of each chapter of the thesis. 1.6 Thesis overview The current thesis conforms with the style of a thesis by publication. The nature of a thesis by publication means that some topic overlap between chapters is 22

unavoidable. This overlap occurs across chapters two to six, primarily within the background and methods sections of the included manuscripts. Chapters two to six present the five publications produced during the candidature period. Each chapter presents one of the five manuscripts that form the program of work conducted for this thesis. Each manuscript is presented in a style consistent with the overall thesis. Further descriptions of each chapter are provided below. Chapter two contains the background, aims, design, results, and discussion for a systematic review that examines how sexuality is being defined across existing later life sexuality research. The manuscript titled “Defining sexuality in later life: A systematic review” was published in a special edition of The Australasian Journal on Ageing (Macleod & McCabe, accepted). Chapter three contains the background, aims, design, results, and discussion for a systematic review that compares the sexuality topics assessed within validated later life sexuality measurement scales with the topics identified in qualitative research by heterosexual adults in later life as relevant to their perceptions of sexuality. The manuscript titled “How well do measurement scales reflect the actual experience of sexuality in midlife and beyond?” was published in Sexual Medicine Reviews (Macleod & McCabe, 2019). Chapter four contains the background, aims, design, results, and discussion for Study 1, a mixed-methods study that uses group concept mapping to develop a new conceptual framework for sexuality in mid and later life, and thematic analysis to examine the perceived changes to the experience of sexuality for heterosexual men and women aged 45 years and above. The manuscript titled “Mapping the perceived sexuality of heterosexual men and women in mid and later life: A mixed-methods study” was published in Sexual Medicine (Macleod, Busija, & McCabe, in press). Chapter five contains the background, aims, design, results, and discussion for Study 2. This paper describes the development and psychometric evaluation of the Over 45s Adult Sexuality and Intimacy Scale (OASIS) using cognitive interviewing and Rasch modelling. The paper then provides an examination of the psychometric properties of the new scale. The manuscript titled “Development and psychometric properties of the Over 45s Adult Sexuality and Intimacy Scale (OASIS)” is in preparation for submission to Psychological Methods. Chapter six contains the background, aims, design, results, and discussion for Study 3. The manuscript examines cohort differences in the experience of sexuality 23

and the factors associated with these experiences for heterosexual men and women in mid and later life based on age cohort membership and sex using data collected with the scale developed and validated in Study 2. The manuscript titled “Similarities and differences in the experience of sexuality for men and women in mid and later life” is in preparation for submission to the Journal of Sex Research. Because this manuscript references previous studies, it cannot be submitted to a journal until the manuscript in chapter five has been accepted for publication. Chapter seven provides a synthesis of the work completed in chapters two through six to address the overall aim of the thesis and each of the three sub-aims. First, the novel approach to understanding sexuality in mid and later life for heterosexual men and women is presented, before discussing the operationalisation of this approach to produce the OASIS. A discussion of areas of similarity and difference in the experience of sexuality and associated factors for men and women, and for individuals in mid and later life follows. These results are then aligned with Levinson’s normative-crisis model of lifespan development and implications of this research on the field of sexuality and ageing are explored. Finally, a summary of the limitations present throughout the studies in this thesis are considered, before recommendations for future research and general conclusions are presented. The findings from this thesis demonstrate that a clear understanding of sexuality for heterosexual men and women in mid and later life is dependent on being able to accurately represent and measure the experience of sexuality and the factors associated with these experiences, independent of the experience of sexuality for younger adults. In particular, the studies in this thesis demonstrate the importance of having an approach to understanding sexuality that is driven by the actual experiences and perspectives of heterosexual adults in mid and later life, and a psychometrically robust measurement tool that reflects these experiences.

Chapter 2. Defining sexuality in later life: A systematic review

2.1 Chapter guide Macleod, A., McCabe, M. P. (accepted). Defining sexuality in later life: A systematic review. Australasian Journal on Ageing. 00:1–10. https://doi.org/10.1111/ajag.12741

This chapter provides information relating to critical methodological issues raised in the previous chapter. Specifically, issues relating to the use of inconsistent terminology within the sexuality and ageing literature are discussed, providing additional justification for why a common lexicon would be beneficial to the field. The chapter contains a published systematic review examining how sexuality has been defined in the sexuality and ageing literature over the last 20 years. The novel contribution made by this systematic review is that it provides a thorough examination of how sexuality and ageing researchers conceptualise sexuality for older adults across publications. Themes present within both implicit and explicit definitions are examined, as are definition origins. The chapter is structured in a way that aligns both with the format of the publication and the style used in the thesis: abstract (Section 2.2), introduction (Section 2.3), method (Section 2.4), results (Section 2.5), discussion (Section 2.6) and conclusions and future directions for research (Section 2.8). At the conclusion of the current chapter, it will be demonstrated that a new approach to understanding the experience of sexuality for heterosexual men and women in mid and later life and the factors associated with these experiences would facilitate a more cohesive understanding of sexuality in the later stages of life than current approaches are able to provide.

2.2 Abstract Objective: To identify whether a consistent definition of sexuality is used across research articles to describe sexuality in later life for heterosexual men and women. Methods: A systematic review was conducted that examined how researchers define sexuality for heterosexual adults. Peer-reviewed journal articles on sexuality published between 1999-2018 were examined for the presence of sexuality definitions. Sexuality themes discussed in each article were recorded, and results tabulated. Results: Few articles explicitly provide a definition of ‘sexuality’ as it pertains to later life. Definitions from articles that defined sexuality explicitly were inconsistent across studies. Topics included in implicit and explicit definitions of sexuality were disparate across studies and included a diverse range of sexuality dimensions. Conclusion: Greater consistency is needed in how ‘sexuality’ in later life is defined across studies and should be informed by data so that it accurately reflects the sexual experiences of older adults.

Keywords: sexuality, aged, review, systematic review, consensus

2.3 Introduction Sexuality is an important part of life, regardless of age (Baldissera et al., 2012; Carpenter & DeLamater, 2012; DeLamater & Koepsel, 2014). Research shows that older adults still think of themselves as sexual beings (Jung & Schill, 2004; Langer, 2009; Lusti-Narasimhan & Beard, 2013), yet only limited research has examined the sexuality of older adults beyond assessments of sexual dysfunction. This limited body of later life sexuality research is compromised by inconsistent definitions of sexuality across studies, and differences in how key terms are operationalised (Bell et al., 2017). Given that our understanding of a topic is dependent on how it is defined and operationalised, these differences can produce conflicting results across studies, and limit our ability to consolidate knowledge effectively. The World Health Organisation defines sexuality as: "...a central aspect of being human throughout life encompasses sex, gender identities and roles, sexual orientation, eroticism, pleasure, intimacy and reproduction. Sexuality is experienced and expressed in thoughts, fantasies, desires, beliefs, attitudes, values, behaviours, practices, roles and relationships. While sexuality can include all of these dimensions, not all of them are always experienced or expressed..." (World Health Organisation, 2006). However, a cursory examination of existing later life sexuality research shows that this definition is rarely referenced in articles. Instead, multiple examples of inconsistency can be found in terms of how sexuality, sexual activity, and other key terms related to sexuality are defined across studies (e.g., Bauer et al., 2007; vs. Langer, 2009). These inconsistencies have implications not only for the types of data collected in research studies, but also for how results are interpreted, how results from different studies are compared, for the types of inferences made on the basis of these data (Bell et al., 2017), and for how sexual problems are defined and understood. The language used to engage with and discuss research is a key component of how researchers can effectively communicate their findings (Mercer, 2002). Having a shared lexicon also helps to ensure that researchers and clinicians can accurately interpret and utilise the results of research. The most basic requirement for understanding sexuality in later life is, therefore, ensuring that the way 'sexuality' and related terms are being conceptualised is consistent.

To more clearly understand the extent to which inconsistencies are present in how sexuality is being defined in later life sexuality research, a systematic review of peer-reviewed journal articles published in the last 20 years was completed. Specifically, this article aims to investigate how sexuality (independent of any diagnosed medical condition or sexual dysfunction) is defined in articles that examine sexuality or an aspect of sexuality as it pertains specifically to community-dwelling heterosexual adults in later life (i.e., aged 65 and above). Such a review will help to identify potential explanations for conflicting and inconclusive results present within the existing research, and will help identify whether the field of later life sexuality research would benefit from a more standardised approach to defining key terms, including what is meant by ‘sexuality’. By focusing on research that only examines the sexuality and sexual experiences of community-dwelling heterosexual older adults outside of the context of medically diagnosed conditions or dysfunctions, the review aims to provide an overview of how sexuality is being defined in research for non- marginalised older adults. 2.4 Methods A broad range of search terms related to (and including) ‘sexuality’ and ‘ageing’ were used in searches across PsycINFO, Medline Complete, PubMed, Scopus and Web of Science databases (see Appendix A for search terms). Search criteria were developed by the first author (AM) and then discussed and refined in consultation with the second author (MM). Figure 2.1 shows the PRISMA flow diagram relevant to this systematic review and details of the refinement process are more closely discussed in the results. After compiling all search results into EndNote X7.8 (Bld 11583, Clarivate Analytics), duplicates were removed, and all articles published more than 20 years ago were excluded. Articles were initially screened for their relevance by the first author (AM). Where any ambiguity was present as to the relevance of an article for inclusion or exclusion, the first and second author collectively examined the paper for its relevance, before a joint decision was made on its inclusion or exclusion from the review. In the first round of screening, all articles that referred to the sexuality of older adults or to sexuality and ageing in the title were retained for further screening. Because of the variations present in how 'later life', 'elderly' or 'older person' is defined across research, an article was included for further screening if it referred to the population as 'older', 'elderly', 'ageing' or related synonyms in the title, or if the 29

title reported a population that was exclusively composed of individuals aged 65 or above, as per the age cut-off points used by the United Nations (United Nations, Statistical Office, 1982). If the title of the article did not meet these criteria, the article was excluded. If the relevance of articles could not be determined at the title level, they were retained for further assessment. In the next stage of screening, article relevance was assessed using the title and abstract combined. Any articles that discussed later life sexuality within the specific context of illness, medical treatment or surgical outcome, chronic health conditions, sexual dysfunction, exclusively non-heterosexual populations, assisted living facilities, sexually transmitted infections, criminal behaviours, sex-seeking behaviours, sex education, professional attitudes, marital satisfaction, body satisfaction, quality of life, or younger age groups (e.g., young adult, undergraduates, first-time parents, adolescents, etc.) were excluded, as were any dissertations, editorials, comments, or articles solely focused on the psychometrics of a measurement scale. All remaining articles were retained for examination using the full text. If the relevance of articles could still not be determined using the abstract, these articles were also retained for further assessment using the full text. For an article to be included in the final review, it needed to discuss sexuality (or an aspect of sexuality) as it pertains to community-dwelling heterosexual older people. This cohort was targeted to highlight definitions of sexuality for non- marginalised older adults. Research examining sexuality as it relates to sexual dysfunction or medically-diagnosed conditions were excluded from the review, as evidence suggests that the presence of sexual dysfunction can change how sexuality is conceptualised and expressed by older adults (Ménard et al., 2014). Because of the variation present in how 'older adult' is used across sexuality research, a decision was made to examine articles that related to the sexuality of later life individuals as per the United Nations age cut-off points (United Nations, Statistical Office, 1982). This decision was made to ensure that the results were not inadvertently influenced by articles that incorrectly presented individuals in midlife (i.e., ages 45-64) as 'older' or 'elderly' adults. Thus, if an article did not provide separate results for the later life cohort (age 65+), it was considered out of scope. If no specific age range was provided in a theoretical and/or review paper, but the population was referred to exclusively as 'older' or 'elderly' (or a related synonym), the article was included in the

Figure 2. 1 PRISMA flowchart for original and updated systematic review 31

review. Because the focus of this paper is on the sexuality of adults in later life, original research articles were only included if they used a measure of sexuality as a dependent variable. The initial searches for this systematic review were conducted in 2016. An update of these searches was performed in December 2018. In line with this, an additional step was taken at the end of the screening process to remove any articles published prior to 1999 from the review, to ensure the review captured only the last 20 years of later life sexuality research. This was done with the expectation that any data lost with the removal of these articles would be retained within the remaining articles. Articles were examined for the use of clear definitions of the term ‘sexuality’. Topics captured within these definitions were collated. In articles where no definitions were provided, a list of sexuality topics discussed was compiled for each article. Demographic information collected from original research studies included: country of origin, sex, sample size, age range(s), and age cohort split (if any) used in the analysis. 2.5 Results In 2016, the searches yielded a total of 25,484 results. After removing duplicates (n = 3,314), 22,170 records remained. By excluding all articles published before 1996 (n = 3,741), the number of records to screen was reduced to 18,429. Screening at title and abstract levels resulted in a further 18,273 records being excluded from the review. Examination of reference lists of relevant articles revealed another eight articles to include in the screening process. After screening articles at title and abstract using the exclusion criteria, a total of 164 articles were included in the final review stage. Of the 164 articles included at the full-text review stage, 127 articles were rejected for meeting one or more of the exclusion criteria, and 15 articles could not be sourced. This left a total of 24 articles to be included in the review. An update of the searches in 2018 yielded an additional 1,271 results. After removing duplicates (n = 34), articles published prior to 1999 (n = 3), and screening at title level, a total of 144 articles published between 2016 and 2018 were reviewed against exclusion criteria at title and abstract level combined. This resulted in 34 articles that were screened for inclusion using the full text. Of these, only seven articles published between 2016 and 2018 were added to the results of the original

systematic review. Examination of reference lists of additional in-scope articles identified one additional article that was included in the review. In total, 32 articles were included in this review. During the examination of the included articles, it was evident that articles fell into one of two groups: articles that provided a clear, explicit definition of sexuality, and articles that did not explicitly define sexuality but relied on the readers’ implicit understanding of sexuality as a construct. This separation of articles into two groups informed the analysis and reporting of results for this review. Articles that did not define sexuality but relied instead on an implicit definition of sexuality will be discussed first, before discussing articles that included an explicit definition of sexuality. 2.5.1 Implicit definitions of sexuality Twenty-two articles failed to define the term 'sexuality' and instead relied on implicit definitions of sexuality. Definitions were absent in 36% of all theoretical articles (Kennedy, Martinez, & Garo, 2010; Marshall, 2010; Trudel, Turgeon, & Piché, 2010; Wylie, Wood, & McManus, 2013), 100% of all qualitative articles (Muruthi, McCoy, Chou, & Farnham, 2018; Sandberg, 2013, 2016), and 83% of all quantitative articles (Arias-Castillo et al., 2009; Beckman et al., 2008; Beckman et al., 2014; Bergeron et al., 2017; Beutel et al., 2018; Bortz & Wallace, 1999; Bortz et al., 1999; Choi, Jang, Lee, & Kim, 2011; Freak-Poli et al., 2017; Hyde et al., 2010; Stulhofer, Hinchliff, Jurin, Hald, & Traeen, 2018; Trudel et al., 2014; Twenge, Sherman, & Wells, 2017; Umidi et al., 2007; Wang et al., 2008). Despite this, most articles acknowledged the complexity of sexuality as a construct, and a broad range of topics were discussed across the included articles. Sexual activity was the only topic to appear consistently across all 22 articles that relied on an implicit definition of sexuality. Desire, passion, and/or sexual interest were referenced in 13 articles, and sexual functioning was referenced in 12 articles. Other common topics referenced in articles that relied on implicit definitions of sexuality included physical intimacy, and attitudes and/or beliefs related to sexuality. Less common topics included sexual satisfaction, arousal and/or pleasure, emotions related to sexuality, and emotional intimacy. A full list of identified themes can be found in Table 2.1. A side-by-side comparison of papers using implicit definitions of sexuality highlights the lack of consistency in discussed themes (see Table 2.1). For example,

one paper (Umidi et al., 2007) discussed sexual satisfaction, sexual motivation, and sexual desire, while another (Beckman et al., 2008) discussed physical intimacy and/or partnered activities, emotions related to sexuality, relationship satisfaction, and sexual satisfaction. 2.5.2 Explicit definitions of sexuality Definitions of sexuality were provided in only ten articles (Bauer et al., 2007; Galinsky et al., 2014; Garrett, 2014; Kessel, 2001; Langer, 2009; Müller et al., 2014; Palacios-Cena et al., 2012; Sharpe, 2004; Syme, 2014; Watters & Boyd, 2009). Definitions of sexuality ranged from broad definitions (e.g., Syme, 2014) to narrow definitions (e.g., Müller et al., 2014). No two articles provided the same definition of sexuality, and only one article discussed the process used to develop the definition used in the article (Garrett, 2014). Definitions of sexuality most often included reference to sexual expression and behaviours, physical intimacy and partnered activities, desire and/or passion, attitudes and beliefs relating to sexuality, and emotional intimacy. Less frequent inclusions were arousal and/or pleasure, sexual functioning, emotions related to sexuality, sexual satisfaction, and sexual orientation. Definitions of sexuality less commonly discussed relationship satisfaction, sensory experiences related to sex and sexuality, fantasy and thoughts about sex, reproduction-specific aspects of sexuality, or gender identity. No definitions discussed sexual knowledge. A full list of identified themes can be found in Table 2.2. Four articles referenced sexuality in older adults specifically in the provided definitions (Bauer et al., 2007; Garrett, 2014; Langer, 2009; Sharpe, 2004). In each of these, reference was made to the 'broadening' definition of sexuality for older adults to include a wider variety of sexual and affectionate behaviours. One article also specifically referenced the changing nature of sexuality as one grows older (Sharpe, 2004). A side-by-side comparison of papers that provided explicit definitions of sexuality highlights the lack of consistency in discussed themes (see Table 2.2). Explicit definitions included broad but vaguely described definitions (e.g., Palacios- Cena et al., 2012), and highly specific descriptions with detailed information about the construct and its included themes (e.g., Bauer et al., 2007).

Table 2.1

Themes discussed in articles that use an implicit definition of sexuality

Article Type

Article Theoretical Qualitative Quantitative interest Desire/passion/sexual expressions/behaviours Sexual Fantasy/thoughts activities intimacy/partnered Physical Arousal/pleasure functioning Sexual experience Sensory sexuality related to Attitudes/beliefs sexuality to related Emotions specific Reproduction satisfaction Relationship orientation Sexual intimacy Emotional identity Gender knowledge Sexual satisfaction Sexual

Arias-Castillo et al. (2009)           Beckman et al. (2008)          Beckman et al. (2014)       Bergeron et al. (2017)        Beutel et al. (2018)       Bortz & Wallace (1999)        Bortz et al. (1999)             Choi et al. (2011)     Freak-Poli et al. (2017)       Hyde et al. (2010a)       Kennedy et al. (2010)            

Article Type

Article Theoretical Qualitative Quantitative interest Desire/passion/sexual Sexual expressions/behaviours Fantasy/thoughts activities intimacy/partnered Physical Arousal/pleasure Sexual functioning experience Sensory sexuality related to Attitudes/beliefs sexuality to related Emotions specific Reproduction satisfaction Relationship Sexual orientation intimacy Emotional identity Gender Sexual knowledge Sexual satisfaction

Marshall (2010)            Muruthi et al. (2018)              Sandberg (2013)               Sandberg (2016)            Stulhofer et al. (2018)     Trudel et al. (2010)             Trudel et al. (2014)        Twenge et al. (2017)        Umidi et al. (2007)     Wang et al. (2008)     Wylie et al. (2013b)      

Table 2.2 Themes captured in explicit definitions of sexuality (by article)

Article Type

Theoretical Theoretical Qualitative Quantitative Article interest Desire/passion/sexual expressions/behaviours Sexual Fantasy/thoughts activities intimacy/partnered Physical Arousal/pleasure functioning Sexual experience Sensory sexuality related to Attitudes/beliefs sexuality to related Emotions specific Reproduction satisfaction Relationship orientation Sexual intimacy Emotional identity Gender knowledge Sexual satisfaction Sexual

Bauer et al. (2007)                  Galinsky et al. (2014)             Garrett (2014)            Kessel (2001)     Langer (2009)             Müller et al. (2014)      Palacios-Cena et al. (2012)           Sharpe (2004)          Syme (2014)                 Watters and Boyd (2009)    

2.6 Discussion Sexuality has been identified as an important aspect of wellbeing across the lifespan (Baldissera et al., 2012), yet relatively little is known about sexuality in later life. This lack of knowledge is further complicated by inconsistencies in how sexuality and related sexual concepts are defined and measured across studies (Bell et al., 2017). The purpose of this review was to examine how sexuality has been defined in later life sexuality research over the last 20 years and to determine whether these definitions are consistent in terms of the themes they capture. The results of this review show that there is limited consistency in how sexuality is being conceptualised in later life sexuality research. Implicit definitions of sexuality were used most frequently. Where explicit definitions of sexuality were provided, large variations were present in the topics discussed. No two articles used the same definition of sexuality, and only four articles explicitly referenced the sexuality of adults in later life (Bauer et al., 2007; Garrett, 2014; Langer, 2009; Sharpe, 2004). Many of the articles discussed later life sexuality without providing a clear definition of sexuality. Instead, they relied on a reader's existing understanding of sexuality to frame their discussion. The problems with the use of implicit definitions are highlighted when a side-by-side comparison of some of the included topics is made, and emphasise that while overlaps are present, there are distinct differences in the topics discussed. This pattern is seen repeatedly across the articles that rely on an implicit definition of sexuality. The use of implicit definitions of sexuality has implications for how results are compared across studies. For example, none of the included qualitative articles that relied on an implicit definition of sexuality discussed arousal and/or pleasure. However, these articles did discuss desire and/or sexual interest (Muruthi et al., 2018; Sandberg, 2016). Without the presence of an explicit definition of sexuality, it is unclear whether the authors would include arousal and/or pleasure to be a part of desire and/or sexual interest, or whether they consider these to be conceptually separate constructs. Likewise, the distinction between arousal and/or pleasure and desire and/or sexual interest was not always clear across the included quantitative articles (e.g., Bergeron et al., 2017; vs. Stulhofer et al., 2018). This makes it difficult for results to be consolidated or compared and leaves room for misinterpreting data.

This problem with comparing data across studies was also apparent in the articles that provided an explicit definition of sexuality. For example, Langer (2009) defines sexuality as “…an individual’s self-perception of being attractive as a sexual partner”. In comparison, Syme (2014) describes sexuality as “…a broad concept, encompassing interest, behaviours, functioning, satisfaction, intimate relationships, and sexual self-esteem”. While both definitions are clear, they do not appear to be describing the same construct, although both papers are referring to sexuality. Explicit definitions were both broad with vaguely described boundaries (e.g., Palacios-Cena et al., 2012), and highly specific, providing detailed information to describe the construct and its included themes (e.g., Bauer et al., 2007). This indicates that there is currently no single agreed-upon definition of sexuality in use across later life sexuality research. The use of more than one definition of sexuality within a single paper was also evident (e.g., Baldissera et al., 2012; Benbow & Beeston, 2012; Johnson, 1996; Kessel, 2001). While combining multiple definitions may be an effective way to ensure the author's conceptualisation of sexuality is clearly conveyed to readers, it also suggests that no existing definition of sexuality sufficiently describes the entire sexuality construct. This is concerning, given how often authors rely on an implied definition of sexuality. The variation across studies in terms of how sexuality is being defined and operationalised has direct implications for how these results are understood. An article by Bell et al. (2017) provides an example of how differences in the operationalisation of constructs can produce results that vary in significance or even contradict each other. In the review by Bell et al. (2017) the presence of a significant relationship between age and sexual activity, and even the significance of identified correlations between age and sexual activity, depended on how the term ‘intercourse’ was being defined (Bell et al., 2017). Many of the explicit definitions identified in this review also failed to specify whether they had been developed for use with older populations. As people age, there is a shift in focus away from intercourse frequency and towards intimacy, companionship and other forms of sexual expression (Hinchliff & Gott, 2004; Hurd Clarke, 2006). Given that much of the sexuality research has been developed based on youthful ideals around sexuality (Ginsberg et al., 2005; Langer, 2009), it is possible that some of these definitions may fail to incorporate changes to the sexual experience over the life course. The absence of information about the origin of definitions and its

relevance to older populations in the included articles raises concerns about whether existing measures of sexuality developed using these definitions are representative of the sexual experiences and attitudes specific to those in later life. The results of this review highlight the inherent problems with assuming readers share the same definitions of sexuality when using implied definitions and provide evidence that the field would benefit greatly from a more consistent approach to how sexuality is being defined. 2.6.1 Limitations At present, there is little consistency in how 'later life', 'older' and 'elderly' is being defined across the literature. To ensure that the results were genuinely reflective of adults in later life, the authors decided to use the United Nations recommendations for age cohort cut-off points (United Nations, Statistical Office, 1982). This meant that some articles that discuss sexuality in later life may have been excluded from this review. The risk of taking such a stringent approach to the review is mitigated by the fact that this review focused on how sexuality was being defined across the literature rather than specific measures of sexual experiences. Further, no comments can be made on whether common definitions are being used to describe sexuality and other key terms for younger populations because this article focuses on definitions of sexuality for those in later life only. However, the lack of consistency in how terms are defined for older adults suggests that these inconsistencies are likely to be present in sexuality research for younger cohorts as well. Articles that examined an exclusively non-heterosexual population were excluded from this review. While the data contained within this review are likely to have some relevance to LGBTQIA+ populations, there are likely to be aspects of the sexual experiences of LGBTQIA+ adults that are not relevant for heterosexual individuals, such as those related to stigma and discrimination. Given the unique sexual and relational experiences of the LGBTQIA+ community, it is important that research on definitions of sexuality is conducted separately to provide a more accurate representation of these individual experiences. Some relevant articles discussing sexuality in later life may have been missed because of poorly described titles. While an attempt was made to mitigate this risk by including any articles where relevancy was unclear in subsequent review stages, it is possible that articles that did not effectively identify the research topic in the title were incorrectly excluded from this review. Articles may also have been excluded from the review that may have been

included by others because initial screening of articles was conducted by only one author. This risk was mitigated by the approach used whereby any article that could not be definitively excluded based on the criteria agreed upon by both authors was discussed in detail by both authors before a decision was made on its inclusion. Additionally, as the updated literature search was conducted in December 2018, articles published after this date have not been included. Consequently, there may be several relevant, recently published articles examining sexuality in later life that are not present in the review. 2.7 Conclusions and future directions The range of sexuality definitions identified in this review was broad and included a variety of detailed and narrow descriptions. The consistency of language in scientific research and the clear operationalisation of measures is vital to ensure reproducibility, comparability, and effective communication of results (Fisher Jr & Stenner, 2011; Pendrill, 2013). This is especially true in research that examines an idea or concept that is vaguely described or can have multiple interpretations (Maul, Torres Irribarra, & Wilson, 2016), as is the case with the term ‘sexuality’. To ensure that later life sexuality research continues to progress with rigour and reproducibility, there needs to be greater consistency in how key terms are defined. These definitions should be developed from determining the actual sexual experience of older people to ensure that they are reflective of the experiences of the target population. To that end, research is needed that examines the aspects of sexuality that are relevant to the sexual experiences of older adults. These data can then be used to create a new definition of sexuality in later life. Developing a definition of sexuality in later life that is based on the sexual experiences of older adults will help to build consensus within the scientific community and remove the need to identify a single existing definition of sexuality in later life that everyone is willing to agree on. This will help to produce a more cohesive knowledge-base, make data comparisons more effective, and will help in the development of more rigorous, scientifically-sound measurement tools.

Chapter 3. How well do measurement scales reflect the actual experience of sexuality in midlife and beyond?

3.1 Chapter guide Macleod, A., & McCabe, M. P. (2019). How well do measurement scales reflect the actual experience of sexuality in mid-life and beyond? Sexual Medicine Reviews, 7(1), 29-45. https://doi.org/10.1016/j.sxmr.2018.04.004

The current chapter highlights specific gaps in the sexuality and ageing literature that prevent the development of a more robust understanding of sexuality for heterosexual men and women in mid and later life than those provided in current research. Specifically, the chapter aims to highlight discrepancies in how older adults describe sexuality and intimacy and the aspects of sexuality assessed in existing later life sexuality measurement scales. In doing so, it provides evidence that a new sexuality measurement scale is needed that adequately reflects the sexuality and sexual experiences of older adults. The systematic review in this chapter highlights the limited ability of current later life sexuality measurement scales to adequately capture the sexual experiences of heterosexual men and women in mid and later life, as perceived by this cohort. It examines the aspects of sexuality that have been explicitly discussed by heterosexual men and women in mid and later life within qualitative sexuality and ageing research published within the last 20 years, and contrasts this with a review of the different aspects captured by existing later life sexuality measurement scales. The published manuscript is presented in a way that aligns with its published format, and that conforms to the style of the thesis: abstract (Section 3.2), introduction (Section 3.3), method (Section 3.4), results (Section 3.5), discussion (Section 3.6), and conclusions (Section 3.7). At the conclusion of the chapter, evidence will be provided to highlight the need for a more robust later life sexuality measurement scale.

3.2 Abstract Introduction: There is limited evidence that current sexuality measures accurately reflect the experience of sexuality among older people. Aim: This review investigates whether sexuality scales that examine aspects of sexuality for heterosexual men and women in midlife and beyond accurately reflect the experiences of these individuals as discussed within qualitative studies. Methods: Online databases and reference lists were searched for articles from 1997 that investigated experiences of sexuality, ageing, and validated sexuality scales. Articles describing scales were included if they reported on a statistically validated measurement scale that was used to measure one or more aspects of sexuality for heterosexual adults aged 45 and above. Qualitative articles were included if they discussed the sexuality experiences of this population. Main Outcomes Measures: Nineteen qualitative articles and 11 sexuality scales were located. Results: Aspects of sexuality identified in the qualitative studies were sexual thoughts and fantasies, sexual desire, sexual beliefs, sexual attitudes, sexual values, sexual behaviours and practices, sexual roles, and thoughts and/or feelings about intimacy and relationships. Aspects assessed in the sexuality scales were sexual desire, sexual beliefs, sexual attitudes, sexual behaviours and practices, sexual roles, and thoughts and/or feelings about intimacy and relationships. Gaps between how sexuality at mid and later life is experienced and how it is measured in existing scales were found. Clinical Translation: Existing validated measurement tools fail to provide a nuanced picture of the sexual experiences of adults in midlife and beyond, which may impact the success of clinical interventions. This, then, is likely to limit the effectiveness of these interventions. Strengths and Limitations: This article demonstrated the need for more comprehensive sexuality measurement scales. However, it focused exclusively on the experiences of heterosexual adults aged 45 and above. Conclusion: To validly evaluate the sexual expression of older people, new measurement scales are needed that assess multiple aspects of sexuality within a single scale.

Keywords: sexuality, ageing, middle aged

3.3 Introduction Sexuality has been identified as an important aspect of wellbeing throughout the lifespan (Baldissera et al., 2012). With the average life span increasing and a greater proportion of older adults living into old age (Australian Bureau of Statistics, 2014b), it is important that the scales that are used to understand sexuality beyond the reproductive years are relevant to the experience of sexuality for older adults. The use of such scales helps ensure that our knowledge of sexuality in mid and later life, and any prescribed interventions relating to the sexual problems of those in mid and later life are based on a validated and robust understanding of sexual experiences at this life stage. There is currently limited evidence that sexuality measures accurately reflect the experience of sexuality among older people (Gott & Hinchliff, 2003b). Existing sexuality measures often focus on a model of sexuality that places penile-vaginal penetration as the 'gold standard' of sexual expression, and thus limits the measurement of other aspects of sexuality (Sanders et al., 2013). As a result, sexuality is often described as declining with age (Drench & Losee, 1996). The prevalence of ageist attitudes in society means that as men and women age, they face significant cultural pressure to conform to an asexual, or post-sexual, stereotype (Benbow & Beeston, 2012; Jung & Schill, 2004). More recently, there has been a change in perspective to a more inclusive, life- long concept of sexuality. As people age, there appears to be a shift in focus away from intercourse frequency and towards intimacy, companionship and forms of sexual expression that do not rely on penile-vaginal intercourse (Hinchliff & Gott, 2004; Hurd Clarke, 2006). Given that much of the sexuality research has been developed based on youthful ideals of sexuality (Ginsberg et al., 2005; Langer, 2009; Renshaw, 1983), it is important to ensure that the measures being used to examine sexuality for adults in mid and later life accurately reflect the aspects that men and women in midlife and beyond consider important, rather than simply focusing on measures of sexual activity frequency and sexual function. The purpose of this review is to examine how well sexuality measures reflect the actual experience of mid and later life sexuality, as defined by the World Health Organisation (WHO, World Health Organisation, 2006). This definition was chosen because it is applicable across disciplines and cultures. The WHO definition of sexuality is:

"...a central aspect of being human throughout life encompasses sex, gender identities and roles, sexual orientation, eroticism, pleasure, intimacy and reproduction. Sexuality is experienced and expressed in thoughts, fantasies, desires, beliefs, attitudes, values, behaviours, practices, roles and relationships. While sexuality can include all of these dimensions, not all of them are always experienced or expressed..."(World Health Organisation, 2006). This review compares the content of validated sexuality measurement scales currently used to quantify aspects of sexuality for heterosexual men and women in mid and later life with the aspects of sexuality that have been identified by this population within the qualitative literature. For this review, the participants were limited to heterosexual men and women over the age of 45. Because there is a lack of consistency in the literature in terms of age cut-offs used for mid and later life, the age of 45 was chosen to align with the provisional guidelines set out by the United Nations, in which midlife is considered to be between ages 45 and 64, with later life starting at age 65 (Australian Bureau of Statistics, 2014a; United Nations, Statistical Office, 1982). Journal articles published between 1997 and 2017 were examined for the qualitative and quantitative measurement of aspects of sexuality, as defined by the Organisation. 3.4 Method This review was conducted in three parts. In part one, qualitative articles that examined aspects of sexuality for community-dwelling heterosexual adults in midlife and beyond were identified. Qualitative articles were examined because they describe the sexual experiences of men and women in mid and later life in their own words, rather than using a scale to examine these constructs. In part two, a second review was conducted to identify validated measurement scales used in articles examining aspects of sexuality in midlife and beyond for community-dwelling heterosexual men and women aged 45 and above. In part three, results from the two reviews were combined to evaluate how well existing validated measurement scales captured the sexual experiences of heterosexual men and women aged 45 and above captured in the qualitative literature. To identify as many relevant articles as possible for this review, a broad range of search terms related to (and including) ‘sexuality’ and ‘ageing’ were used in searches across Academic Search Complete, CINAHL Complete, MEDLINE

Complete, PsycINFO, PsycTESTS, SocINDEX, PubMed, Scopus and Web of Science databases. Figure 3.1 shows the PRISMA flow diagram relevant to the search for articles and measurement scales. Details of the refinement process are discussed in the results. For a qualitative article to be included in the review, it had to provide a novel exploration of one or more aspects of sexuality as defined in the WHO working definition of sexuality for community-dwelling heterosexual adults aged 45 or above only and be published within the last 20 years. For a quantitative article to be included in the review, it had to include a validated measurement scale that measured one or more aspects of sexuality as defined in the WHO working definition of sexuality for community-dwelling heterosexual populations aged 45 or above. Based on the WHO definition of sexuality, the following categories were used to define aspects of sexuality for this review: sexual thoughts and fantasies, sexual desire (including arousal and passion), sexual beliefs (including personal perspectives and knowledge), sexual attitudes (including satisfaction and importance), sexual values (including moral standards towards sexuality), sexual behaviours and practices, sexual roles, and personal thoughts and feelings related to intimacy and relationships (see Table 3.1 for a full list of inclusion and exclusion criteria). If any ambiguity was present regarding the relevance of an article, it was included in the full-text review stage for further assessment. Articles that examined a clinical sample only were not included, as this paper aimed to examine the experiences of the wider ageing population as a whole. Likewise, articles that included non-community-dwelling individuals, or discussion of attitudes towards the target population (including professional perspectives) were excluded from the analysis. Articles were only included if the sample did not include LGBTQIA+ populations, as gender identity and sexual orientation were not considered within the scope of this review. Articles related to sexual functioning were also excluded from this review, as the measurement of sexual functioning remains consistent across the lifespan (e.g., the International Index of Erectile Function [IIEF] and the Female Sexual Function Index [FSFI]; C. Rosen et al., 2000; R. C. Rosen et al., 1997). All included articles were assessed for their quality using the National CASP Appraisal Tool for Qualitative Studies (Critical Appraisal Skills Program) or the NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies

Figure 3. 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart of searches for articles examining aspects of sexuality at midlife and later life

(National Institutes of Health). Qualitative articles received a quality score out of ten (where 1-3/10 = low quality; 4-7/10 = moderate quality; 8-10/10 = high quality). Articles discussing measurement scales received a corrected quality score out of eight (instead of 12), as six questions were not applicable based on study design (where 1-3/8 = low quality; 4-6/8 = moderate quality; 7-8/8 = high quality). Because sexuality for this age cohort is understudied, no articles were excluded based on their assessed quality. 3.5 Results The first search for both the aspects of sexuality and the measurement scales yielded a total of 6,155 results. After removing duplicates (n = 1,310), 4,845 records remained. The second search yielded a total of 1,802 results. After removing duplicates (n = 213), 1,589 records remained and were combined with the results of the first search. Combining the results from both searches led to an additional 1,578 duplicates being removed, leaving a total of 4,856 articles to be screened for inclusion in the systematic review. Screening at title and abstract levels resulted in the exclusion of 4,690 records from the review. After screening articles by comparing title and abstract against the exclusion criteria, a total of 166 articles were included in the final review stage. A further four articles were identified through additional sources, resulting in 170 articles included in the full-text review stage. Using the inclusion and exclusion criteria outlined in Table 3.1, articles were examined for their relevance. Of the 170 articles included at the full review stage, only 27 articles were included in the final analysis, with 18 articles that qualitatively examined aspects of sexuality in mid and later life, eight articles that used validated measurement scales to assess aspects of sexuality in mid and later life, and one article that assessed aspects of sexuality in mid and later life both qualitatively and using a validated measurement scale. Quality assessments of all included articles can be found in Tables 3.2 and 3.3. A number of articles containing valuable research on sexuality in midlife and beyond were excluded from this review based on the specific criteria used. For instance, 40 articles were excluded because they did not identify the sexual orientation of the sample, 31 articles were excluded because they included non-heterosexual individuals in the overall analysis or were focused exclusively on non-heterosexual individuals, and nine articles were excluded because the definition of 'midlife' used in the article included adults aged 40 and above.

Table 3.1 Inclusion and exclusion criteria Include if… Articles Scales The peer-reviewed journal article uses qualitative The scale was a validated measurement scale that research methods to explore one or more aspects of measured one or more aspects of sexuality as defined sexuality as defined in the WHO working definition in the WHO working definition of sexuality for of sexuality for community-dwelling heterosexual community-dwelling heterosexual populations aged adults aged 45 or above 45 or above

The peer-reviewed journal article was published in the last 20 years

Exclude if… Qualitative Articles Articles with Scales The article uses quantitative research methods only to The scale was designed to measure sexual dysfunction explore sexuality The article focuses exclusively on sexuality as it The scale was designed to only measure specific relates to dysfunction, illness, disability, or disease biological functions related to sexuality such as physiological arousal The article focuses on specific non-community-based The measurement method used is not a scale populations (e.g., hospitals, aged care facilities) or highly unique populations (e.g., Olympic athletes, veterans)

The article was published more than 20 years ago The scale or measurement method used was not tested for validity or reliability The article focuses on non-heterosexual populations, The scale was designed to measure sexuality within or includes non-heterosexual populations in the the context of an Aged Care or hospital setting analysis as one group

The article does not specify the population as only The scale measures factors associated with sexuality heterosexual but does not measure aspects of sexuality as defined in the WHO working definition of sexuality

The article focuses on a discussion of attitudes towards mid and later life sexuality by those not within the target population (e.g., health professionals or those in younger age groups)

The article fails to analyse participants aged 45+ as a separate subgroup

The article does not present novel data (e.g., book chapter or review, editorial, correction, or comments paper, literature review)

The article focuses on measures of sexual function/biology only

The article examines non-human sexuality

The article examines sexuality as a non-sexual gender-based construct only

The article focuses on the evaluation of treatment outcomes The article examines factors associated with sexuality but does not examine aspects of sexuality as defined in the WHO working definition of sexuality

Because this review was concerned specifically with articles that used validated measurement scales to assess aspects of sexuality in midlife and beyond, 30 articles were also excluded for using a non-validated measurement scale. This includes well-known articles that used national survey data (e.g., articles that used data from the National Social Life, Health, and Aging Project). In total, 143 articles were considered out of scope on further examination of the text. Only scales that quantified aspects of sexuality were included in the analysis. Any validated measurement scales that were used to assess sexual function and/or factors associated with sexuality were not considered within the scope of this review. 3.5.1 Aspects of sexuality identified within the qualitative articles The age of participants within the qualitative articles ranged from 48 to 92 years. Twelve articles examined females only, one article examined males only, three examined both male and females, and two articles examined heterosexual dyads in the form of married or partnered couples. Studies were conducted across a range of countries including Brazil, Canada, Australia, UK, USA, Korea, Iran, Israel, and Sweden, although studies were predominantly Western (see Table 3.4 for further details). All aspects discussed in the WHO definition of sexuality were present within the qualitative literature. The most frequently discussed aspect of sexuality was sexual desire (Drummond et al., 2013; Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; Gonçalves & Merighi, 2009; Gott & Hinchliff, 2003b; Hinchliff & Gott, 2004, 2008; Hurd Clarke, 2006; Koren, 2014; Lagana & Maciel, 2010; Lodge & Umberson, 2012; Loe, 2004; Ravanipour, Gharibi, & Gharibi, 2013; Sandberg, 2013; Watson, Bell, & Stelle, 2010; Watson & Stelle, 2011; Wood et al., 2007; Youn, 2009; Yun, Kim, & Chung, 2014). This was closely followed by sexual behaviours and practices (Drummond et al., 2013; Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; Gonçalves & Merighi, 2009; Gott & Hinchliff, 2003b; Hinchliff & Gott, 2004, 2008; Hurd Clarke, 2006; Koren, 2014; Lagana & Maciel, 2010; Lodge & Umberson, 2012; Loe, 2004; Ravanipour et al., 2013; Sandberg, 2013; Watson et al., 2010; Watson & Stelle, 2011; Wood et al., 2007; Yun et al., 2014). These topics, however, were often discussed within the context of other sexuality aspects. For example, in the study by Hinchliff and Gott (2004), the discussion of sexual behaviours occurred within the broader context of relationship dynamics.

Table 3.2 Quality assessment of included qualitative articles

n toappropriate address the aims of

atelyconsidered?

Was there clearstatementa of the ofaims theresearch? qualitativea Is methodology appropriate? Was the research desig the research? Was the recruitment strategy appropriate to the aims of the research? Were the datacollected in way a that addressed theresearch issue? theHas relationshipbetweenresearcher and participants been adequ Have ethical issues been intotaken consideration? Was the data analysis sufficiently rigorous? thereIs clear statementa of findings? theIs valuable research a contribution to knowledge? Article Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Score Baldissera, Bueno & Hoga. Improvement of older women's sexuality through Y Y Y Y Y N Y Y Y Y 9 emancipatory education (2012)

Drummond, et al. The impact of caregiving: Older women's experiences of Y Y Y Y Y N Y Y Y Y 9 sexuality and intimacy (2013).

Fileborn, Thorpe, Hawkes, Minichiello & Pitts. Sex and the (older) single girl: N Y Y N Y N N Y Y Y 6 Experiences of sex and dating in later life. (2015) Gonçalves & Merighi. Reflections on Y Y Y N Y N Y N N Y 6 sexuality during the climacteric. (2009) Gott & Hinchliff. How important is sex in later life? The views of older people. Y Y Y Y Y Y Y Y Y Y 10 (2003b) Hinchliff & Gott. Intimacy, commitment, and adaptation: Sexual relationships Y Y Y Y Y Y Y Y Y Y 10 within long-term marriages. (2004) Hinchliff & Gott. Challenging social myths and stereotypes of women and Y Y Y Y Y N Y Y Y Y 9 aging: Heterosexual women talk about sex. (2008) Hurd Clarke. Older women and sexuality: experiences in marital relationships across Y Y Y Y Y N N Y Y Y 8 the life course. (2006) Koren. Together and apart: A typology of Y Y Y Y Y N Y Y N Y 8 re-partnering in old age. (2014) Lagana & Maciel. Sexual desire among Mexican-American older women: A N Y Y Y Y N Y Y Y Y 8 qualitative study. (2010) Lodge & Umberson. All shook up: Sexuality of mid‐ to later life married Y Y Y Y Y N N Y Y Y 8 couples. (2012) Loe. Sex and the senior woman: Pleasure N Y Y Y Y N N Y Y Y 7 and danger in the Viagra era. (2004)

n appropriate to address the ofaims

ately considered? ately considered?

Was there a clear statement of the of aims the research? qualitative Is a methodology appropriate? Was the research desig the research? Was the recruitment strategy appropriate to aims the of the research? Were thedata incollected that addressed a theway research issue? theHas relationshipbetween and researcher participants been adequ Have ethical issues been taken consideration?into Was the data analysis sufficiently rigorous? Is there clearstatement a of findings? Is the research a valuable contribution to knowledge? Article Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Score Ravanipour et al. Elderly women's views about sexual desire during old age: A Y Y Y N Y N Y Y Y Y 8 qualitative study. (2013) Sandberg. Just feeling a naked body close to you: Men, sexuality and intimacy in Y Y Y Y Y N N Y Y Y 8 later life. (2013) Watson & Stelle. Dating for older women: Experiences and meanings of Y Y Y Y Y N N Y Y Y 8 dating in later life. (2011) Watson et al. Women narrate later life remarriage: Negotiating the cultural to Y Y Y Y Y N N Y Y Y 8 create the personal. (2010) Wood et al. Negotiating sexual agency: Postmenopausal women's meaning and Y Y Y Y Y N N Y Y Y 8 experience of sexual desire. (2007) Youn. Marital and sexual conflicts in Y Y Y Y Y N N N Y Y 7 elderly Korean people. (2009) Yun et al. The sexuality experience of Y Y Y Y Y N Y Y Y Y 9 older widows in Korea. (2014)

Table 3.3 Quality assessment of included quantitative articles

once over time? time? once over

-up after baseline 20% or less? 20% -up after baseline applied uniformly to all participants? to all participants? applied uniformly

e.g., categories of exposure, or exposure of exposure, categories e.g.,

participation rate of eligible persons at least 50%? least 50%? rate at of participation persons eligible sufficient so that couldreasonably one timeframe

paperor this questionin clearly the objective Was research stated? specifieddefined? andpopulation the clearly Was study the Was or from or the selected the same recruited all Were subjects period)? time Were populationsthe same similar (including being criteria thestudyfor inclusion in exclusion and prespecified and sample description, or justification, size a Was power provided? and estimates variance effect of the exposure(s) inpaper, were For this the analyses being prior theoutcome(s) measured? interest tomeasured the Was and outcome exposure between expect anto see association if it existed? level,did the varyor that canamount For exposures in therelated of differentas to studyexposure levels examine ( the outcome continuous as measured variable)? (independent variables) clearly the exposure Were measures consistently implemented across defined, valid,and reliable, all study participants? more than the Was assessed exposure(s) (dependent variables) the outcome measures Were clearly consistently implemented across defined, valid,and reliable, all study participants? tothe status blinded exposure the outcome assessors Were of participants? to loss Was follow variables measured and confounding key Were potential on impact the adjustedrelationship for their statistically and outcome(s)? between exposure(s) Reference Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Score

Flynn & Gow. Examining associations between sexual Y Y Y Y N N.A. N.A. N.A. Y N.A. Y N.A. N.A. N 6 behaviours and quality of life in older adults. (2015) Gagnon, Hebert, Leclerc & Lefrancois. Development and validation of a sexual actualization measuring instrument for Y Y N Y N N.A. N.A. N.A. N.A. N.A. Y N.A. N.A. N 4 the elderly. (2002) Gott & Hinchliff. How important is sex in later life? The Y Y N Y N N.A. N.A. N.A. Y N.A. Y N.A. N.A. N 5 views of older people. (2003b)

Harper, Schaalje & Sandberg. Daily hassles, intimacy, and Y Y N Y N N.A. N.A. N.A. Y N.A. Y N.A. N.A. N 5 marital quality in later life marriages. (2000)

once time? over

-up after baseline 20% or less? or 20% -up less? after baseline applied uniformly to all participants? allparticipants? to applied uniformly

e.g., categories of exposure, or exposure of exposure, categories e.g.,

participation rate of eligible persons at least 50%? 50%? rateat least ofparticipation persons eligible oneso thatcouldsufficient timeframe reasonably

Was the research question or objective paper questionin clearlyorthis the objective Was research stated? andpopulationspecified defined? the clearly Was study the Was or the selectedsame all orWere the recruited from subjects period)? time Werepopulations same similar (including the being criteriathefor inclusion in and exclusion study prespecified and orjustification, size a Was power sample description, provided?variance and estimates effect of the paper, exposure(s) were For this analyses in the being prior outcome(s) interestto the measured? measured the Was outcome exposure between and expect tosee association an if itexisted? did thelevel, varyor that amount For in can exposures of the to study exposurerelated levelsas examine different the outcome ( continuous measured as variable)? (independent variables)theexposure clearly Were measures across consistently defined, implemented valid,and reliable, all studyparticipants? than more the Was assessed exposure(s) (dependent variables) theoutcome measures Were clearly across consistently defined, implemented valid,and reliable, all studyparticipants? the statusblinded to exposure theoutcome assessors Were of participants? loss Was to follow measured and variables key Were potential confounding onimpact the adjusted relationship for their statistically outcome(s)? and between exposure(s) Reference Q1 Q2 Q3 Q4 Q5 Q6 Q7 Q8 Q9 Q10 Q11 Q12 Q13 Q14 Score

Kazer et al. Validity and reliability of the geriatric sexuality Y Y N.A. Y N N.A. N.A. N.A. N.A. N.A. Y N.A. N.A. N 4 inventory. (2013) J. S. Kim & Kang A study on body image, sexual quality of life, depression, and quality of life in middle-aged adults. Y Y Y Y Y N.A. N.A. N.A. Y N.A. Y N.A. N.A. N 7 (2015) Mazo & Cardoso. Sexual satisfaction and correlates among Y Y Y Y N N.A. N.A. N.A. Y N.A. Y N.A. N.A. N 6 elderly Brazilians. (2011) Shkolnik & Iecovich. Health, body image, gender, and migration status: Their relationship to sexuality in old age. Y Y N Y N N.A. N.A. N.A. Y N.A. Y N.A. N.A. N 5 (2013) E. Weinstein. Senior adult sexuality scales. (2013) Y Y Y Y N N.A. N.A. N.A. Y N.A. Y N.A. N.A. N 6

Table 3.4 Demographic details of qualitative articles and the aspects they discussed Sexual thoughts & fantasies & thoughts Sexual fantasies (includesdesire arousal/passion) Sexual knowledge) beliefs perspectives, (e.g., Sexual importance/satisfaction) (e.g., Sexual attitudes (e.g.,values moral standards) Sexual & practices behaviours Sexual (e.g., masculinity/femininity) Sexual roles relationships & feelings Thoughts/ intimacy re:

Theme count 2 18 12 12 13 17 12 9 Mean Age Article Sample Age Range Country Approach Aspects (s.d.) Baldissera, Bueno & Hoga. Improvement of older women's sexuality through emancipatory 6F 62-73 - Brazil Action Research (AR)       education (2012)

Drummond, et al. The impact of caregiving: Older women's experiences of sexuality and 6F 60-80 - Canada Phenomenological approach         intimacy (2013).

Fileborn, Thorpe, Hawkes, Minichiello & Pitts. Sex and the (older) single girl: Experiences of 15F 55-81 - Australia None provided        sex and dating in later life. (2015) Gonçalves & Merighi. Reflections on sexuality 7F 48-55 - Brazil Phenomenological approach      during the climacteric. (2009)

Sexual thoughts & fantasies thoughts Sexual fantasies & desire arousal/passion) Sexual (includes beliefs perspectives, knowledge) (e.g., Sexual (e.g., Sexual attitudes importance/satisfaction) valuesstandards) Sexual (e.g., moral practices behaviours & Sexual masculinity/femininity) Sexual roles (e.g., & relationships feelings Thoughts/ intimacy re:

Theme count 2 18 12 12 13 17 12 9 Mean Age Article Sample Age Range Country Approach Aspects (s.d.) Gott & Hinchliff. How important is sex in later 69 50-92 - UK Between-methods triangulation        life? The views of older people. (2003b) (21M; 23F)

Hinchliff & Gott. Intimacy, commitment, and 28 Thematic analysis "Framework" adaptation: Sexual relationships within long- 50-86 - UK        (8M; 10F) (National Centre for Social Research) term marriages. (2004)

Hinchliff & Gott. Challenging social myths and stereotypes of women and aging: Heterosexual 19F 50-83 61.9 UK Foucauldian discourse analysis       women talk about sex. (2008)

Hurd Clarke. Older women and sexuality: experiences in marital relationships across the 24F 52-90 - Canada Grounded theory        life course. (2006)

Koren. Together and apart: A typology of re- 20 dyads 66-92 - Israel None provided        partnering in old age. (2014)

Theme count 2 18 12 12 13 17 12 9 Mean Age Article Sample Age Range Country Approach Aspects (s.d.) Lagana & Maciel. Sexual desire among Mexican-American older women: A qualitative 25F 59-89 - USA None provided       study. (2010) Lodge & Umberson. All shook up: Sexuality of 17 dyads 50-86 - USA Grounded theory        mid‐ to later life married couples. (2012) Loe. Sex and the senior woman: Pleasure and 8F 67-86 - USA Grounded theory        danger in the Viagra era. (2004) Ravanipour, Gharibi, & Gharibi. Elderly women's views about sexual desire during old 15F 60+ 67 Iran Content analysis       age: A qualitative study. (2013)

Sandberg. Just feeling a naked body close to you: Men, sexuality and intimacy in later life. 22M 67-87 - Sweden Thematic analysis        (2013)

Theme count 2 18 12 12 13 17 12 9 Mean Age Article Sample Age Range Country Approach Aspects (s.d.) Watson & Stelle. Dating for older women: Experiences and meanings of dating in later 8F 65-80 - USA Interpretive Phenomenological Analysis (IPA)        life. (2011) Watson et al. Women narrate later life remarriage: Negotiating the cultural to create 14F 64-77 - USA Phenomenological approach       the personal. (2010)

Wood et al. Negotiating sexual agency: Postmenopausal women's meaning and 22F 58-65 62 USA Grounded theory        experience of sexual desire. (2007)

Youn. Marital and sexual conflicts in elderly 24 73.8 68-79 Korea Grounded theory     Korean people. (2009) (12M; 12F) (3.2)

Yun et al. The sexuality experience of older 10F 65-75 68 Korea Thematic analysis     widows in Korea. (2014)

The aspects of sexuality discussed least often were sexual thoughts and fantasies (Drummond et al., 2013; Lagana & Maciel, 2010). All other aspects of sexuality identified within the WHO definition of sexuality appeared in at least half of the included articles. A full list of aspects identified within the included articles can be found in Table 3.4. 3.5.2 Aspects of sexuality measured using validated measurement scales The age of participants in the articles using a validated measurement scale to examine aspects of sexuality ranged from 45 to 94 years. One article examined females only, no articles examined males only, seven articles examined both male and females, and one article examined heterosexual dyads in the form of married or partnered couples. Studies were conducted across a range of countries including the UK, Canada, Brazil, Korea, USA and Israel (see Table 3.5 for further details). All but two aspects discussed in the WHO definition of sexuality were measured using validated measurement scales. The most frequently measured aspect of sexuality was sexual attitudes (Gagnon, Hébert, Leclerc, & Lefrançois, 2002; Gott & Hinchliff, 2003b; Kazer et al., 2013; J. S. Kim & Kang, 2015; Mazo & Cardoso, 2011; Shkolnik & Iecovich, 2013). The second-most frequently measured aspect of sexuality was sexual behaviours and practices (Flynn & Gow, 2015; Shkolnik & Iecovich, 2013; E. Weinstein, 2013). Eleven validated measurement scales were used to quantify aspects of sexuality in the quantitative literature. A full list of the scales used and their corresponding Cronbach’s alpha scores for each study can be found in Table 3.5. Of these scales, only three were validated using exclusively older populations (SAQ, Gagnon et al., 2002; GSI, Kazer et al., 2013; SASS, E. Weinstein, 2013; S. Weinstein & Rosen, 1988). Variations of the World Health Organisation Quality of Life scales were used most frequently - the WHOQOL-100 including the WHOQOL Importance Scale (Gott & Hinchliff, 2003b); and the WHOQOL-BREF (J. S. Kim & Kang, 2015; Mazo & Cardoso, 2011). No other scales were present within more than one article. None of the measurement scales measured more than three aspects of sexuality, with six scales assessing only one aspect of sexuality (Flynn & Gow, 2015; Gott & Hinchliff, 2003b; J. S. Kim & Kang, 2015; Malloy, 2010; Mazo & Cardoso, 2011; Shkolnik & Iecovich, 2013), two scales assessing two aspects of sexuality(Gagnon et al., 2002; J. S. Kim & Kang, 2015), and two scales assessing three aspects of sexuality (Kazer et al., 2013; E. Weinstein, 2013). The aspects of

Table 3.5 Demographic details of quantitative articles, the validated measurement scales used, and the aspects they measure

Sexual thoughts & fantasies & fantasies thoughts Sexual arousal) desire (includes Sexual knowledge) perspectives, (e.g., beliefs Sexual importance/satisfaction) (e.g., attitudes Sexual moral standards) (e.g., values Sexual practices & behaviours Sexual masculinity/femininity) (e.g., roles Sexual & relationships intimacy re: feelings Thoughts/

Theme count 0 2 2 7 1 4 0 1 Mean Age Article Design Sample Age Range (s.d.) Country Scale Cronbach’s α Aspects Flynn & Gow. Examining associations 133 74 Sexual Behaviour between sexual behaviours and quality of Cross-sectional - UK 0.84   (73M, 65F) (7.1) Frequency Scale life in older adults. (2015) Sexual Gagnon et al. Development and validation M = 68.9 Scale validation 115 Actualization of a sexual actualization measuring - Canada 0.97     study (63M, 52F) Questionnaire instrument for the elderly. (2002) F = 67.3 (SAQ) Gott & Hinchliff. How important is sex in 0.8 *(Sexual 44 later life? The views of older people. Cross-sectional 50-92 - UK WHOQOL-100 Activity Facet   (21M, 23F) (2003b) only) Personal Harper et al. Daily hassles, intimacy, and Assessment of marital quality in later life marriages. Cross-sectional 472 dyads 55-75 - USA Intimacy in 0.7  (2000) Relationships (PAIR) Kazer et al. Validity and reliability of the Scale validation Geriatric Sexuality 34 60-91 - USA 0.74     geriatric sexuality inventory. (2013) study Inventory (GSI)

Sexual thoughts & fantasies & fantasies thoughts Sexual arousal) (includes desire Sexual knowledge) perspectives, (e.g., beliefs Sexual importance/satisfaction) (e.g., attitudes Sexual moral standards) (e.g., values Sexual practices & behaviours Sexual masculinity/femininity) (e.g., roles Sexual & relationships intimacy re: feelings Thoughts/

Theme count 0 2 2 7 1 4 0 1 Mean Age Article Design Sample Age Range (s.d.) Country Scale Cronbach’s α Aspects J. S. Kim & Kang A study on body image, 367 SQOL (Korean Cross-sectional 45-60 - Korea 0.91     sexual quality of life, depression, and (181M, 186F) Language) quality of life in middle-aged adults. (2015)  WHOQOL- BREF 0.94  (Korean Language)

0.82, 0.81, 0.68, Mazo & Cardoso. Sexual satisfaction and Cross-sectional 198F 65+ - Brazil WHOQOL-BREF 0.80 (4   correlates among elderly Brazilians. (2011) domains)  Sexual Activity Shkolnik & Iecovich. Health, body image, 200 Index (customised gender, and migration status: Their Cross-sectional 61-94 73 (6.65) Israel 0.84  (45% M, 55% F) from Ginsberg et al. relationship to sexuality in old age. (2013) 2005) Index of Sexual 0.94   Satisfaction

Senior Adult 0.87, 0.90, 0.84, E. Weinstein. Senior adult sexuality scales. Scale validation 314 60+ - USA Sexuality Scales 0.66 (subscales    (2013) study (SASS) A-D)

sexuality that were not assessed using validated measurement scales included sexual thoughts and fantasies, and sexual roles. All other aspects of sexuality identified within the WHO definition of sexuality appeared in only one or two of the included articles. A full list of all aspects identified within the included articles can be found in Table 3.5. 3.6 Discussion Although the current body of research is clearly capturing important information about the sexuality of those in midlife and beyond, existing validated measurement scales fail to capture the full spectrum of sexuality experiences. Exacerbating this problem is the limited range of validated measurement scales available to researchers that assess aspects of sexuality for adults in mid and later life. This review has identified that there are currently gaps in how sexuality aspects in mid and later life are being measured when compared to the aspects of sexuality described by heterosexual men and women aged 45 and above. This has implications for how sexuality in mid and later life is understood and for how sexual problems in this cohort are diagnosed and managed. All themes within the WHO definition of sexuality were discussed in the qualitative articles, and all but two themes were present within the measurement scales. No measurement scales assessed sexual thoughts and fantasies, or sexual roles of heterosexual men and women aged 45 and above, despite their presence in the qualitative literature. Although scales do exist that measure sexual thoughts and fantasies (e.g., Sexual Daydreaming Scale of the Imaginal Processes Inventory, Giambra & Singer, 1998) and sexual roles (e.g., Bem Sex Role Inventory, Malloy, 2010), these have been developed using predominantly or exclusively younger populations, so may be skewed towards the sexual experiences of younger adults. Although sexuality aspects identified differ between studies, it is clear that if given the opportunity to discuss sexuality, adults in mid and later life discuss a broad range of their sexual experiences. This contrasts sharply with attitudes that portray older adults as asexual (Benbow & Beeston, 2012; Jung & Schill, 2004), or that see older adults as unwilling to discuss sexuality (Giambra & Singer, 1998). This suggests that a key obstacle to understanding sexuality in mid and later life comes from misguided perceptions of older adults being either asexual or reticent to embrace their sexuality.

In contrast to the breadth of sexuality aspects captured in the qualitative studies, the scales identified within this review only captured a limited range of sexuality aspects. Scales developed and validated using younger populations focus on only one or two areas. In contrast, two of the three scales developed and validated using older populations included the largest range of sexuality aspects (GSI, Kazer et al., 2013; SASS, E. Weinstein, 2013; S. Weinstein & Rosen, 1988). However, at best, these scales only capture three areas of sexuality. This shows that the scales that are available fail to adequately capture the complexity of sexuality in mid and later life. Such a discrepancy is likely to result in only a partial picture of the sexual experiences of older adults. Thus, if existing approaches to understanding and treating sexual problems have been developed using scales that provide only a partial picture, it follows that such treatments are likely to only partially address these problems and that there are aspects of sexual problems that are not adequately being evaluated and treated in sexual medicine. Aspects of sexuality, such as sexual roles, sexual beliefs and sexual values were often linked within the qualitative literature yet were either assessed in a limited way by sexuality scales or not assessed at all. For example, sexual roles were discussed often in the qualitative literature but were not evaluated in any measure (Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015), and there is evidence that suggests the sexual experiences of men and women in mid and later life become more similar over time (Gott & Hinchliff, 2003b; Hinchliff & Gott, 2008; Sandberg, 2013). Valuable quantitative information about the association between sexual roles and the sexual experience of older adults is missing from our current understanding of sexuality in mid and later life. Overall, it was found that validated measurement scales that assess sexual desire, sexual beliefs, sexual attitudes, sexual values, sexual behaviours and practices, and thoughts and/or feelings about intimacy and relationships are currently in use. No scales that assess the sexual roles or the sexual thoughts and fantasies of adults in mid and later life were identified. At best, only three of the above areas are measured using a single scale. Thus, for a robust assessment of multiple aspects of sexuality to occur using validated scales, several measurement scales would be needed, increasing the length of the questionnaire, and increasing the risk of participant burden.

3.6.1 Limitations There are several limitations with this review. At present, there is a lack of consistency in the literature for what ages are used by researchers to capture 'midlife' and 'later life'. By defining midlife as 45-64 years, as per the UN guidelines, articles that defined midlife to be 40+ years were excluded. Several articles were also excluded because they failed to specify the sexual orientation of the sample, and this information could not be determined through an analysis of the provided data. In addition, several well-known national surveys were not included in this review. This was because the focus of this article was on the use of validated measurement scales. However, by not including these studies, there may be other aspects of sexuality assessed using quantitative techniques that were not captured in this review. This review is only reflective of heterosexual adults in midlife and beyond and may not be more broadly reflective of the general population, which includes those who do not identify as heterosexual. Because of the small number of relevant articles identified, articles were not excluded or weighted based on their quality. This may affect the interpretation of results, as all data is being treated equally in the analysis. Finally, some interpretation by the authors was necessary to align aspects identified within the literature into the categories developed from the WHO definition. While these interpretations were made as objectively as possible, there is always a risk that other readers may classify certain aspects differently to the authors. 3.7 Conclusion The results of this review show that existing measurement scales developed to assess sexuality in adults during mid and later life fail to include important aspects of sexuality, as identified by heterosexual men and women aged 45 and above. In particular, no validated measurement scale was found that measured sexual thoughts and fantasies, or sexual values. Validated sexuality measurement scales were also found to be narrow in scope. It is clear from this review that additional validated sexuality measurement scales are needed that reflect the perspectives and experiences of adults aged 45 and above that include a greater number of sexual experiences within a single instrument. In particular, validated scales that assess sexual thoughts and fantasies, and sexual roles are needed to help quantify aspects of sexuality only captured currently within the qualitative literature. The expansion of such scales will be of benefit to the medical profession, by providing a more detailed picture of sexuality for heterosexual men and women in mid and later life, thus providing

additional avenues for professionals to address the sexual problems of their patients. Future research is also needed to explore sexuality in midlife and beyond among non- heterosexual populations, as there are likely to be factors related to stigma and discrimination associated with non-heterosexuality and ageing that may influence the way in which sexuality is experienced and expressed within these groups.

Chapter 4. Study 1: Mapping the perceived sexuality of heterosexual men and women in mid and later life: A mixed-methods study

4.1 Chapter guide Macleod, A., Busija, L., & McCabe, M. P. (in press). Mapping the perceived sexuality of heterosexual men and women in mid and later life: A mixed-methods study. Sexual Medicine, 8(1), 84-99. doi: http://dx.doi.org/10.1016/j.esxm.2019.10.001

This chapter provides important information about how heterosexual men and women in mid and later life perceive the experience of sexuality and the factors associated with these experiences. The novel contribution made by this chapter is the development of a new approach to understanding the experience of sexuality and the factors associated with these experiences that more accurately reflects the perspectives of the target cohort than current approaches. Group concept mapping and thematic analysis are used to describe how men and women aged 45 years and above understand their own experience of sexuality, intimacy and desire. Representative statements are collated, and a statistically generated ‘map’ of the perceived relationships between these topics as determined by men and women in mid and later life is presented. The current chapter contains a published manuscript that is presented to conform to the style of the thesis and includes the abstract (Section 4.2), introduction (Section 4.3), method (Section 4.4), results (Section 4.5), discussion (Section 4.6), and conclusions and future research (Section 4.7). At the conclusion of the current chapter, it will be clear that the development of a new approach to how sexuality in mid and later life is understood will help to bridge the gap between sexuality topics highlighted in the available qualitative literature and current approaches for measuring sexuality in mid and later life. Ethics approval and related documentation for this study can be located in Appendix B.

4.2 Abstract Background: There is currently limited research that examines the meaning of sexuality at midlife and later life. Aim: This study investigates how heterosexual men and women in mid and later life perceive their sexuality and the factors that influence it. Methods: Group concept mapping was used to produce a conceptual map of the experience of sexuality for heterosexual adults aged 45 and above. Data for the group concept mapping stage were collected using six open-ended survey questions that asked about sexuality, intimacy, and desire. Thematic analysis was used to examine how the participants perceived their sexuality to have changed as they aged. Data for the thematic analysis were collected using a single open-ended survey question. Outcomes: The responses to the six open-ended survey questions were rated by participants using a five-point Likert scale for how important participants felt that each statement was to themselves personally. Participants responses to the seventh open-ended survey question were examined using thematic analysis to understand whether participants felt that their sexual experiences had changed over time, and if so, how they had changed. Results: Eight themes were identified across the different phases of group concept mapping. These were, in order of importance: partner compatibility, intimacy and pleasure, determinants of sexual desire, sexual expression, determinants of sexual expression, barriers to intimacy, sexual urges, and barriers to sexual expression. Seven areas of change were identified in terms of perceived changes to sexuality with age. These included changes to perspective, relationship dynamics, environment, behaviour, body/function, sexual interest/desire, and sexual enjoyment. Clinical Application: Health practitioners and researchers can use the conceptual framework produced in this study to engage with adults in mid and later life on a range of topics relating to their sexual health and sexual experiences. Strengths and Limitations: Group concept mapping reduces researcher and clinician bias by engaging directly with the target population to examine complex social phenomena and provides a detailed visual map of the topic under investigation using both qualitative and quantitative data. However, further validation of these results is necessary because of the small sample size. Conclusion: The results highlight the prioritisation of interrelationship dynamics in mid and later life sexuality over the more frequently examined areas of sexual functioning and sexual urges. These findings may facilitate the development of new perspectives on how sexuality is

experienced in the later years and provide new avenues for intervention in situations where sexual problems arise.

Keywords: sexuality, heterosexuality, ageing, psychology, qualitative methods, quantitative/statistical/survey

4.3 Introduction Sexuality is an integral and constantly changing part of the human experience (Baldissera et al., 2012; Hurd Clarke, 2006; Sandberg, 2013; Wylie et al., 2013). Research indicates that sexual expression is still desired by older adults and that they continue to think of themselves as sexual beings well into later life (Jung & Schill, 2004; Langer, 2009; Lusti-Narasimhan & Beard, 2013). However, as people age, the quality of sexual experiences becomes increasingly more important than their frequency (Deacon et al., 1995; Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; McCarthy, Farr, & McDonald, 2013). This change indicates that the current emphasis of research and medicine on the maintenance or re-establishment of ‘youthful’ sexual functioning and frequency of sexual behaviour in later life may not accurately reflect the sexual priorities of older adults (Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015). The absence of a clear and consistent definition of sexuality in mid and later life, however, makes it difficult to determine the actual sexual wants, needs, and experiences of older adults. Current approaches to understanding sexuality in the later years often imply that declines in sexual activity frequency and sexual functioning are problematic side effects of the ageing process (Beutel et al., 2018; Bortz & Wallace, 1999; Marshall, 2012). In line with this assumption, much of the quantitative later life sexuality research has focused on sexual activity frequency (e.g., Beckman et al., 2008; DeLamater & Moorman, 2007; Karraker, DeLamater, & Schwartz, 2011), sexual dysfunction prevalence (e.g., Bacon et al., 2003; Beckman et al., 2008), and the impact that changes in sexual functioning have on the sex lives of older adults (Corona, Rastrelli, Maseroli, Forti, & Maggi, 2013; Kingsberg, 2002; L. J. Smith, Mulhall, Deveci, Monaghan, & Reid, 2007). This research has been valuable in advancing our understanding of sexual dysfunction and for identifying how sexuality in mid and later life correlates with a broad range of biological, psychological, social, and environmental factors. However, this approach may promote a view of sexuality that focuses on the sexual functioning and sexual activity frequency of older adults to the exclusion of other data, and risks influencing the interpretation of results so that the presence of sexual dysfunction and decreases in penile-vaginal intercourse frequency becomes evidence of an unsatisfactory sex life, despite evidence to the contrary (e.g., Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; Fileborn,

Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; Hurd Clarke, 2006; McCarthy et al., 2013; Ménard et al., 2014; Sandberg, 2013). In contrast, qualitative studies often highlight the importance of intimacy, warmth, and closeness between partners as part of older adults’ sexual experiences (Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; Hurd Clarke, 2006; McCarthy et al., 2013; Ménard et al., 2014; Sandberg, 2013). For example, a study by Ménard et al. (2014) asked men and women over the age of 60 in long term relationships to describe their experiences of ‘great sex’. In their descriptions, participants focused on the role of relationship dynamics, partner availability, openness to new experiences, and overcoming learned ageism in their experiences as opposed to sexual activity frequency, intercourse duration, or sexual functioning capacity in a sexual encounter. These results suggest the current tendency to pathologise any decrease in penile-vaginal intercourse (thus equating optimal sexual functioning with the presence of a satisfying sex life; e.g., Carvalheira, Traeen, & Stulhofer, 2014; Cheung & Rensvold, 2002; Marshall, 2010) may not accurately reflect the way that adults in mid and later life perceive their own sexual experiences. Despite evidence that the quality of sexual experiences is increasingly important to individuals as they age, there is still little research that explores what this means in terms of how sexuality is conceptualised by adults in the late stages of life (Bouman & Kleinplatz, 2014; DeLamater, 2012). This suggests that crucial pieces of the sexuality ‘puzzle’ may be missing in terms of how sexuality in both mid and later life is understood, and how sexual and/or relationship problems are treated. These gaps in our understanding have significant implications for understanding healthy sexuality in later life, on the topics that researchers choose to focus on, and for how researchers and health professionals understand the role of intimacy and non- penetrative sexual expressions in the sex lives of older adults. To better understand how sexuality is conceptualised by heterosexual men and women in mid and later life, it would be helpful to examine the sexual experiences of these adults in a broad, open-ended way. Because the experience of sexuality is highly personal, giving individuals the chance to describe these experiences in their own words allows for a greater depth of understanding of those experiences than a yes/no answer. Qualitative and mixed-method approaches, such as thematic analysis and

group concept mapping, are often used to better understand the experiences or perspectives of individuals (Hammarberg, Kirkman, & de Lacey, 2016). Group concept mapping is a mixed-methods, multi-phase approach to understanding complex social phenomena and generating a conceptual framework of a topic by engaging directly with the population under investigation (Trochim & Kane, 2005). Group concept mapping is an approach that can integrate differing viewpoints from individuals with various attitudes, experiences, and characteristics (Trochim & Kane, 2005). As such, group concept mapping is well placed to examine heterogeneous populations such as heterosexual men and women over the age of 45. Thematic analysis is a qualitative methodology that can be used to examine whether any underlying patterns are present in how a population discusses a topic of interest (Braun & Clarke, 2006). Because group concept mapping produces a conceptual ‘map’ of how ideas interrelate rather than identifying thematic patterns in responses across a sample population, thematic analysis is a more appropriate methodological approach for uncovering similarities within groups. This paper aims to use the voices of heterosexual adults aged 45 and above to obtain an expanded understanding of how sexuality is conceptualised in mid and later life, identify the level of importance placed on the different themes contained within this conceptualisation, and provide further clarity around whether adults in midlife and later life have perceived any changes to their sexual experiences as they have aged. In this study, group concept mapping was used to produce a conceptual ‘map’ that reflected the lived experiences of sexuality for heterosexual adults in mid and later life, and thematic analysis was used to examine how heterosexual men and women in mid and later life perceived changes to their sexuality over time. For this study, it was decided that individuals who do not identify as heterosexual would not be included, and that sample subgroups (i.e., men/women, midlife/later life) would not be examined separately, despite the presence of recognised differences in the sexual experiences for these cohorts. These investigations are currently beyond the scope of this study; however, it is hoped that the results of this study can help to provide a framework for future research in these areas in a way that helps improve the way that comparisons are made between different cohorts. 4.4 Method This study was divided into two parts: a group concept mapping exercise, and a thematic analysis of perceived changes to sexuality with increasing age. The group

concept mapping utilised the thoughts and experiences of heterosexual adults aged 45 and above to develop an overarching conceptual ‘map’ of sexuality in mid and later life that identifies: a) The major themes that describe the sexual experiences of heterosexual adults aged 45 and above, and the determinants that influence these experiences; b) The interrelationships between the themes found above; and c) How these themes are prioritised by heterosexual adults in mid and later life. In the thematic analysis, patterns of perceived changes to sexuality and sexual experiences of heterosexual men and women in mid and later life were investigated. The group mapping exercise will be discussed first, followed by a discussion of the thematic analysis. Because group concept mapping is not a widely used methodology, an overview of the methodological approach is presented before presenting a description of the study methods. 4.4.1 Group concept mapping Group concept mapping (GCM) methodology was used to develop a conceptual model of sexuality in midlife and beyond. Because GCM is an iterative and intensive process of identifying similarities in how a concept is understood by a group of individuals (Braun & Clarke, 2006; Trochim & Kane, 2005), a decision was made to produce an overarching conceptual map of sexuality for both men and women across mid and later life cohorts, with the expectation that further research would be needed to more closely examine differences between these cohorts. The process of group concept mapping comprises three major phases: data collection, data structuring (sorting) and rating, and data analysis (Trochim & Kane, 2005). It is not necessary for the same participants to take part in all three phases, nor is it problematic for the same participants to participate in all phases of the study (Trochim, 1989), as the three phases are structured and analysed independently of each other while remaining interconnected. In the data collection phase, qualitative methods are used to understand experiences in relation to the topic of interest. The information is then consolidated to identify unique statements describing the concept of interest. In the second phase, the statements are presented back to participants who are asked to sort the statements into themes by grouping the statements together in a way that they feel makes sense.

Rosas and Kane (2012) recommend using between 20 to 30 sorters for optimal results in the statistical analysis phase. Once the sorting task has been completed, participants are asked to reflect on the meaning of statements in each grouping and to suggest a short descriptive label that captures the essence of the grouping. Participants are then asked to rate the relative importance of each statement on a Likert-type scale. In the final phase, the data from the sorting and rating tasks are subjected to multidimensional scaling and hierarchical cluster analysis to identify clusters of related statements and to calculate the average importance of different clusters. The results of the quantitative analysis are then presented as a pictorial concept map of major themes (clusters) that describe the concept of interest and show interrelationships between the themes. Details of the group concept mapping process utilised in the current study are presented in the following sub-sections. 4.4.1.1 Phase one: Data collection. 4.4.1.1.1 Participants. Forty participants took part in the data collection phase (43% men, 57% women, mean age = 65, s.d. = 7.97; see Table 4.1). Eligible participants were community-dwelling men and women aged 45 or above who identified as heterosexual, who lived in Australia currently, and who could read and write English competently. The age cut-off was set at 45 years to align with cut-off points for midlife in the guidelines set by the United Nations (United Nations, Statistical Office, 1982). No restrictions were set for current relationship status, ethnicity, or the presence of any chronic health conditions. This was done so that a broad range of perspectives from the population of interest could be sampled, thus increasing the likelihood that the resultant model would be as broadly inclusive of the wider population. 4.4.1.1.2 Materials. Participants were asked to complete a survey that contained seven open-ended focus questions (see Table 4.1). The first six open-ended focus questions of this survey were used in phase one of the study. These six questions asked participants about their views on what sexuality, intimacy and sexual desire meant to them presently, whether they felt anything currently prevented them from being able to express their sexuality and what these obstacles were, what they felt was of particular importance to their own sexual expressions at present, and what they currently perceived to be erotic (see Table 4.1). The questions were developed using the current

Table 4.1 Demographic properties of study sample across phase one and two

Phase 1 Phase 2 Age 45-64 % 65+ % Total 45-64 % 65+ % Total Single, never married 0 0% 0 0% 0 0 0% 0 0% 0 Married without children 6 86% 10 100% 16 0 0% 0 0% 0 Married with children 0 0% 0 0% 0 0 0% 0 0% 0 Divorced 0 0% 0 0% 0 0 0% 0 0% 0 Separated 0 0% 0 0% 0 0 0% 0 0% 0 Widowed 0 0% 0 0% 0 0 0% 0 0% 0 Living with partner 1 14% 0 0% 1 0 0% 0 0% 0 Data unavailable 0 0% 0 0% 0 5 100% 5 100% 10 Living alone 1 14% 3 30% 4 0 0% 0 0% 0 Living with partner 6 86% 6 60% 12 0 0% 0 0% 0 Men Living with children and/or parents 0 0% 1 10% 1 0 0% 0 0% 0 Shared accommodation with non-family 0 0% 0 0% 0 0 0% 0 0% 0 Data unavailable 0 0% 0 0% 0 5 100% 5 100% 10 Australia 6 86% 10 100% 16 0 0% 0 0% 0 Netherlands 0 0% 0 0% 0 0 0% 0 0% 0 New Zealand 1 14% 0 0% 1 0 0% 0 0% 0 Spain 0 0% 0 0% 0 0 0% 0 0% 0 UK 0 0% 0 0% 0 0 0% 0 0% 0 USA 0 0% 0 0% 0 0 0% 0 0% 0 Mauritius 0 0% 0 0% 0 0 0% 0 0% 0 Data unavailable 0 0% 0 0% 0 5 100% 5 100% 10 Women Single, never married 0 0% 1 11% 1 0 0% 0 0% 0

Phase 1 Phase 2 Age 45-64 % 65+ % Total 45-64 % 65+ % Total Married without children 8 57% 6 67% 14 0 0% 0 0% 0 Married with children 2 14% 1 11% 3 0 0% 0 0% 0 Divorced 2 14% 0 0% 2 0 0% 0 0% 0 Separated 1 7% 0 0% 1 0 0% 0 0% 0 Widowed 1 7% 1 11% 2 0 0% 0 0% 0 Living with partner 0 0% 0 0% 0 0 0% 0 0% 0 Data unavailable 0 0% 0 0% 0 7 100% 8 100% 15 Living alone 3 21% 6 67% 9 0 0% 0 0% 0 Living with partner 9 64% 3 33% 12 0 0% 0 0% 0 Living with children and/or parents 2 14% 0 0% 2 0 0% 0 0% 0 Shared accommodation with non-family 0 0% 0 0% 0 0 0% 0 0% 0 Data unavailable 0 0% 0 0% 0 7 100% 8 100% 15 Australia 8 57% 6 67% 14 0 0% 0 0% 0 Netherlands 1 7% 1 11% 2 0 0% 0 0% 0 New Zealand 0 0% 0 0% 0 0 0% 0 0% 0 Spain 1 7% 0 0% 1 0 0% 0 0% 0 UK 2 14% 1 11% 3 0 0% 0 0% 0 USA 2 14% 0 0% 2 0 0% 0 0% 0 Mauritius 0 0% 1 11% 1 0 0% 0 0% 0 Data unavailable 0 0% 0 0% 0 7 100% 8 100% 15

literature on sexuality in mid and later life as well as existing later life sexuality scales (e.g., Geriatric Sexuality Inventory, Kazer et al., 2013; and Senior Adult Sexuality Scales, E. Weinstein, 2013), the input of two psychologists with experience in the areas of sexuality and ageing, and a biostatistician experienced in group concept mapping. The six questions were designed to gather information about the way that individuals thought, felt, and interacted with their sexuality currently, including the experiences of desire and intimacy. No limits were set on the type or length of responses that participants could give in response to the focus questions. A comments section was included to allow participants to provide additional notes about the topic of sexuality that they felt were not captured by the focus questions. 4.4.1.1.3 Procedure. Ethics approval was obtained for all phases of the study from the University Human Research Ethics Committee prior to phase one data collection. Informed consent for phase one was obtained prior to the start of the study. Prior to data collection, participants were provided with an information letter that outlined the nature of the study including information about the type of sensitive information collected, preservation of participant anonymity, data storage procedures, and instructions for how participants could withdraw from or raise concerns about the study. Participants were recruited from across Australia using a promotional strategy that relied predominantly on radio interviews, as well as media coverage in the form of news articles in printed and online media across National news publications. No incentives or tokens of appreciation were offered to participants for this study. Interested individuals were invited to contact the researchers directly, who then assessed the prospective participants against the eligibility criteria. During this initial assessment, participants were also queried about the preference for data collection method (online or via a postal pack). Participants who opted for the online data collection option were referred to the online survey tool Qualtrics version 10.2016 (Qualtrics, 2017). Participants who expressed a preference for a pen-and- paper format were posted a data collection pack. The data collection pack included instructions for how to complete the data collection task, a data collection sheet, and a self-addressed reply-paid envelope. At the end of the survey, participants were provided with the contact details of relevant help organisations in case they found the experience of reflecting on their sexuality distressing.

The data were collected via open-ended survey responses. Responses from all participants across all questions were consolidated in a single Microsoft Excel (Microsoft Office 360 ProPlus 2016, Version 16.0.7726.1042) spreadsheet. Where multiple ideas were present in a response, these ideas were separated into single-idea statements. For example, in response to the question “what kinds of things are most important to you in terms of sexual expression?”, a participant may list multiple things as being important. Each point listed was separated into a single-idea statement for coding. Each single-idea statement was then summarised using a single word by the first author (AM) to capture the core meaning of the statement. Each single statement summary was then used to produce a list of overarching themes. Once a theme had been assigned to each item, all items were sorted using the identified themes and examined collectively by the three authors to identify redundancies within the themes. Redundancies were identified as any statement that shared the same meaning as another statement, as agreed upon by all three authors. After consolidation and removal of redundancies, the remaining unique statements were reviewed by the first author for correct spelling and grammar, and, where necessary, minor changes were made to improve readability prior to using the statements in the next phase. 4.4.1.2 Phase two: Sorting and rating. 4.4.1.2.1 Participants. Twenty-five individuals took part in phase two (40% men, 60% women; see Table 4.1). Because of an error in the collection of demographic data, only sex and age cohort data were available for phase two. Participants from phase one that wished to be contacted about future recruitment for the study were invited to take part in phase two via email, and new participants were recruited using radio and other media outlets. Due to the anonymous nature of the study, no information is available about how many participants from phase one also took part in phase two. All eligibility criteria and recruitment approaches used in phase one were repeated in phase two (see phase one ‘Participants’ section). 4.4.1.2.2 Materials. Participants who chose to participate using a post-pack were provided with a participant information letter, an instruction sheet, a set of statement cards (one card for each unique statement from the brainstorming phase), a set of blank envelopes, a rating task activity sheet, and a self-addressed reply-paid envelope. For online

participants, the same information and tasks were presented in electronic form using the online survey tool Qualtrics version 10.2016 (Qualtrics, 2017). 4.4.1.2.3 Procedure. Informed consent for phase two was obtained prior to the commencement of the study. Participants were asked to complete a sorting task followed by a rating task. In the sorting task, participants were presented with the unique statements identified in phase one and asked to arrange statements into groups of items that they felt were similar in meaning. Participants were directed to create between three and 15 groups of statements and advised that no group should contain only one statement, and no group should be made up of items deemed ‘miscellaneous’. The decision to request between three and 15 groups was made to optimise a nuanced representation of sexuality in midlife and beyond while keeping group numbers manageable for participants. Participants were requested to have more than one item in each group to encourage them to think about the similarities and differences in the ideas captured in statements. The decision to request that no group contained only a single item was made to encourage participants to think about how each statement interconnected with others. No other instructions were provided to participants in terms of how to sort the statements. After participants had sorted all statements into groups, they were directed to name each of the groups. Participants were asked to provide names that best represented the overall theme of statements in each of the groups, but no further direction was given in terms of naming conventions. The participants who completed the sorting task by post were instructed to place each group of statements into a separate envelope, seal the envelope, and write the label describing the group of statements on the envelope, before posting the results back to researchers. After completing the sorting task, participants were asked to rate each statement on how important they felt that statement was to themselves personally on a five-point Likert-type scale (where: 1 = not at all important, 2 = slightly important, 3 = moderately important, 4 = very important, and 5 = extremely important). Again, no token of appreciation or financial incentive was offered to participants. 4.4.1.3 Phase three: Statistical analyses. The sorted data were used to produce individual sort matrices for each participant in a Microsoft Excel (Microsoft Office 360 ProPlus 2016, Version 16.0.7726.1042) spreadsheet. The sort matrices had the number of rows and columns

equal to the number of statements used in the sorting task and had ‘1’ for any pair of items that were sorted in the same group by a given participant and ‘0’ for any pair of items that were sorted into separate groups. The individual sort matrices were then combined to produce a composite dissimilarity matrix for the entire sample. A dissimilarity matrix calculates a numerical value that represents the amount of (dis)similarity between each statement using the chi-square test of equality for two sets of frequencies. In a dissimilarity matrix, smaller numbers indicate that statements appear together in sort groups more often, and larger numbers indicate that statements do not often appear together in sort groups. The composite dissimilarity matrix was then imported into IBM SPSS Statistics for Windows, Version 23 (IBM Corporation, 2013) and used to perform multidimensional scaling (MDS) analysis. This produced a set of Euclidean distances that provided coordinate information about the relative spatial ‘distance’ on a two- dimensional plane between statements based on how often the statements were sorted together by the study participants. Results of MDS were represented as a bivariate point map that showed the frequency with which statements were sorted together by participants. Items that were sorted together more often were distributed more closely together on the bivariate point map and produced smaller squared Euclidean distances. Thus, similar items appeared closer together on the point map, while less similar items appeared further apart on the point map. The MDS distances were used as input into a hierarchical cluster analysis (HCA) conducted using Ward’s algorithm in IBM SPSS Statistics for Windows, Version 23 (IBM Corporation, 2013). HCA is preferred in GCM over other common ‘clustering’ approaches (e.g., principal components analysis or exploratory factor analysis) because it aligns more readily to the results of MDS and is less likely to produce overlapping themes/clusters (Trochim, 1989). Goodness-of-fit of MDS solution was assessed with Kruskal’s stress value formula 1. Stress values represent how well the distances generated in MDS correspond to the values present in the dissimilarity matrix. A fair correspondence between the derived MDS solution and the initial dissimilarity matrix is consistent with stress values equal to .10 or lower, and a good correspondence is consistent with stress values equal to .05 or lower (Kruskal, 1964). The objective was to group items into clusters of conceptually similar statements. By using HCA, statement ‘clusters’ that are most reflective of the

groupings formed by participants are mathematically identified, thus minimising researcher bias in the identification of themes. Cluster solutions with the number of clusters ranging from three to 15 were examined for how well items within a cluster appeared to represent a single theme. The decision to examine only solutions with three to 15 clusters was made to align with directions given to participants during the sorting task to sort statements into between three and 15 groups. Each solution was examined jointly by the authors for whether the clustering of statements produced meaningful themes. Solutions were discarded if one or more of the following conditions were met: (a) one or more clusters within the solution contained only a single item; (b) the solution contained conceptually overlapping clusters; and (c) items within at least one cluster did not form a conceptually cohesive theme. The preferred cluster solution was the solution that optimally identified nuanced groupings of interpretable statements without repetition or redundancies among the identified clusters, as determined by the authors. A final cluster name was chosen using cluster names suggested by participants during the sorting task in conjunction with discussion of identified themes within each cluster by the author team. Once a cluster name was chosen, theme definitions were developed for each cluster that was representative of the type of statements captured within the cluster and other statements that would be considered conceptually similar. Data collected from the rating task were used to calculate mean rating scores for each statement across all participants. The statement ratings were then used to produce mean importance ratings for each cluster. Following this, the final named cluster map, cluster definitions, and cluster importance ratings were presented to participants for review via the project newsletter. Participants were encouraged to contact the research team via email to provide feedback on whether they felt that the results represented their own experiences of sexuality. No participants chose to return written feedback; however, four participants returned feedback via phone. The responses from participants confirmed that the themes and ratings presented were a good representation of their own thoughts and experiences in relation to sexuality in midlife and beyond. For example, a female participant aged 74, commented that the model “covered most areas well”.

4.4.2 Thematic analysis of perceived change to sexual experience with increasing age Thematic analysis (TA) was used to examine whether participants perceived a change in their sexual experiences as they aged. TA allows researchers to use a systematic and iterative approach to identify patterns or repeating themes within the collected data (Attride-Stirling, 2001; Braun & Clarke, 2006). In this study, an inductive approach to TA was used whereby the identification of themes was data- driven rather than theory-driven (Braun & Clarke, 2006). 4.4.2.1 Participants. Refer to participant data in phase one of group concept mapping. 4.4.2.2 Materials. Data were collected using a single question included in the survey during the group concept mapping phase (Table 4.1, question 7). This question was designed to allow participants to describe changes to their sexuality over time in their own words. No instructions were provided in terms of how participants should respond to this question. 4.4.2.3 Procedure. The data were collected via open-ended survey responses. Responses from all participants were consolidated in a single Microsoft Excel (Microsoft Office 360 ProPlus 2016, Version 16.0.7726.1042) spreadsheet. Responses were first examined for whether participants identified a perceived change in their sexual experiences compared to their earlier years. Where changes were detailed by participants, each idea was separated into a single idea statement. These statements were then read and re-read by AM, and a code list was developed using an axial approach that reflected the main themes discussed in the statements. Statements were then re-examined using the code list, and final coding of the data was agreed upon by the author team. Once the key themes relating to any changes to the sexual experiences of participants were identified, responses in their original form were re-examined (AM) to develop a table of themes. Emergent themes were reviewed by the author team in meetings at regular intervals throughout the thematic analysis process, and again at the end of the thematic analysis to ensure agreement on the identified themes and confirm that all pertinent topics were captured.

4.5 Results 4.5.1 Group concept mapping A total of 708 single-idea statements were generated using the responses to the focus questions. After removing repeated ideas, 75 unique statements were identified. These statements represented both attitudes and experiences relating to sexuality, and determinants that influence sexual experiences. Statements encompassed a range of sexual behaviours, sexual attitudes, sexual and relational preferences, and barriers to sexual expression. A full list of unique statements can be found in Table 4.2. The 75 unique statements formed the basis of the sorting and rating tasks in phase two. Sorting data were combined in a dissimilarity matrix and a set of Euclidean distances generated using multidimensional scaling (Kruskal’s stress value = .07). Kruskal’s stress values indicated a fair correspondence between the raw data and distances between points on the bivariate point map. Hierarchical cluster analysis produced a shortlist of cluster solutions that contained seven, eight, or nine clusters for examination. Solutions with more than nine clusters were discarded because they contained conceptually redundant clusters or included at least one cluster with only one item. Solutions with fewer than seven clusters were discarded due to a lack of conceptually cohesive themes in some of the clusters. After examination of shortlisted solutions, the seven-cluster solution was also rejected because it contained one cluster that was not clearly interpretable. Consequently, the preferred cluster solution was identified as the eight-cluster solution because it contained no single item clusters, had no conceptually overlapping clusters, and statements in each cluster formed cohesive, interpretable themes. A cluster map of the eight-cluster solution can be found in Figure 4.1. The eight clusters were named based on identified themes within each cluster and were: partner compatibility, intimacy and pleasure, determinants of sexual desire, sexual expression, determinants of sexual expression, barriers to intimacy, sexual urges, and barriers to sexual expression. 4.5.1.1 Cluster definitions. 4.5.1.1.1 Sexual expression. Seven statements formed the sexual expression cluster and reflected the different ways individuals may express their sexuality. This cluster encompassed statements that referred more specifically to physical acts or partnered activities as a way of engaging with and expressing sexuality. Statements captured in this cluster

Table 4.2 Distribution of clustered statements and mean importance scores across identified clusters

Importance Rating

Cluster Cluster ID Statement Rating* Range M SD M (SD) A sense of freedom of expression and appreciation of S2 V 2-5 4.00 0.83 our differences S6 Being able to communicate openly V 3-5 4.62 0.56 S9 Compatibility with each other V 1-5 4.07 1.14 S17 Feeling cared for and valued V 2-5 4.37 0.83 Partner 4.37 compatibility S18 Feeling safe V 2-5 4.33 0.87 (0.24) S21 Freedom to be myself V 2-5 4.22 0.93 S22 Friendship/positive regard for you and your partner V 3-5 4.48 0.70 S54 Respect is important V 3-5 4.70 0.54 S60 Someone whose mind and personality are attractive V 3-5 4.55 0.57 S4 An emotional response to a physical need of touch V 3-5 4.11 0.64 Anticipation for when we share those special S5 V 3-5 4.07 0.72 moments together S7 Being able to hold and be held V 3-5 4.59 0.57 S11 Enjoying each other’s bodies V 3-5 4.44 0.64 S13 Expression of desire or interest M 3-5 3.92 0.82 S24 General petting and intimacy V 2-5 4.11 0.80 S25 Giving and receiving pleasure V 3-5 4.37 0.68 S34 Knowing that my partner finds me attractive V 1-5 4.07 0.95 S43 Mutual attraction and satisfaction V 3-5 4.51 0.64 S44 Mutual trust and understanding V 3-5 4.81 0.55 S50 Partner's smile V 1-5 4.14 1.13 Physical contact of some sort, hugging, holding S51 V 2-5 4.55 0.80 hands, kissing Intimacy and Playful activities leading up to being intimate (e.g., 4.22 S53 V 3-5 4.25 0.81 pleasure pats and hugs throughout the day) (0.31) S55 Romance V 3-5 4.14 0.71 S56 Sensuality V 3-5 4.18 0.68 Sharing the things that are the most personal part of S59 V 1-5 4.07 1.07 you S61 Something that brings emotional satisfaction V 2-5 4.33 0.73 Speaking softly to each other is an important part of S63 M 1-5 3.44 1.21 intimacy S64 The ability to enjoy sexual activity V 3-5 4.44 0.64 To be appreciated as a whole person is important to S68 V 2-5 4.18 1.00 my sexual expression To be intimate with a particular person, with or S69 V 3-5 4.29 0.66 without actual intercourse S70 To explore my sexuality with a particular person M 1-5 3.59 1.30 Touching, kissing and being attentive to the other S71 V 2-5 4.40 0.79 person

Importance Rating

Cluster Cluster ID Statement Rating* Range M SD M (SD) A feeling I have when I get a certain look, touch, or S1 M 1-5 3.74 1.02 conversation from someone I find attractive Finding another person attractive and physically S19 M 1-5 3.81 0.96 appealing Determinants 3.87 of sexual S29 How comfortable in my body or how attractive I feel V 2-5 4.03 0.80 (0.26) desire How I relate to someone in terms of making eye S30 contact, touching them, how I speak to them and the V 3-5 4.22 0.69 words I use S65 The atmosphere and surrounds M 2-5 3.55 0.84 S27 Having intercourse M 1-5 3.62 1.07 S38 Massage M 1-5 3.07 1.07 S45 Nakedness M 1-5 3.66 1.03 Sexual 3.42 S47 Oral sex M 1-5 3.29 1.40 expression (0.35) S57 Sensual body movements M 1-5 3.85 0.98 S73 Visual stimulation M 1-5 3.51 1.08 S75 Watching my partner in the shower S 1-5 2.88 1.25 It is important that my partner is clean when S8 M 1-5 3.18 1.44 expecting me to respond S23 Gender is important in terms of sexual expression M 1-5 3.29 1.65 Going along because of the other person's needs at S26 M 1-5 3.07 1.10 the time S28 Health and wellbeing V 2-5 4.33 0.91

Determinants S31 I tend to "fly solo" S 1-5 2.11 1.52 3.22 of sexual Moral and privacy standards (e.g., there is a time and (0.64) expression S42 a place to express your sexuality that is acceptable in S 1-5 2.70 1.32 our society) Sexual orientation and how you express or act upon S58 M 1-5 3.14 1.56 your sexual feelings Through behavior and the way in which I present S66 myself particularly in the company of the opposite M 1-5 3.18 1.11 sex S67 Time and opportunity for sex M 2-5 3.96 0.93 S10 Energy levels M 1-5 3.37 0.92

Barriers to S15 Fear of being hurt S 1-5 2.51 1.45 2.84 intimacy S16 Fear of forming a new relationship S 1-4 2.11 1.12 (0.63) S41 Mood or mental state M 1-5 3.37 1.44 Enjoying erotic print and film media alone or with a S12 S 1-5 2.44 1.31 partner S14 Fantasising and thoughts M 1-5 3.18 1.30 S20 Flirting S 1-5 2.92 1.14 2.72 Sexual urges S33 It is an underlying driving force that never goes away S 1-5 2.77 1.55 (0.31)

S37 Manual orgasm S 1-5 2.77 1.39 S72 Urge for physical release of body fluids S 1-5 2.62 1.47 S74 Voyeurism S 1-5 2.25 1.31

Importance Rating

Cluster Cluster ID Statement Rating* Range M SD M (SD) Alcohol or drug use prevents me from expressing my S3 N 1-5 1.55 1.31 sexuality S32 Illness or pain interferes with sexual expression S 1-5 2.51 1.67 S35 Lack of opportunity S 1-5 2.33 1.35 S36 Lack of suitable company N 1-5 1.70 1.23

Barriers to S39 Medications N 1-4 1.88 1.18 2.36 sexual S40 Mismatched libido with my partner N 1-5 1.92 1.38 (0.55) expression S46 Not feeling the urge S 1-5 2.51 1.28 S48 Partner's illness and medical treatment S 1-5 2.85 1.72 S49 Partner's lack of interest or tiredness M 1-5 3.29 1.51 S52 Physical disabilities S 1-5 2.44 1.57 S62 Sometimes I do not want to express my sexuality S 1-5 2.92 1.38 * N = Not at all important; S = Slightly important; M = Moderately important; V = Very important

included “[other than sexual intercourse, I also express my sexuality through] visual stimulation” and “[to me, sexuality means] having intercourse”. 4.5.1.1.2 Sexual urges. Seven statements formed the sexual urges cluster and reflected the urges that underlie sexual expressions. This cluster included statements relating to sexual desires, physical sexual urges, and other forms of sexual expression. Statements captured in this cluster included “[other than sexual intercourse, I also express my sexuality through] fantasising” and “[sexuality] is an underlying driving force that never goes away”. 4.5.1.1.3 Intimacy and pleasure. The intimacy and pleasure cluster encompassed 23 statements and was the largest cluster of statements. Eighteen of these statements reflected the sexual experiences related to how an individual is intimate with others, and experiences of pleasure. Five statements within the intimacy and pleasure cluster reflected determinants that influence how an individual is intimate with others, and their experiences of pleasure. Alternative solutions failed to separate this cluster into smaller clusters while retaining meaning in other clusters. The experiences captured in the intimacy and pleasure cluster encompassed both the ability and the need for affectionate, intimate, and some sexual behaviours, while statements that reflected determinants that influence sexual expression emphasised the quality of the intimate

Figure 4. 1 Depiction of cluster proximity using a bivariate point map.

relationship and partner engagements, mutual trust, and mutual attraction. Statements captured in this cluster included “to be appreciated as a whole person is important to my sexual expression”, “[to me, sexuality and intimacy] is giving and receiving pleasure”, and “intimacy is sharing the things that are the most personal part of you”. 4.5.1.1.4 Partner compatibility. Nine statements were grouped together to form the partner compatibility cluster. Statements in this cluster encompassed the determinants associated with feeling comfortable with and cared for by a romantic partner. It also included statements about being able to communicate with a partner, and the acceptance and mutual respect between partners. Statements captured in this cluster included “feeling cared for and valued [is important to me in terms of sexual expression]” and “a sense of freedom of expression and appreciation of our differences [is important to me in terms of sexual expression]”. 4.5.1.1.5 Determinants of sexual desire. Five statements were grouped together to form the determinants of sexual desire cluster and referred to factors associated with the experience or onset of sexual desire. Statements captured in this cluster included “[I can express my sexuality through] how I relate to someone in terms of making eye contact, touching them, how I speak to them and the words I use” and “finding another person attractive and physically appealing [is important to me in terms of how I express my sexuality]”. 4.5.1.1.6 Determinants of sexual expression. Nine statements were included in the determinants of sexual expression cluster. These referred to the factors that influenced the way people thought about and expressed themselves sexually, their sexual standards, and the ways they engaged with others sexually. Statements captured in this cluster included “it is important that my partner is clean when expecting me to respond” and “moral and privacy standards [can prevent me from expressing my sexuality]”. 4.5.1.1.7 Barriers to intimacy. Four statements were grouped together to form the barriers to intimacy cluster. This cluster related to determinants that impinged upon an individual’s willingness to engage with another person on an intimate and/or sexual level. Statements captured in this cluster included “a fear of being hurt [prevents me from being able to express my sexuality]” and “energy levels [can prevent me from being able to express my sexuality]”.

4.5.1.1.8 Barriers to sexual expression. Eleven statements were grouped together to form the barriers to sexual expression cluster. These related primarily to physical, medical, or social factors that act as barriers to sexual expression. Statements captured in this cluster included “illness or pain interferes with sexual expression” and “sometimes I do not want to express my sexuality”. 4.5.1.2 Cluster proximity. Clusters were arranged on the bivariate point map as determined by MDS and HCA (see Figure 4.1). Clusters that were most proximal to each other included the partner compatibility and intimacy and pleasure clusters, the sexual expression and sexual urges clusters, and the barriers to intimacy and barriers to sexual expression clusters. Proximal clusters could be seen to share similar ideas. For example, the item “freedom to be myself” from the partner compatibility cluster is similar to, yet still distinct from, the item “to be appreciated as a whole person is important to my sexual expression” in the intimacy and pleasure cluster. Clusters that were located further apart on the cluster map were less likely to share similar themes. For example, the barriers to sexual expression cluster had little similarities with the partner compatibility cluster, and these two clusters were found on opposite sides of the cluster map. 4.5.1.3 Cluster importance. An overall importance rating for each cluster was produced by using the mean importance rating scores of all statements included within the cluster (see Table 4.2). The partner compatibility and intimacy and pleasure clusters produced average importance ratings above 4 (very important); the determinants of sexual desire, sexual expressions, and determinants of sexual expression clusters produced average importance ratings above 3 (moderately important); and the barriers to intimacy, sexual urges, and barriers to sexual expression clusters produced average importance ratings above 2 (slightly important) overall by participants. No clusters produced average importance ratings equal to 5 (extremely important) or below 2 (not important at all). 4.5.2 Thematic analysis of perceived change to sexual experience with increasing age Twenty-two participants stated that they had experienced changes to their sexuality over time, and seven participants stated that they had not experienced any

changes to their sexuality over time. Most participants that reported no change to their sexuality were women (71% female), and all participants that reported no change to their sexuality were in midlife. Eleven participants chose to not provide a response to this question. The changes discussed by participants encompassed seven themes: changes to perspective, changes to relationship dynamics, environmental changes, behaviour changes, physical changes, changes to sexual interest/desire, and increased sexual enjoyment. A summary of themes identified by age cohort and gender can be found in Table 4.3.

Table 4.3 Sum totals of described changes to the experience of sexuality with age across gender and age cohorts

Theme Midlife (M/F) Later life (M/F) Total (M/F) No change 7 0 7 (2/5) (0/0) (2/5) Changes to perspective 7 15 22 (3/4) (8/7) (11/11) Changes to relationship dynamics 7 7 14 (2/5) (4/3) (6/8) Environmental changes 3 3 6 (0/3) (1/2) (1/5) Behavioral changes 2 11 13 (0/2) (5/6) (5/8) Physical changes 4 12 16 (1/3) (8/4) (9/7) Change to interest/desire 7 9 16 (2/5) (4/5) (6/10) Increased enjoyment 2 7 9 (0/2) (2/5) (2/7)

4.5.2.1 Changes to perspective. Changes to perspective were discussed by 22 participants. Changes to perspective were discussed equally between men and women but were discussed more frequently by individuals in later life (68% later life). Positive changes to perspective included a better sexual awareness, learning to accept each other’s limitations, becoming more open-minded, overcoming old beliefs about sex being dirty and/or wrong, having less expectations for intercourse, expanding definitions of sexuality to include companionship and love, feeling a greater sexual freedom because there is no

risk of unwanted pregnancy, becoming more relaxed in discussing sexual topics, and a greater focus on the quality of sexual experiences over quantity. Negative changes to perspective included feeling less attractive with age and feeling less confident about being able to express sexuality. 4.5.2.2 Changes to relationship dynamics. Fourteen participants discussed changes that related to their relationships. Changes to relationship dynamics were discussed equally between men and women and between those in mid and later life. Positive changes included feeling like a more compassionate lover, changes in sexual activity and sexual activity frequency related to starting a new relationship, and an improvement in the quality or depth of connection with a partner. Negative changes focused on a mismatch in libido between partners. 4.5.2.3 Environmental changes. Six participants attributed changes to their sexuality to changes in their environment. Environmental changes were discussed more frequently by women (83% women) but were discussed equally between those in mid and later life. Specific environmental changes included children moving out of the family home, partners having more time together after retirement, and increased privacy and/or an increase in the time available to think about or engage in sexual experiences. Almost all environmental changes were described positively. Only one participant stated that environmental factors negatively impacted on their sex life but did not expand on what these environmental factors were. 4.5.2.4 Behavioural changes. Thirteen participants discussed behavioural changes that largely reflected developing alternative forms of intimacy to gain sexual fulfilment without intercourse. Women were slightly more likely to discuss behavioural changes than men (62% women), and those in later life were more likely to discuss behavioural changes than those in midlife (85% later life). This included a focus on love and appreciation of each other, and expanding the types of sexual behaviours engaged in, in response to physical changes over time, such as a reduction in erection rigidity. One participant discussed seeking out sex regularly after the breakdown of a long- term relationship, resulting in “more sex than ever”.

4.5.2.5 Physical changes. Sixteen participants discussed physical changes that had influenced their changed sexual experiences. Roughly equal amounts of men and women discussed physical changes (44% women). Individuals in later life were more likely to discuss physical changes than those in midlife (75% later life). Most of the physical changes discussed focused on a decrease in sexual functioning, issues with sexual expression because of medical or health barriers, and a change in the physical experience of sex. For example, several men reported a reduction in erection rigidity. Several participants also commented on physical impediments to certain sexual positions. 4.5.2.6 Change to sexual interest/desire. Sixteen participants reported a change to their sexual interest or desire. Changes to sexual interest/desire were discussed slightly more often by women compared with men (63% women) but were discussed roughly equally between those in mid and later life (44% midlife, 56% later life). For most participants, this change was a decrease in the intensity or frequency of desire. For many of these participants, a decrease in their desire did not mean the absence of desire altogether, although one participant reported a complete loss of desire. 4.5.2.7 Increased sexual enjoyment. Nine participants felt that they experienced a greater amount of sexual enjoyment now than they had in the past. Most of those reporting an increase in sexual enjoyment were women (78% women) and those in later life (78% later life). These changes included a greater enjoyment of sexuality generally, a greater enjoyment of sexuality related to better emotional aspects of a relationship and enjoying their body more in sexual engagements. 4.6 Discussion The use of a mixed-methods approach in this study allowed for the sexual experiences of heterosexual men and women aged 45 and above to be examined in a broad, open-ended way. Group concept mapping helped elucidate how these men and women understood and prioritised their own sexuality and sexual experiences, as well as how they categorised these experiences in a way that was meaningful to them. The thematic analysis gave added clarity about whether heterosexual men and women in this life stage perceived a change to their sexuality as they aged and identified the different types of change they had experienced. The discussion will first explore the themes captured within the concept map, their relative importance, and the way they

combine to form a new picture of sexuality in the later years of life before discussing how sexuality is perceived to change over time. This will be followed by a discussion about how this new representation of sexuality in mid and later life can offer further insight into how sexuality for heterosexual men and women in these life stages is understood, and how it can help clinicians identify new avenues for improving the treatment of sexual and relationship problems. The eight-cluster conceptual model generated in the study captured a broad range of experiences associated with sexuality and the different factors that influence them and highlighted the interpersonal elements of sexual experiences, their underlying drivers, and the different barriers to sexual expression experienced by heterosexual men and women aged 45 and above. However, the way that ideas were grouped by participants indicated that much of what is known about sexuality in mid and later life is not organised in a way that accurately reflects the way that older adults conceptualise their own sexuality and sexual experiences. For example, the conceptual map highlighted the importance of interpersonal relationships in sexuality in mid and later life (e.g., partner compatibility and intimacy). This emphasis is consistent with research that shows that intimacy and relationships become more important to adults as they age (e.g., Hinchliff & Gott, 2004; Hurd Clarke, 2006; Kleinplatz et al., 2009), but challenges the idea that intimacy is a separate but related construct (e.g., Kalra, Subramanyam, & Pinto, 2011). In the proposed conceptual framework, intimacy and trust played a central role to the experience of sexuality and to sexual expression in mid and later life. However, this greater relative importance placed on the partner compatibility and intimacy clusters over the barriers to sexual expression and sexual urges clusters challenges the assumption that penile-vaginal penetrative intercourse is the centre-point to sexuality throughout the lifespan, and that sexual problems stem primarily from physiological dysfunctions (Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; Sanders et al., 2013). While still considered somewhat important by the participants in the present study, sexual activity and intercourse were rated less important to their sexual experiences than partner compatibility or intimacy and pleasure. This questions the current research focus on sexual frequency and sexual functioning, while supporting research that shows that the quality of sexual experiences becomes more important for older adults over time than the frequency with which they occur (Deacon et al., 1995;

Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; McCarthy et al., 2013). This key difference also shows that an individual’s sexual experiences are subject to change as partnership dynamics change over time, and the proposed concept map provides suggestions as to what factors may be associated with these changes. However, further research examining the interpersonal aspects of sexuality and their associated determinants is hindered by a lack of validated measurement tools that assess sexuality for older adults (Macleod & McCabe, 2019). The importance of interpersonal relationships in mid and later life sexuality is further highlighted by the distinction made between barriers to intimacy and barriers to sexual expression in the concept map. Currently, research that examines the decline in sexual activity frequency with age has focused predominantly on the presence of sexual dysfunction or the absence of a partner (e.g., Huang et al., 2009; A. Karraker et al., 2011) rather than interpersonal factors such as a fear of rejection, or a lack of opportunities to engage in sexual expression. The splitting of barriers to sexuality into two sub-categories indicates a clear separation of ideas by participants, and further suggests that interpersonal problems can be just as influential in the decline of sexual behaviours as the presence of sexual dysfunction or the absence of a partner. However, it is worth noting that other exploratory methodologies that do not involve the direct involvement of participants in identifying similarities between statements (e.g., exploratory factor analysis) may not have identified this thematic split in barriers to sexuality and intimacy. Unsurprisingly, negative changes related to ageing identified by the thematic analysis included a reduction or loss of desire or sexual interest, health impediments to sexual expression, and decreased sexual functioning. However, many individuals also discussed positive changes to their sexuality over time. These included feeling like a more considerate lover, and increased enjoyment of their sexuality. For many adults in mid and later life, there is a message that sex and sexuality is only for the ‘young’ (Bouman & Kleinplatz, 2014; Langer, 2009) and that intercourse is the primary ‘goal’ of sexual expression (Ménard et al., 2014). However, in a study by Ménard et al. (2014), participants discussed an increase in satisfaction and enjoyment in their sexual encounters once they were able to overcome learned attitudes about sexuality, gender roles, and sexual behaviours. This suggests that learning to disregard socially prevalent messages about sex being the domain of the young may be an important part of being able to maintain a positive sex life well into old age.

For all but seven individuals in the midlife cohort, sexual experiences in mid and later life were perceived to be different to those from the younger years. This suggests that individuals are likely to perceive a change to their experience of sexuality at some point prior to entering the later life stage (i.e., 65+). Existing research fails to agree on what age this change occurs, with studies finding an increased focus on intimacy in populations in individuals from their late 20s into their 80s (Carmichael, Reis, & Duberstein, 2015; Hammarberg et al., 2016; Hurd Clarke, 2006). This variability in results may suggest that this change is related more to life events that predominantly occur prior to entering later life than to chronological age. Interestingly, more women discussed an increase in sexual enjoyment as they aged than did men. For these women, increased sexual enjoyment was attributed to changing environmental factors, including a reduction in parental responsibilities as children become adults. This is in line with research that suggests that many women attribute positive changes to their sex lives in midlife to environmental changes (Mansfield, Koch, & Voda, 2000) and that women report that being able to juggle work and family becomes easier as children get older (Craig & Sawrikar, 2009). The improvement of sexual enjoyment by women but not men specifically in midlife may also be explained by research that suggests midlife is a greater time of environmental flux for women than men (Burns & Leonard, 2005; K. R. Smith & Moen, 1988). The results from this study are particularly relevant to health professionals and researchers who are working in the area of sexuality with heterosexual men and women aged 45 years and above. In particular, the conceptual framework developed in this study may be an effective tool for examining sexual and relational problems in research and health care settings. The effective treatment of sexual and relational problems relies on the health practitioner having as much contextual information about the issue as possible (Althof, 2010). By structuring information-gathering activities to align with the presented conceptual framework, health practitioners can be confident that they have a broad understanding of the context in which the sexual problem exists. To demonstrate how the results of this study could help with the collection of contextual information about a sexual or relational problem by health practitioners, the conceptual framework will be discussed within the context of narrative therapy. Narrative therapy is sometimes used in counselling to help identify and counteract repeating ‘scripts’ (Muruthi et al., 2018). In such a setting, the conceptual framework

can be used to help a clinician gently prompt a client to discuss topics that may not outwardly appear to be related to the problem being treated but may provide important contextual information in developing a treatment plan. By guiding conversations using the proposed conceptual framework, clinicians can work with clients to identify problematic scripts that are impacting their sexuality that may not have been identified in sessions otherwise. However, clinicians should not assume that using the proposed conceptual model as a roadmap for information-gathering will eliminate the need to collect contextual information using other approaches as well. The proposed conceptual framework provides researchers and clinicians with a structured yet flexible approach to understanding and engaging with sexuality in mid and later life. The framework is able to incorporate individual differences in terms of how an individual prioritises the different sexuality aspects and related determinants while still providing a framework for organising existing knowledge. This framework may also assist in the development of an agreed-upon definition of sexuality in mid and later life that is currently lacking in the field (Macleod & McCabe, 2019). Currently, there is limited data available to produce a reliable picture of what constitutes ‘normal’ sexuality in mid and later life, let alone effectively capture the spectrum of potential sexual experiences and preferences of older adults (Bouman & Kleinplatz, 2014). This is partly due to the limited number of validated measurement tools available to assess sexuality in mid and later life (Macleod & McCabe, 2019). By reframing our understanding of sexuality in mid and later life within the proposed framework, gaps in current approaches to assessing sexuality in mid and later life are highlighted, and opportunities for further research identified. 4.6.1 Limitations This study was conducted solely within Australia, with men and women who identified exclusively as heterosexual. As such, further investigation is needed to determine the generalisability of results to other populations with different cultural values and social mores, those who do not identify as heterosexual, and for those that do not wish to identify their sex using a binary descriptive system (i.e., man/woman). The need for additional investigation as to the generalisability of these results is further warranted based on the low response rate at the conclusion of phase three when participants were asked to comment on the conceptual map produced in the study. In addition, because the study required that participants read and respond to a written survey, participant inclusion was limited to those who could read and write

English confidently. As a result, the results may not reflect the views of individuals within the community who are less proficient in English. This study was designed to be exploratory in nature, rather than descriptive. Therefore, examining differences between sub-groups of the sampled population was considered out of scope for this paper. This may mean that generational and/or sex differences, if they exist, have been masked by examining these cohorts collectively, and more research is needed that explores potential differences between these cohorts. Exploration of sex and/or gender differences was also prevented by the small sample size, although the sample size is within or above recommended guidelines for group concept mapping methodologies (Rosas & Kane, 2012; Trochim, 1989). Phrasing of demographic questions also limited the exploration of gender differences in this study, as respondents were asked to identify their sex as male or female, as opposed to identifying the gender they choose to identify with. It is also important to note that because this research focuses exclusively on the attitudes and perceived experiences of heterosexual adults aged 45 and over, no data is available that compares these results with younger individuals or with LGBTQIA+ populations. As such, no comments can be made on the relevance of the proposed model to younger populations or to LGBTQIA+ populations of any age. 4.7 Conclusions and future research The present study produced a conceptual ‘map’ of sexuality for heterosexual men and women in mid and later life that reflects the sexual experiences of these individuals beyond simplistic assessments of sexual function and intercourse frequency. The eight themes identified in the group concept mapping, paired with their calculated importance ratings, provide new information about the domains that heterosexual men and women in midlife and beyond consider to be important to their sexual experiences and how these are prioritised. Additionally, the conceptual ‘map’ produced in this study provides a theoretical framework for sexuality in mid and later life that helps distinguish between the internally driven parts of an individual’s sexual experience from the parts that are, or can be, influenced by external factors. Such an approach allows for new perspectives on how sexuality in the later years is studied and may provide new avenues of intervention in situations where sexual problems arise. Future research needs to build upon this representation of sexuality in mid and later life to examine whether there are differences between men and women, or

between mid and later life groups in terms of the areas they consider important to their sexual experiences. Additional research examining the influence of relationship length, relationship status, and cultural background on the sexuality and sexual experiences of adults in mid and later life would also be beneficial. The concept map could also be used to guide the development of a comprehensive measurement tool that captures a broader range of aspects and factors related to sexuality in midlife and beyond. It would also be beneficial to conduct a similar study to examine the relevance of the proposed model to older LGBTQIA+ populations, as this area is largely understudied. Such a study would help identify areas of overlap for heterosexual and LGBTQIA+ populations, while still providing enough flexibility in the research design to identify themes that may be unique to older LGBTQIA+ populations. It would also be interesting to conduct the same analyses with a younger cohort to see if the current representations of sexuality in younger adulthood are valid.

Chapter 5. Study 2: Development and psychometric evaluation of the Over 45s Adult Sexuality and Intimacy Scale (OASIS)

100

5.1 Chapter guide Macleod, A. & McCabe, M. P. (in preparation [1]). Development and psychometric evaluation of the Over 45s Adult Sexuality and Intimacy Scale (OASIS). To be submitted to Psychological Methods.

The current chapter describes the development of the Over 45s Adult Sexuality and Intimacy Scale (OASIS) and provides a psychometric assessment of the final scale. This process builds on work from chapters two through four to facilitate the development of a new sexuality measurement scale according to best practice guidelines (Boateng, Neilands, Frongillo, Melgar-Quiñonez, & Young, 2018). The chapter begins with the study abstract (Section 5.2) and introduction (Section 5.3), before separately describing the methods, results and a brief discussion for each of the three studies (Sections 5.4-5.6). In the first study, cognitive interviewing is used to refine item phrasing (Section 5.4). The second study uses Rasch modelling to identify the strongest items for inclusion in the final scale and to confirm the appropriateness of the assigned Likert-type rating scale (Section 5.5). In the third study, confirmatory factor analysis (CFA) is used to examine the test dimensionality of the OASIS, before the internal consistency, test-retest reliability, and concurrent validity of the OASIS is examined (Section 5.6). This section is followed by a general discussion of the three studies, a summary of limitations (Section 5.7), and lastly, conclusions (Section 5.8). At the close of the current chapter, a refined, psychometrically robust version of the Over 45s Adult Sexuality and Intimacy Scale is presented. Ethics approval and related documentation for this study can be located in Appendix C.

101

5.2 Abstract Background: Existing later life sexuality measurement scales reflect only a narrow spectrum of experiences relevant to sexuality in mid and later life. New, comprehensive scales are needed that have been developed for, and with, older adults. This paper outlines the development and psychometric evaluation of the Over 45s Adult Sexuality and Intimacy Scale (OASIS) across three studies. Methods: Conceptual clarity of draft items of the OASIS was assessed through cognitive interviews in the first of three studies. Psychometric properties of the items were assessed with Rasch modelling in the second study. Confirmatory factor analysis (CFA) was used to evaluate the hypothesised nine-factor structure of the OASIS. Cronbach's alpha and Intra Class Coefficients (ICC) were used to evaluate the test- retest reliability, internal consistency and scale reliability in the third study, before the converged construct validity of the OASIS was evaluated using scales measuring similar constructs to the OASIS subscales. Results: The OASIS is a 78 item sexuality measurement scale for heterosexual adults aged 45 years and over distributed across nine subscales that assess the experience of sexuality (aspects of intimacy and pleasure, sexual expressions and sexual urges), and factors related to the experience of sexuality (partner compatibility, factors associated with intimacy and pleasure, determinants of sexual desire and sexual expression, and barriers to intimacy and sexual expression). CFA supported a nine-subscale structure (RMSEA = .07, SRMR = .09). Psychometric evaluation established good internal consistency (α’s between .74 and .93) and test-retest reliability (ICCs between .85 and .94), convergent validity for all subscales with existing sexuality measures, and discriminant validity for five of the nine subscales. Conclusions: The present study confirmed that the OASIS is a psychometrically sound measures of sexuality and the factors associated with sexuality experiences for heterosexual men and women in mid and later life.

Keywords: scale development, factor analysis, sexuality, ageing, midlife, measurement, psychometric properties, Rasch modelling

102

5.3 Introduction Sexuality is broadly recognised as an essential part of life (DeLamater & Moorman, 2007; World Health Organisation, 2006) but is understudied across the later stages of life. Currently, the small number of later life sexuality measurement scales that exist examine only a narrow range of topics related to sexuality (e.g., Kazer et al., 2013; E. Weinstein, 2013). As a result, a large portion of knowledge relating to sexuality across the later life stages has been developed using scales designed for younger populations or with survey items that have no published evidence of validity or reliability (e.g., Matthias et al., 1997; Minichiello et al., 2004; Momtaz et al., 2014; Penhollow et al., 2009). To understand the sexuality of older men and women more clearly in mid and later life (i.e., those aged 45 years and above, as per UN guidelines; United Nations, Statistical Office, 1982), a measurement scale that is capable of assessing the sexual experiences relevant to adults in mid and later life and the factors associated with these sexual experiences is needed. Current approaches to understanding sexuality in mid and later life primarily focus on the sexual functioning (or dysfunction) of older adults (Killinger et al., 2014). This research often portrays changes in sexual functioning with age as a symptom of ageing that needs to be medically managed (Bradway & Beard, 2015; Malatesta, 2007). However, this focus on sexual functioning fails to acknowledge that not all older adults experience sexual dysfunction (Laumann & Waite, 2008; Nicolosi et al., 2004) and that not all individuals who experience sexual dysfunction consider this an impediment to a fulfilling sex life (Ménard et al., 2014). Further, using an approach to understanding sexuality throughout the life-course that emphasises sexual functioning and sexual activity ignores evidence that older adults often prioritise aspects of their sexuality (including sexual functioning) differently in later life than in their younger years (Deacon et al., 1995; Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; McCarthy et al., 2013; Sandberg, 2013). As such, the over-reliance on sexual functioning as the centre-point of sexuality throughout the adult life-course may not reflect the full range of sexual experiences or preferences of older adults (Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015). As with those in early adulthood, the sexual experiences of men and women in mid and later life are composed of and shaped by a broad range of biological, psychological and social factors (Lehmiller, 2017). These factors include age, gender, health status, relationship status, quality of life, partner health and attitudes, and

103

partner availability (Ginsberg et al., 2005; Gott, 2004; Guan, 2004; Hyde et al., 2010; Jung & Schill, 2004; Lindau et al., 2007; Lusti-Narasimhan & Beard, 2013; Malakouti et al., 2012; Marshall, 2010, 2012; Ménard et al., 2014; Müller et al., 2014; Palacios- Cena et al., 2012; Sandberg, 2013; Trudel et al., 2014). However, the dominant focus of existing later life sexuality measures on assessments of sexual activities, attitudes relating to sexual permissiveness, sexual satisfaction, and sexual interest for older adults (e.g., Kazer et al., 2013; E. Weinstein, 2013) leaves significant gaps in the ability of research to capture the experience of sexuality for individuals in mid and later life and the factors associated with these experiences comprehensively. There is clear evidence that an approach to understanding sexuality that emphasises the sensual and intimacy-related aspects of sexuality would be more relevant to older populations than a functional, orgasm-oriented approach that focuses on sexual activity (Beckman et al., 2014; Deacon et al., 1995; Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; McCarthy et al., 2013). Having a clear understanding of the experience of sexuality for adults across mid and later life would help to improve the identification of individuals who are negatively impacted by changes to their sexual functioning as they age, as opposed to those who have adapted well to changes in their sexual functioning over time. As such, the development of a new sexuality measurement scale that incorporates the perspectives of older men and women is needed to fill gaps in the current understanding of sexuality for adults in mid and later life. Such a scale would help generate a more precise representation of how older adults experience sexuality and the factors associated with these experiences than is currently available. This paper describes three studies that outline the development and psychometric assessment of a new sexuality measurement scale for heterosexual men and women in mid and later life according to best practice guidelines. According to Boateng et al. (2018), best practice scale development occurs across three stages: item development (Macleod et al., in press; Macleod & McCabe, accepted, 2019), scale development, and scale evaluation. Three previous publications align with the first stage (item development), thereby providing a strong foundation for the second and third stages of scale development reflected in this manuscript. These prior publications highlighted the need for a new approach to how sexuality in the later stages of life is understood and measured (Macleod & McCabe, accepted, 2019), before using group concept mapping methodologies to provide a robust

104

assessment of sexuality for men and women in mid and later life life (Macleod et al., in press). The group concept mapping process concluded with the identification of 88 representative statements about the experience of sexuality in mid and later life and the factors associated with these experiences for heterosexual men and women aged 45 years and above. These representative statements were grouped into themes that provided a statistical representation of how these cohorts perceived the representative statements to be related to each other, thereby generating a conceptual ‘map’ of the sub-constructs within the broader experience of sexuality for adults in the later stages of life. The current manuscript builds on the work conducted across these previous publications to complete the second and third stages of scale development according to best practice guidelines: Scale development and scale evaluation. In the first study, the representative statements generated in the group concept mapping study (Macleod et al., in press) form the pool of draft items for the proposed scale. These items are then refined using cognitive interviewing (Study 1), before Rasch modelling is used to identify the most psychometrically sound items for inclusion in the final scale (Study 2). The psychometric properties of the finalised scale are then assessed to determine the reliability and validity of the OASIS (Study 3). 5.4 Study one: Pre-testing of developed items Cognitive interviewing (CI) was used to assess the comprehensibility and conceptual interpretability of a pool of 88 draft items of the emergent scale. Draft items were representative statements that reflected major themes related to the experience of sexuality and the factors associated with these experiences for heterosexual men and women aged 45 years and over identified in a previous study (Macleod et al., in press). CI is a method of identifying and correcting potential phrasing issues when developing survey or measurement scale items (Collins, 2003; Peterson, Peterson, & Powell, 2017). This approach is particularly useful when developing complex survey items, or where items relate to a sensitive topic (Drennan, 2002) such as sexuality. The CI process involves discussing each draft survey item with participants to examine the conceptual clarity, comprehensibility, and interpretability of items by the target population. This typically involves: examining the way a participant interprets items based on their wording; collecting information on the way a participant selected their response to the item(s); identifying difficulties

105

participants may have had in answering an item; and discussing the perceived relevance of items to the participant (Beatty & Willis, 2007). 5.4.1 Method 5.4.1.1 Participants. A convenience sample of 19 heterosexual men and women aged 46-79 (32% men, 68% women; 68% midlife, 32% later life) were recruited using a combination of radio interviews, social media, and printed and online media. Eligibility criteria were: at least 45 years of age, able to speak and read English competently, living independently within the community, and identify as heterosexual. Those interested in participating in the study were asked to register their interest online via a participant registration sheet or to contact the project team directly by phone or email. The study team contacted each person who indicated their interest in the study using the phone number provided in the registration process and confirmed their interest in participating in the study prior to conducting the phone interviews. 5.4.1.2 Materials. Cognitive interviews utilised a pool of 88 draft survey items developed using responses to open-ended survey items asking about sexuality and intimacy from a previous study (Macleod et al., in press). The draft items comprised nine separate subscales that reflected preciously identified themes describing older adults’ sexuality experiences (Macleod et al., in press). Three of the nine subscales reflected topics associated with the experience of sexuality, including the intimacy and pleasure – aspects (IPA), sexual expression (SE), and sexual urges (SU) subscales. The remaining six subscales reflected factors associated with the experience of sexuality. These were the intimacy and pleasure – factors (IPF), partner compatibility (PC), determinants of sexual desire (DSD), determinants of sexual expression (DSE), barriers to intimacy (BI), and barriers to sexual expression (BSE) subscales. Responses to the draft items were scored on a 5-point Likert-type scale where 1 = 'does not apply at all', and 5 = 'completely applies'. Four separate interview schedules were created to minimise the burden on respondents during the cognitive interviews (CIs). Each schedule comprised 22 draft survey items randomly drawn from the total pool of 88. The subsets were created to ensure that each item from the total pool of items was allocated to one interview schedule only. Thus, respondents could be allocated to one of four CI groups, allowing multiple assessments to be made on each draft item, while limiting

106

participant burnout overall. In total, each draft item was reviewed by between four and five respondents. 5.4.1.3 Procedure. Ethics approval was obtained from the Human Research Ethics Committee prior to recruitment. CIs were conducted by telephone. Prior to the interview, individuals were read information about the purpose of the study and asked if they still wished to participate in the study. No individuals chose to opt-out of the study. Participants were then randomly allocated to one of the four CI schedules and asked to provide feedback on the corresponding subset of survey items. At the start of the interview, only information on the participant's sex, age, and relationship status were collected. During the interview, the interviewer read out each draft survey item and the response options on the assigned interview schedule. A female psychology graduate with tertiary training and professional experience in qualitative data collection conducted the interviews. The item and/or response scale was repeated as often as necessary for the participants to respond to the questions about the item. After hearing each survey item and response scale, participants were asked questions about the clarity and relevance of the item and response scale, and encouraged to give as detailed feedback as possible. The interviewer took written notes during the interview, recording the participant's responses to each interview question as carefully as possible, before reading back the response to participants to verify the accuracy of these notes. Interviews took an average of 30 minutes to complete. At the end of the interview, participants were thanked for their time and asked if they had any follow- up questions for the researcher. 5.4.1.4 Data analysis. Responses from the CIs were collated in Microsoft Excel, and a 'scorecard' for each draft survey item was produced. Each item was scored against six categories: clearly worded item; item misinterpreted by respondent; rephrasing suggestions offered; assigned scale appropriate to the item; clear response scale categories provided; and relevance of the item to the target cohort. "Yes" responses for each category received a score of 1, and "no" responses received a score of 0. Notes were made describing any item misinterpretations by participants. Where participants identified poor item clarity, alternate suggestions for phrasing provided by participants were recorded and later used to refine item phrasing. The final scorecard

107

and all comments made by participants were then reviewed by the authors collectively to improve the wording of items if necessary and split any double-barrelled items. At the end of the refinement process, items were randomly arranged into a single list of measurement items and allocated a new scale ID number. 5.4.2 Results CI scorecards can be found in the supplementary materials (Appendix D, Table D1). Of the 88 draft items, 56 items were considered clearly worded by all respondents. A further 23 items were considered clearly worded by all but one of the CI participants. Of these, only six items were misinterpreted by a given CI respondent, despite citing concerns about the item's clarity. Nine items were considered unclear by two or more CI participants. Of these nine items, only one was misinterpreted by a single CI participant, despite concerns about the item's clarity. Each of the seven misinterpreted items was misinterpreted by only one of the five CI respondents assigned to the relevant CI schedule. Seventy one items received at least one ‘yes’ response to the question “is there a better way to phrase the item?”. Specific suggestions made by participants for how to rephrase unclear items were incorporated for 44 items. No suggestions were offered by participants for improving the clarity of the remaining items. As such, item phrasing for these items was refined using interview notes, with revisions aiming to address participant misinterpretations for these items. Three items were identified as double-barrelled and split into separate items. A single item was identified as triple- barrelled and split into three items. This added a further five items to the draft item pool, creating a final item pool of 93 items. All respondents considered the assigned rating scale to be appropriate to the items and easy to understand. Two participants considered one item each to be too unclear to make comments on the appropriateness of the rating scale, or on the relevance of the item to the intended cohort. Two participants considered one item each to be irrelevant to the intended cohort. Given the small proportion of respondents who identified some of the items as irrelevant, no items were removed at this stage. The five-point Likert-type response scale was also retained for all items. 5.4.3 Discussion Cognitive interviewing was used to assess conceptual clarity and, if necessary, refine the wording of scale items before undertaking a quantitative assessment of survey items. According to the Standards for Educational and Psychological Testing,

108

cognitive interviewing is an important part of scale development, and helps ensure that respondents can understand and respond to scale items in a consistent way (Linn, 2011; Peterson et al., 2017). The 93 items that resulted from the CIs were used to form the draft version of the Over 45s Adult Sexuality and Intimacy Scale (OASIS). Because item misinterpretation by participants can directly impact the construct validity of a measurement scale (Peterson et al., 2017), refining scale items using CIs provides scale developers an opportunity to identify whether items are being interpreted by respondents in the manner intended and creates opportunities to ensure that items are relevant and answerable by respondents (Peterson et al., 2017). The refined items were subsequently used in quantitative testing of a measurement scale to identify the best-functioning (and most informative) items for inclusion in the final scale. 5.5 Study two: Quantitative assessment of items and construction of subscales Rasch modelling was used to assess the measurement properties of the 93 draft items developed in the first study to select the most psychometrically robust items for inclusion in the final version of the OASIS. Rasch modelling aims to produce a measurement scale that is unidimensional, contains items that are free from measurement bias, and that uses a response scale that is correctly ordered (i.e., response options are ordered from least- to most-easily endorsed and are interpreted consistently by respondents). Rasch modelling involves a specific process of analysis whereby items that do not function well (or that do not add new information) can be identified and, where appropriate, removed so that the remaining items provide a parsimonious and precise assessment of the underlying trait (Bond & Fox, 2015). The original Rasch model was designed for use with dichotomous data. When working with polytomous data (e.g., those that use a Likert-type response scale), one of two Rasch model extensions can be used: the Rasch-Andrich rating scale model (RSM), or the Rasch-Masters partial credit model (PCM). While both approaches produce a similar result, RSM assumes that the distances between response scale points are equal across all items. In contrast, PCM makes no assumptions about the distance between response scale points prior to analysis and allows locations of response categories to vary across items. Since the PCM model allows for the examination of response category locations separately for each item, it is better suited to the identification of items with potentially disordered response categories (i.e., response options are not interpreted consistently by respondents from least- to most-

109

easily endorsed than the rating scale model. Consequently, PCM was used for this study. 5.5.1 Method 5.5.1.1 Participants. A convenience sample of 181 community-dwelling heterosexual men and women aged 45 years and above (45% men, 55% women; 61% midlife, 39% later life, see Table 5.1 for further demographics) was recruited to complete the OASIS items using a combination of radio interviews, social media, and printed and online media articles. Eligibility criteria were: at least 45 years of age, able to speak and read English competently, living independently within the community, and identify as heterosexual. Interested individuals were able to access the survey directly via a web- link or could contact the study team via email or phone to request a hard-copy of the survey. One hundred and seventy-six (97.2%) participants took part online, and five (2.7%) completed a hard-copy version of the survey.

Table 5.1 Study 2 participant demographics

Male % Female % Total % Midlife 35 43% 76 77% 111 61% Age group Later life 47 57% 23 23% 70 39% Present 66 80% 84 85% 150 83% Health condition Not present 16 20% 15 15% 31 17% Single 21 26% 26 26% 47 26%

Relationship status Partnered 52 63% 60 61% 112 62% Dating, not living 9 11% 13 13% 22 12% together Present 13 16% 23 24% 36 21% Partner health condition Not present 67 84% 72 76% 139 79% Perimenopause - - 25 100% 25 100% Menopausal status Post menopause - - 73 100% 73 100% Australia 59 74% 65 66% 124 48% Country of birth Overseas 21 26% 34 34% 55 38%

110

5.5.1.2 Materials. The participants responded to 15 demographic questions, in addition to the 93 OASIS draft items that were randomly ordered into a single survey. This number was reduced to 92 draft items after data cleaning identified one item as double-barrelled. Demographic information collected for the study included: sex (male/female), age cohort (45-64 years/65+ years), presence and type of any diagnosed health condition/s, relationship status (single/married/de facto/widowed/divorced/separated/dating but not living together), country of birth, presence and type of any diagnosed health condition/s in partner, and menopausal status (for females only). The online questionnaire was accessible via an anonymous web-link. A second anonymous web-link was located at the end of the online survey to redirect interested participants to a separate form to enter a prize draw for a gift card worth $50AUD. The (offline) post-pack included a print-out of the full questionnaire with a set of instructions and a postage-paid return envelope. A separate form and reply-paid envelope were included for participants to enter the prize draw. All prize draw data were collected and stored separately to ensure the anonymity of survey responses. 5.5.1.3 Procedure. Ethical approval from the Human Research Ethics Committee was obtained prior to recruitment. Data were collected both online and offline using post-packs. Online participants responded to screening questions before beginning the survey. Individuals who did not meet all eligibility criteria were redirected to a message thanking them for their time, and the survey was then terminated. Potential offline participants were screened for eligibility via phone or email before being sent the post-pack. If interested parties were not eligible, the research team thanked the individual for their interest and answered any further questions about the study or the eligibility criteria. Eligible participants responded to demographic questions before completing the OASIS items. 5.5.1.4 Data analysis. Using Microsoft Excel, raw score 'applicability' ratings for each item were calculated by summing item responses across all participants. Item responses were sorted into their respective subscales, and WINSTEPS software 3.70.0.5 (Linacre, 2009) was then used to perform Rasch modelling using PCM on OASIS items. For this analysis, each of the nine subscales was treated as an independent scale, as each

111

subscale was designed to assess a separate domain of sexuality and be able to be used on its own if desired. Assessing subscale functioning involved tests of dimensionality, reliability, and targeting using Rasch analysis. Content validity of subscales was assessed using raw score applicability measures for items followed by comparisons between subscale items. This was done to ensure that no items were removed that measured an aspect of the subscale’s overall theme that was not captured by other items within the subscale. Items within each subscale were examined for response category order, fit with the Rasch model, and the potential presence of measurement bias. Items that displayed disordered response categories, item misfit against the Rasch model and/or the presence of measurement bias were considered for removal or revision. To begin, the dimensionality of each subscale was examined by comparing the calculated empirical standard residual variance (i.e., how much variance in the data can be explained by the variance present within the Rasch model), modelled standard residual variance (i.e., how much variance can be attributed to the Rasch model, if the data were a perfect fit to the Rasch model), and the eigenvalue of the first residual contrast. A difference between the empirical standard residual variance and modelled standard residual variance of approximately 0-2% and principal component analysis of residual correlations producing an eigenvalue of the first contrast below 2.0 were considered indicative of unidimensionality (Bond & Fox, 2015). If the principle component analysis produced an eigenvalue of the first contrast equal to or greater than 2.0, the standardised variance scree-plot was used to examine whether subscale items clustered into underlying sub-topics. If sub-topic clusters were present on the scree-plot, this indicated that the subscale might not be unidimensional. Where sub- topic clusters were present within subscales that failed tests of unidimensionality, subscale items were split into separate groups, and each group analysed separately to determine whether splitting the items across two subscales produced unidimensional subscales. After examining the dimensionality of the subscale, subscale items were examined for whether the assigned Likert-type response scale functioned as expected for each item. If the assigned response scale was functioning as expected, category locations would be ordered from lowest (i.e., 'applies very little') to highest (i.e., 'applies very much'), or the presence of disordered response categories could be explained by a low endorsement rate of some response categories. If low endorsement

112

rates did not explain disordered response categories, neighbouring response categories were combined and the analysis was re-run. If collapsing neighbouring response categories did not correct the disordered response categories and/or did not improve the overall fit of the subscale with the Rasch model, subsequent analyses were performed using the original response scale and items with disordered response categories were flagged for potential removal from the subscale, pending further investigation. Once the response scale categories were finalised, person-item distributions were examined to see how effectively subscale items were targeting different levels of the underlying trait being measured. The person-item distribution histogram maps person ability against item difficulty to identify the range of abilities captured by the items in the subscale (Bond & Fox, 2015). Subscale reliability was then calculated using the person separation index (PSI), which estimates how well results can be generalised to different samples (Van Zile-Tamsen, 2017). PSI is comparable to Cronbach’s alpha (or KR-20 for dichotomous items) and has similar cut-off points for acceptable values (Clauser & Linacre, 1999). A subscale was considered reliable if it produced PSI reliability scores above 0.7. In the next step, the fit of individual items with the Rasch model was investigated using mean-square (MNSQ) infit/outfit measures. Misfitting items can be categorised as overfitting (i.e., items that produce data that is 'too predictable') or underfitting (i.e., items that produce data that is 'too unpredictable'). Overfitting items were those that had MNSQ infit/outfit values below 0.6, and underfitting items were those that had MNSQ infit/outfit values above 1.4 (Wright & Linacre, 1994). Any misfitting items were flagged as potentially problematic items. After identifying misfitting items, potential measurement bias was assessed with differential item functioning (DIF) to determine whether participant sub-groups responded to items in different ways. Sub-groups examined included sex, age group (45-64 years/65+ years), presence of one or more diagnosed health condition/s (yes/no), country of birth (Australia/overseas), relationship status (single/partnered/dating, not living together) presence of one or more diagnosed health condition/s in partner (yes/no), and menopausal status (yes/no) in women. No DIF analysis was possible for the mode of response (online/offline) because of the limited number of offline respondents. Any items that functioned differently across sub-groups were flagged for potential removal and ranked from largest DIF size and

113

greatest statistical significance to smallest DIF size and smallest statistical significance. Statistical significance tests for sub-group DIF assessments were performed using a Bonferroni correction (α = 0.05/k, where k is the number of items in a subscale). Calculation of the DIF contrast in Winsteps is equivalent to the Mantel-Haenszel significance test but can account for missing data where the Mantel- Haenszel test does not (Linacre, 2009). All items flagged for potential removal were ranked from most problematic to least problematic. Issues were weighted as follows: the size of misfit present (with underfit weighted more heavily than overfit), presence of a problematic response scale as indicated by disordered response categories, and the presence of significant DIF contrasts. The content of the most problematic item was then compared against all other items in the subscale for topic similarity and raw-score applicability measures. Higher raw-score applicability measures indicated a greater item content relevance for respondents, and lower raw-score values indicated lower item content relevance to respondents. Content validity was considered compromised if the most problematic item identified in the subscale reflected an aspect of sexuality that was not captured by the remaining items in the subscale. If it was deemed that the content validity of the subscale would not be affected by the removal of the most problematic item, the item was removed. After removing the item, each step of the data analysis process was repeated for the shortened subscale. Removal of items from the subscale ended when removal of the most problematic item was countered by a reduction in reliability scores and/or a reduction in the content validity of the subscale. At the end of Rasch modelling, the optimised subscales were combined and renumbered to produce the final version of the OASIS. 5.5.2 Results 5.5.2.1 Overall fit of original subscales. Statistical outputs from Rasch modelling are detailed in Tables 5.2 and 5.3. Standardised residual variances and eigenvalues of the first residual contrast provided evidence of unidimensionality for each of the nine subscales (see Table 5.2). Differences between empirical standardised residual variance and modelled standardised residual variance ranged between -1.0% and 1.5% in the initial versions of each subscale, supporting the unidimensionality of each subscale. Three subscales (SU, BI, and BSE) produced eigenvalues above 2.0 for the first residual contrast,

114

Table 5.2

Dimensionality assessment of original and final versions of the OASIS

Location (logits) Empirical raw Modelled raw Person Person (non-extreme) Item variance variance Reliability Subscale Version (by measure) (by measure) Difference First contrast (PSI) Mean S.D. n Mean S.D. n Intimacy and pleasure - aspects Original 48.10% 48.70% 0.60% 1.9 0.87 1.55 1.21 178 0 0.53 18

Final 52.80% 53.30% 0.50% 1.9 0.86 1.81 1.37 172 0 0.69 14

Change 4.70% 4.60% -0.10% 0 -0.01 0.26 0.16 -6 0 0.16 -4

Intimacy and pleasure - factors Original 47.40% 48.90% 1.50% 1.5 0.79 2.14 1.38 172 0 0.81 8

Final ------

Change Nil Nil Nil Nil Nil Nil Nil Nil Nil Nil Nil

Sexual Expression Original 48.60% 49.20% 0.60% 1.7 0.81 0.78 1.09 178 0 0.62 7

Final 43.90% 44.30% 0.40% 1.4 0.76 1.03 1.14 172 0 0.16 6

Change -4.70% -4.90% -0.20% -0.3 -0.05 0.25 0.05 -6 0 -0.46 -1

Sexual Urges Original 46.10% 45.50% -0.60% 2.2 0.76 0.42 0.92 179 0 0.54 8

Final 53.20% 53.10% -0.10% 1.8 0.74 0.39 0.99 177 0 0.63 7

Change 7.10% 7.60% 0.50% -0.4 -0.02 -0.03 0.07 -2 0 0.09 -1

115

Location (logits) Empirical raw Modelled raw Person Person (non-extreme) Item variance variance Reliability Subscale Version (by measure) (by measure) Difference First contrast (PSI) Mean S.D. n Mean S.D. n Partner Compatibility Original 38.70% 40.00% 1.30% 1.8 0.76 2.21 1.39 160 0 0.32 9

Final 45.40% 45.50% 0.10% 1.5 0.75 2.5 1.6 158 0 0.36 7

Change 6.70% 5.50% -1.20% -0.3 -0.01 0.29 0.21 -2 0 0.04 -2

Determinants of Sexual Desire Original 40.00% 39.30% -0.70% 1.4 0.77 0.67 1.05 179 0 0.39 8

Final ------

Change Nil Nil Nil Nil Nil Nil Nil Nil Nil Nil Nil

Determinants of Sexual Expression Original 37.50% 38.30% 0.80% 1.8 0.66 0.18 0.53 181 0 0.49 12

Final 40.50% 41.30% 0.80% 1.6 0.74 0.36 0.71 180 0 0.48 9

Change 3.00% 3.00% 0.00% -0.2 0.08 0.18 0.18 -1 0 -0.01 -3

Barriers to Intimacy Original 61.30% 60.30% -1.00% 2.4 0.79 -0.94 0.99 176 0 0.93 9

Final 62.20% 61.60% -0.60% 2.4 0.76 -0.89 0.97 176 0 0.97 8

Change 0.90% 1.30% 0.40% 0 -0.03 0.05 -0.02 0 0 0.04 -1

Barriers to Sexual Expression Original 46.10% 45.30% -0.80% 2.2 0.79 -0.73 0.75 164 0 0.62 13

Final 49.20% 48.70% -0.50% 1.9 0.78 -0.75 0.82 164 0 0.69 11

Change 3.10% 3.40% 0.30% -0.3 -0.01 -0.02 0.07 0 0 0.07 -2

116

indicating potential non-unidimensionality. Examination of the standardised residual variance scree-plot for these subscales showed that clustering items did not appear to capture any meaningful sub-topics within the subscales, ruling out potential non- unidimensionality of the subscales. 5.5.2.2 Response category ordering. Examination of empirical item-category measures showed that the rating scale functioned as expected for 75 of the 92 items across the nine subscales. Low response category endorsement rates accounted for the disordered rating scale categories present for 15 of the 17 items identified as having poorly functioning rating scales. Disordered response categories indicate that there may be a problem with how the response categories are being endorsed by respondents, in that disordered response categories cannot be confirmed as measuring a greater (or lesser) amount of the underlying trait than its neighbouring disordered response category. For example, category disordering was present for two items from the BSE subscale (OASIS_61 and OASIS_68). Examining empirical item-category measures, and response category endorsement rates were unable to confirm whether the highest response category (5 = ‘completely applies’) measured a greater amount of the underlying trait than the second-highest response category (4 = ‘applies very much’). Similarly, category disordering for one item from the PC subscale (OASIS_18) could not be explained by the low endorsement of the top three response categories (3 = ‘moderately applies’; 4 = ‘applies very much’, and 5 = ‘completely applies’). Combining neighbouring response categories (i.e., response categories 4 and 5, or response categories 3 and 4) failed to correct the disordering. Because the rating scale categories functioned as expected for all but three items, a decision was made to leave the response scale in its current form, and the three items with unexplained disordered response categories were flagged for possible removal from their respective subscales following subsequent analysis. 5.5.2.3 Reliability and targeting. Reliability of the initial subscales produced PSI scores that ranged from 0.66 (DSE subscale) to 0.87 (IPA subscale). For further details of subscale performance, please refer to Table 5.2. An assessment of person-item distributions for each subscale showed that items on the IPA, IPF, SU, PC, and DSD subscales target low levels of the measured traits, items on the SE and DSE subscales target midrange levels of the

117

measured traits, and items on the BI and BSE subscales target midrange to high levels of their respective traits (Figure 5.1). 5.5.2.4 Item fit. Two overfitting items (as indicated by MNSQ infit/outfit values below 0.6; Wright & Linacre, 1994) were present in the BI subscale (OASIS_54 and OASIS_91). Eight underfitting items (as indicated by MNSQ infit/outfit values above 1.4; Wright & Linacre, 1994) were present in the IPA (OASIS_28 and OASIS_55), DSE (OASIS_79 and OASIS_86), and the IPF, SU, PC, and BSE subscales (OASIS_11, OASIS_85, OASIS_33, and OASIS_61, respectively). No items were identified as misfitting for the SE, PC, DSD, or BI subscales. The individual item fit for each subscale, as determined by MNSQ infit and outfit scores, is provided in the supplementary material (Appendix D, Table D2). 5.5.2.5 Measurement bias. Tests of measurement bias using DIF analysis indicated that most items functioned similarly across the different subgroups, excepting 13 items (see Table 5.3). Predominantly, significant DIF values were produced by relationship status (single vs. partnered subgroups) and sex (male vs. female subgroups). Single respondents more frequently provided a higher applicability rating than partnered respondents for two items each from the BI (OASIS_46 and OASIS_93) and BSE (OASIS_36 and OASIS_41) subscales. Partnered respondents more frequently provided a higher applicability rating than single respondents for a single item on the BI subscale (OASIS_66). Men more frequently provided a higher applicability rating than women on a single item within the SE subscale (OASIS_71), and women more frequently provided a higher applicability rating than men on a single item within each of the IPF, SE, SU and DSE subscales (OASIS_78, OASIS_14, OASIS_38, and OASIS_16 respectively). Respondents born outside of Australia more frequently provided a higher applicability rating than respondents born in Australia for a single item within the DSD subscale (OASIS_16). 5.5.2.6 Scale refinement. Results for item assessment tests can be found in Table 5.3 and supplementary tables (Appendix D, Table D2). In total, 14 items were removed without compromising the content validity of subscales. The final version of the OASIS retained 78 of the original 92 draft items (Appendix D, Table D3), with a total of four

118

Figure 5. 1 Person-item distribution of the final version of the OASIS subscales.

119

Table 5.3 DIF analysis of original and final versions of the OASIS subscales

t-value Bonferroni Version Scale Subgrouping Item # Item Direction DIF contrast (d.f.) p correction Original Intimacy and pleasure - aspects Nil significant - - - - - 0.001

(n = 18)

Intimacy and pleasure - factors Sex (male/female) OASIS_78 It is important to me that I know my partner m < f -11.1 - 4.21 0.000 0.001 finds me attractive (166) Relationship status* OASIS_11 My partner's smile can be a sexual turn on s > p -11.9 - 4.15 0.000 (n = 8) for me (91) Sexual expression Sex (male/female) OASIS_14 I feel that visual stimulation is an important m < f 7 3.26 0.001 0.001 part of sexual desire (170) OASIS_71 I use sensual body movements as a way to m > f -8.5 4.66 0.000 (n = 7) express myself sexually (172) Sexual urges Sex (male/female) OASIS_38 I can express my sexuality through viewing m < f -7.6 4.45 0.000 0.001 of either print media or film (173)

(n = 8)

Partner compatibility Nil significant - - - - - 0.001

(n = 9)

Determinants of sexual desire Sex (male/female) OASIS_16 How comfortable I feel in my body is m < f -7.4 -3.41 0.001 0.001 important to how I engage with my (174) sexuality

OASIS_60 How attractive I feel is important to how I m < f -8.7 -4.24 0.000 (n = 8) engage with my sexuality (175)

Birth country (AU/Overseas) OASIS_16 How comfortable I feel in my body is AU < Overseas 8.3 3.42 0.001 important to how I engage with my (119) sexuality Determinants of sexual expression Relationship status* OASIS_79 I express my sexual behaviours on my own s > p -7.2 -4.35 0.000 0.001 (108) (n = 12) Barriers to intimacy Relationship status* OASIS_46 I avoid engaging with others sexually s > p -9.5 - 4.55 0.000 0.001 because I do not want to be in a relationship (128) right now

120

t-value Bonferroni Version Scale Subgrouping Item # Item Direction DIF contrast (d.f.) p correction OASIS_66 My energy levels can be a barrier to being s < p 10.2 4.86 0.000 (n = 9) intimate (101) OASIS_93 Fear of commitment prevents me from s > p -10.8 - 4.16 0.000 expressing myself sexually (117) Barriers to sexual expression Relationship status* OASIS_36 I find that I do not express myself sexually s > p -6.4 - 3.64 0.000 0.001 because of a lack of opportunity (96) OASIS_41 I find that I do not express myself sexually s > p -11.3 - 6.53 0.000 (n = 13) because of a lack of a suitable partner (106)

Final Intimacy and pleasure - aspects Nil significant - - - - - 0.001 (n = 14)

(n = 6)

Sexual urges Sex (male/female) OASIS_38 I can express my sexuality through viewing m < f -8.9 +/-4.91 0.000 0.001 of either print media or film (170)

(n = 7)

Partner compatibility Nil significant - - - - - 0.001

(n = 7)

Determinants of sexual desire Sex (male/female) OASIS_16 How comfortable I feel in my body is m < f -7.4 -3.41 0.001 0.001 important to how I engage with my (174) sexuality

OASIS_60 How attractive I feel is important to how I m < f -8.7 -4.24 0.000 (n = 8) engage with my sexuality (175)

(n = 9)

121

t-value Bonferroni Version Scale Subgrouping Item # Item Direction DIF contrast (d.f.) p correction Barriers to intimacy Relationship status* OASIS_46 I avoid engaging with others sexually s > p -10.2 - 4.90 0.000 0.001 because I do not want to be in a relationship (129) right now

OASIS_66 My energy levels can be a barrier to being s < p 9.69 4.69 0.000 (n = 8) intimate (101) OASIS_93 Fear of commitment prevents me from s > p -11.6 - 4.53 0.000 expressing myself sexually (117)

Barriers to sexual expression Relationship status* OASIS_36 I find that I do not express myself sexually s > p -6.6 - 3.67 0.000 0.001 because of a lack of opportunity (96)

OASIS_41 I find that I do not express myself sexually s > p -11.7 - 6.55 0.000 (n = 11) because of a lack of a suitable partner (105)

OASIS_75 Sometimes I do not want to express my s < p 7.1 3.45 0.001 sexuality (91) Sex (male/female) OASIS_49 I find that my partner's lack of interest m > f 6.1 4.01 0.000 interferes with my own sexuality (159)

* relationship status = single / partnered / dating, not living together

122

items removed from the IPA subscale (OASIS_8, OASIS_28, OASIS_55, and OASIS_63), three items removed from the DSE subscale (OASIS_53, OASIS_79, and OASIS_86) two each from the PC (OASIS_18, and OASIS_33) and BSE (OASIS_61, and OASIS_75) subscales, and one each from the SE, SU, and BI subscales (OASIS_71, OASIS_70, and OASIS_54, respectively). No items were removed from the IPF or DSD subscales. The most common reasons for item removal were item misfit, disordered response categories unexplained by low endorsement rates of response categories, and the presence of measurement bias as indicated by significant DIF values. 5.5.2.7 Overall fit of finalised subscales. Reliability of the final subscales was comparable to that of the initial subscales and ranged from 0.74 to 0.86 (see Table 5.2). All but one subscale (barriers to intimacy) produced an eigenvalue below 2.0 for the first residual contrast, indicating that the variance identified within the OASIS subscales could be explained by the natural variance present within the Rasch model. Item misfit was present for one item within the IPF subscale (OASIS_11) and significant DIF values were present on a total of 12 items across seven of the nine subscales (see Table 5.3). However, examination of the content of these 13 potentially problematic items indicated that their removal from their corresponding subscale would compromise the subscale's content validity and hence the items were retained. 5.5.2.8 Proposed scoring of OASIS. In scoring the OASIS in future applications, it is recommended that subscales scores are produced by calculating an average score for a given respondent, across all items answered in a subscale, provided the respondent has answered a minimum of 50% of the subscale items. Higher scores on each subscale will correspond to greater applicability of a given subscale to a respondent, in line with the categories used in the 5-point Likert-type response scales. Scores on each subscale can be compared to identify a ranked list of topics considered more or less relevant by the respondent. The final version of the OASIS can be found in the supplementary material (Appendix D, Table D3). 5.5.3 Discussion Rasch modelling was used to refine the draft version of the OASIS to produce a final version of the OASIS containing 78 psychometrically robust items across nine subscales. Each subscale was confirmed as measuring a single underlying construct,

123

as indicated by unidimensionality assessments. In general, scales that assess only a single trait produce a more precise assessment of the measured trait (Bond & Fox, 2015). Response categories for the 5-point Likert-type response scale functioned as desired across all items included in the OASIS, indicating that item scores can be meaningfully combined into a total subscale score (Bond & Fox, 2015). The failure of items to capture a complete range of scores, including both low and high extremes suggests that the scale is best suited for use with non-clinical samples because items are not calibrated to distinguish between extreme scorers clearly. Item targeting is used to examine how well the 'spread' of items can differentiate between individuals with differing levels of an underlying trait (Bond & Fox, 2015). This spread of items has implications on what types of changes are more precisely identified by the subscale in question, in that change will be best identified when it occurs within the target range of items. For example, the IPA subscale contains items that target the lower ranges of the measured topic. As such, changes for individuals who score on the lower range of this subscale will be more precisely detected than changes within the higher ranges because the subscale items are less well-calibrated to distinguish between high scorers. Item fit with the Rasch model was established for all but one item in the final version of the subscales, indicating that each item measures the unidimensional trait captured by the corresponding subscale. Each subscale was shown to be a reliable and stable measure of the underlying trait, as indicated by person-item distributions and PSI assessments. However, the results of DIF assessments suggest that some items of OASIS are endorsed differently by individuals based on their sex, relationship status, or their birth country. As such, comparisons of groups in which a significant DIF value exists warrants caution, as these comparisons may produce biased results. 5.6 Study 3: Scale evaluation The psychometric evaluation of measurement scales is an important step in confirming how well the scale performs across different time points and settings, and for determining how well the scale measures the subject matter it sets out to measure. This third study describes the psychometric evaluation of the OASIS and reflects the third phase of scale development, as outlined by Boateng et al. (2018). Firstly, the dimensionality of the scale is re-examined using confirmatory factor analysis (CFA) to provide further statistical support for dimensionality assessments conducted during Rasch modelling in the second study. Assessing dimensionality provides a statistical

124

evaluation of the distribution of items across the nine subscales produced in study two. Next, assessments of the internal consistency, test-retest reliability, and concurrent (convergent and divergent) validity of the final version of the OASIS are conducted, to determine whether the OASIS is a psychometrically sound alternative to current later life sexuality measurement scales. Because the OASIS was designed to produce a measure of sexuality that more adequately reflects the experience of sexuality for adults in mid and later life and the factors associated with these experiences over existing sexuality measures, it was predicted that all OASIS subscales would be significantly correlated, at various strengths, with both matched and unmatched scales. This was expected because the way that ideas were grouped together in the OASIS has been shown to differ from existing conceptualisations of sexuality (Macleod et al., in press; Macleod & McCabe, 2019). The convergent and divergent validity of OASIS subscales were assessed against matched and unmatched existing sexuality measurement. To demonstrate convergent validity, OASIS subscales would need to produce significant moderate to strong correlations (i.e. r between .3-.8) with their matched subscales. Based on the presence of topic overlap, moderate to strong positive correlations were predicted between OASIS subscales with their matched subscales. Correlations were also predicted between OASIS subscales and unmatched subscales as a result of some topic overlap between these subscales. However, these correlations were predicted to be weaker than those produced between OASIS subscales and their matched subscales, demonstrating the lesser amount of overlap present between OASIS subscales with the unmatched subscales. Divergent validity was only predicted between subscales that measured barriers to sexuality and subscales that did not measure barriers to sexuality. The presence of divergent validity would be demonstrated by nonsignificant correlations between subscales, or the presence of weak negative correlations between subscales. Because no subscales were matched directly with the DSD subscale, predictions about correlations between this subscale and other matched subscales were made based on the content of existing subscales matched with other OASIS subscales. Given the limited overlap of topics included in the DSD subscale with other existing scales, it was expected that only weak to moderate correlations would be found between the DSD subscale and unmatched subscales. Specific predictions

125

for each OASIS subscale with matched and unmatched subscales are provided in Table 5.4. 5.6.1 Method 5.6.1.1 Participants. A convenience sample of 406 community-dwelling heterosexual men and women aged 45 years and over were recruited for this study (see Table 5.5 for demographics). Eligible participants were those who identified as heterosexual, were at least 45 years of age, could read English competently, were living independently within the community, and were able to provide consent legally. No exclusion criteria were set for current relationship status, medically diagnosed conditions, or nationality. Recruitment was targeted to sample approximately equal numbers across four subgroups. The final sample included 128 women and 95 men in midlife (45-64 years), and 86 women and 97 men in later life (65+ years). The age cut-off points used in this study align with those proposed by the United Nations (United Nations, Statistical Office, 1982). 5.6.1.2 Measures. Demographic questions and seven sexuality measurement scales were used to assess the experience of sexuality for heterosexual men and women aged 45 years and above and the factors associated with these experiences. Matched subscales were selected for validity assessments based on the face validity of included items with items from the matched OASIS subscale. Details of the OASIS and each matched subscale are outlined below. 5.6.1.2.1 The OASIS. The OASIS (developed in study two) measures a wide range of topics relevant to the sexuality and sexual experiences of heterosexual adults aged 45 years and over. It contains a series of demographic questions, followed by 78 items that formed three subscales that reflect the experience of sexuality in mid and later life and six subscales that reflect factors associated with the experience of sexuality in mid and later life. 5.6.1.2.2 Matched subscales for testing convergent validity. Six subscales taken from four existing sexuality measurement scales were matched with eight of the OASIS subscales to assess convergent validity. No existing scale or subscale was found that assessed a similar construct to that captured by the DSD subscale. Subscales were matched with OASIS subscales based on the

126

Table 5.4 Predicted correlations between OASIS subscales with matched and unmatched existing sexuality measurement subscales

Predicted correlation ASCS - Sexual PAIR i PAIR iii SASS III, iiB MOSIEC - Synthesis ASCS - Commitment OASIS subscales performance Barriers to intimacy Weak (-) Weak (-) Weak (-) Weak (-) Moderate (+) Weak (+)

Barriers to sexual expression Weak (-) Weak (-) Weak (-) Weak (-) Weak (+) Moderate (+)

Intimacy and pleasure - aspects Moderate/strong (+) Moderate (+) Moderate (+) Weak/moderate (+) Weak/moderate (-) Weak (-)

Intimacy and pleasure - factors Moderate/strong (+) Moderate (+) Moderate (+) Weak/moderate (+) Weak/moderate (-) Weak (-)

Determinants of sexual desire Weak/moderate (+) Weak/moderate (+) Weak/moderate (+) Weak/moderate (+) Weak (-) Weak (-)

Determinants of sexual expression Weak/moderate (+) Weak/moderate (+) Weak/moderate (+) Moderate (+) Weak (-) Weak (-)

Partner compatibility Moderate/strong (+) Moderate (+) Moderate (+) Weak/moderate (+) Weak/moderate (-) Weak (-)

Sexual expression Moderate (+) Moderate (+) Moderate/strong (+) Weak (+) Weak (-) Weak (-)

Sexual urges Moderate (+) Moderate/strong (+) Moderate (+) Weak (+) Weak (-) Weak (-) - negative correlation predicted, whereby scores on one scale increase as scores on the other decrease + positive correlation predicted, whereby scores on both scales increase together bold text denotes matched subscales

127

Table 5.5

Study 3 participant demographics

Male Female TOTAL No. % No. % No. % Midlife 95 49% 128 60% 223 55% Age group Later life 97 51% 86 40% 183 45% Present 63 33% 51 24% 114 28% Health condition Not present 129 67% 163 76% 292 72% Diagnosed sexual dysfunction Present 28 15% 2 1% 30 7% Not present 164 85% 212 99% 376 93% Single 47 24% 61 29% 108 27% Relationship status Partnered 133 69% 138 64% 271 67% Dating, not living together 12 6% 15 7% 27 7% Australia 180 94% 158 74% 338 83% Country of birth Overseas 12 6% 56 26% 68 17%

128

conceptual similarity of the construct being measured. However, because some items in each of the matched subscales measure similar topics as items included in unmatched OASIS subscales, it was expected that all included scales would correlate to some degree. An outline of each scale and matched subscale is provided below. The Personal Assessment of Intimacy in Relationships (PAIR) Inventory (Schaefer & Olson, 1981) is a measure of how an individual perceives their relationship across five different types of intimacy (emotional, social, sexual, intellectual and recreational), as well as how they would like it to be. The emotional intimacy subscale (PAIR i) was considered to be convergent with the IPA, IPF, and PC subscales of the OASIS. The sexual intimacy subscale (PAIR iii)was considered to be convergent with the SU subscale of the OASIS. The Senior Adult Sexuality Scale (SASS; E. Weinstein, 2013) measures sexual attitudes, sexual interests, and a small number of sexual activities for adults aged 60 years and over. The traditional sexual interest subscale (SASS - Part III, Scale II, Part B) was used as a convergent measure for the SE subscale of the OASIS. The Measures of Sexual Identity Exploration and Commitment (MoSIEC) scale (Worthington, Navarro, Savoy, & Hampton, 2008) is a measure of the processes of sexual identity development, as they relate to Marcia's (1966) model of identity development. The synthesis subscale was considered to be convergent with the DSE subscale. The Attitudes Related to Sexual Concerns Scale (ASCS; Cowden & Koch, 1995) assesses an individual's attitudes about sexual concerns and/or dysfunctions. The commitment and sexual performance subscales were used to assess convergent validity of the BI and BSE subscales, respectively. 5.6.1.3 Procedure. Ethics approval was obtained from the Human Research Ethics Committee prior to recruitment. Participants were recruited globally by circulating information about the study via seniors' associations, media outlets (including radio, online publications, and newsletters), social media, and participant word-of-mouth. A prize draw for a gift card worth $250AUD was included in the promotion of the study to increase participation rates. Participants could take part online or offline via post- pack. Online data were collected via the survey tool Qualtrics XM (Qualtrics, 2018). Offline data were collected by sending a hard-copy of the questionnaire to an address nominated by the participant with a set of instructions, a postage-paid return envelope

129

for the survey, and a postage-paid return envelope for the prize draw. All survey responses were consolidated in an MS Excel spreadsheet and cleaned before analysis. All prize draw data were stored in a separate file. Data cleaning involved removing incomplete responses from the survey data file and checking that all item responses were correctly coded. The cleaned data files were then uploaded into IBM SPSS Version 25 and IBM SPSS Amos Version 25 (IBM Corporation, 2017) for analysis. 5.6.1.4 Test-retest sub-study. 5.6.1.4.1 Participants. Seventy-three individuals who took part in the first round of data collection for the validation study volunteered to take part in the second round of data collection. Of these, complete test-retest data were collected from 62 participants (85% response rate). Further demographics are provided in Table 5.5. 5.6.1.4.2 Measures. Test-retest data collection included all demographic questions and the 78 OASIS items included in the first phase of data collection. Participants were not asked to complete the convergent scales contained in the first round of data collection a second time. 5.6.1.4.3 Procedure. Participants who completed the OASIS as part of the validation study were asked at the end of the survey if they would be willing to receive an email asking them to complete the survey again two weeks later. This length of time was selected to provide enough time for participants to forget their exact responses to the original survey while providing a limited window of opportunity for changes to the experience of sexuality or the factors related to these experiences to occur. To ensure all results remained anonymous, any individual who was willing to complete the survey a second time was redirected to a new webpage and asked to provide an email address for the second survey invitation to be sent to two weeks later. Participants then had an additional two weeks from the time of the invitation to complete the second survey. Because only online participants were invited to take part in the second round of data collection, all data were collected online using Qualtrics XM (Qualtrics, 2018). After all responses were collected, data were consolidated in an MS Excel spreadsheet and cleaned. Data cleaning involved removing incomplete responses from the data file and creating a second data file with a sub-sample of matching test-retest responses. Test-retest responses were matched using participant

130

age, sex, and mother's maiden name, with all retest responses successfully matched with original test results. The cleaned data files were then uploaded into IBM SPSS Version 25 and IBM SPSS Amos Version 25 (IBM Corporation, 2017) for analysis. 5.6.1.5 Data analysis. CFA was conducted for a nine-factor model, reflecting the subscale structure of the OASIS using IBM SPSS Amos Version 25 (IBM Corporation, 2017). Hu and Bentler (1999) recommend using a two-index presentation strategy in evaluating CFA to reduce the rate of Type I and Type II errors occurring than those associated with a single-index presentation strategy. Thus, for this analysis a two-index combination is reported for Root Mean Square Error of Approximation (RMSEA) and standardised root mean square residual (SRMR). Hu and Bentler (1999) have reported that this combination of fit-indices (with RMSEA > .06 and SRMR > .09) generated the least sum of Type I and Type II error rates as compared with other two-index combinations. Internal consistency of the OASIS was assessed using Cronbach's alpha (α), with values ≥.7 indicating good internal consistency of the scale (Heale & Twycross, 2015). Concurrent and discriminant validity of the OASIS was assessed by examining the correlation of each OASIS subscale with matched and unmatched subscales. Values of <.3, .3-.5, and >.5 for Pearson's r were used to indicate weak, moderate, and strong correlations between subscales, respectively (Heale & Twycross, 2015). ICC estimates for test-retest reliability and their 95% confidence intervals were generated using an absolute agreement, two-way mixed-effects model. Values of <.5, .5-.75, .75-.9 and >.9 were used to indicate poor, moderate, good, and excellent reliability, respectively (Koo & Li, 2016). Higher ICC scores indicated a higher amount of consistency in test scores across the two time points. 5.6.2 Results CFA demonstrated that the 9-factor model fit the data well when referencing RMSEA (0.07) and SRMR (0.09). Standardised and unstandardised loadings (standard errors) for items in the OASIS are provided in the supplementary material (Appendix D, Table D4). Each OASIS subscale demonstrated good internal consistency and good to excellent test-retest reliability (see Table 5.6). Data for matched scales were missing for 18 participants, producing a sample size of 388 for assessments of construct validity of the OASIS subscales. Table 5.7 outlines the convergent and discriminant validity measures of each OASIS subscale

131

Table 5.6 OASIS subscale reliability assessments using Cronbach’s α and ICC

Internal consistency Test-retest reliability 95% Confidence Interval F Test with True Value 0 Subscales α n ICC n Lower Bound Upper Bound F df1 df2 p Intimacy and pleasure - aspects .93 .94 .90 .96 15.88 61 61 *** Sexual expressions .82 .92 .86 .95 11.84 61 61 *** Sexual urges .81 .92 .87 .95 13.14 61 61 *** Intimacy and pleasure - factors .85 .93 .88 .96 14.00 61 61 *** Partner compatibility .87 406 .92 62 .87 .95 12.78 61 61 *** Determinants of sexual desire .83 .88 .80 .93 8.14 61 61 *** Determinants of sexual expression .77 .85 .75 .91 6.54 61 61 *** Barriers to intimacy .74 .88 .79 .93 8.00 61 61 *** Barriers to sexual expression .75 .90 .84 .94 10.14 61 61 *** ***p<.001

132

Table 5.7

Concurrent validity assessments of OASIS subscales and matched later life sexuality measurement scales using Pearson’s r

MOSIEC - ASCS - ASCS - Sexual PAIR i PAIR iii SASS III, iiB Synthesis Commitment performance Predicted Predicted Predicted Predicted Predicted Predicted OASIS subscales r r r r r r n range range range range range range < -.3 < -.3 < -.3 Barriers to intimacy < -.3 .27** .12* 0.02 -.16** .3-.5 .22** < .3 .14** < -.3 < -.3 < -.3 Barriers to sexual expression < -.3 .31** .16** .11* -.14** < .3 .15** .3-.5 .15** .3-.5 .3-.5 .15-.5 .15-.5 < -.3 Intimacy and pleasure - aspects .3-.8 .39** .44** .74** .49** -.37** -.22** .3-.5 .3-.5 .15-.5 .15-.5 < -.3 Intimacy and pleasure - factors .3-.8 .40** .38** .60** .45** -.31** -.23** .15-.5 .15-.5 .15-.5 < -.3 < -.3 Determinants of sexual desire 388 .15-.5 .45** .38** .60** .33** -.20** -.15** .15-.5 .15-.5 < -.3 < -.3 Determinants of sexual expression .15-.5 .34** .40** .56** .3-.5 .39** -.23** -0.07 .3-.5 < -.3 Partner compatibility .3-.8 .38** .35** .3-.5 .53** .15-.5 .44** -.15-.5 -.36** -.26** .3-.5 < -.3 < -.3 Sexual expression .3-.5 .32** .48** .3-.8 .74** < .3 .37** -.23** -0.08 < -.3 < -.3 Sexual urges .3-.5 .33** .3-.8 .42** .3-.5 .50** < .3 .20** -.14** -0.01 ** Correlation is significant at the .01 level (2-tailed) * Correlation is significant at the .05 level (2-tailed) Bold text denotes matched subscales, italic text denotes lower than predicted correlations

133

with matched and unmatched sexuality measurement scales, respectively. All OASIS subscales produced significant correlations with their matched subscales within the strengths predicted (See Table 5.4 and Table 5.7), except for the BI and BSE subscales. Both of these subscales produced lower than expected correlations with their matched subscales (ASCS – Commitment and ASCS – Sexual performance, respectively). This indicates that convergent validity is present for all OASIS subscales with their matched subscales, although the convergent validity between subscales that measure barriers to sexuality with their matched subscales is poor. Convergent validity of the OASIS subscales with unmatched measurement subscales was also demonstrated with significant correlations that were equal to or stronger than predicted, excepting correlations between the BI and SASS III, iiB, and the SE and SU subscales with the ASCS – Sexual performance subscale. These three correlations were identified as non-significant (see Table 5.7). The strongest correlations were found between the seven OASIS subscales that did not measure barriers to sexuality (IPA, IPF, PC, DSD, DSE, SE and SU subscales) with the traditional sexual interest subscale (SASS - Part III, Scale II, Part B). Discriminant validity was established between subscales measuring barriers to sexuality and subscales that did not measure barriers to sexuality. Significant weak to moderate negative correlations (p<.001) were found between all seven OASIS subscales measuring the experience of sexuality and associated factors (IPA, IPF, PC, DSD, DSE, SE and SU subscales) with the ASCS – Commitment subscale. Significant weak negative correlations (p<.001) were found for four of the seven OASIS subscales (IPA, IPF, PC, and DSD subscales) with the ASCS – Sexual performance subscale, with non-significant weak negative correlations found for the remaining three OASIS subscales (DSE, SE and SU subscales) with the ASCS – Sexual performance subscale (see Table 5.7). 5.6.3 Discussion This study provides evidence of the internal consistency, test-retest reliability, and the convergent and discriminant validity of the OASIS. These results suggest that the OASIS is an effective and psychometrically sound measure of sexuality for heterosexual men and women aged 45 years and above. It can be used by clinicians and researchers to investigate the experience of sexuality and the factors related to the experience of sexuality for heterosexual men and women aged 45 years and above. CFA results support the use of a nine-subscale structure in the final scale based on

134

scores produced on RMSEA, and SRMR fit indices. The nine subscales of the OASIS capture a range of topics related to sexuality and intimacy not well-reflected in other later life sexuality scales using an approach that aligns with the perspectives of mid and later life adults. Each subscale of the OASIS has been designed to function independently of other subscales. The OASIS subscales demonstrated good internal consistency and strong test- retest reliability, indicating that the OASIS scores tend to remain stable in a stable population (Boateng et al., 2018). The OASIS is also strongly correlated with other relevant measurement scales such as the PAIR inventory, SASS, MoSEIC, and ASCS (Cowden & Koch, 1995; Schaefer & Olson, 1981; E. Weinstein, 2013; Worthington et al., 2008), making it a reasonable alternative to these measurement scales when assessing sexuality and intimacy for heterosexual men and women in mid and later life. As such, the oasis provides a single sexuality measurement scale capable of measuring a range of sexuality sub-constructs that previously required the use of multiple sub-scales cherry-picked from existing sexuality measurement scales frequently not validated for use with older samples. Differences in the strength of correlation between existing sexuality measures and OASIS subscales may reflect differences in how topics are categorised in existing scales and how they are categorised in the conceptual framework that underpins the OASIS. Previous research has shown that many measurable sexuality constructs are correlated (e.g., Arias-Castillo et al., 2009; Deacon et al., 1995). As such, these differences are also likely to explain the moderate significant correlation between OASIS subscales and unmatched subscales used in this study. 5.7 General discussion To better understand the sexuality of older adults, it is essential to capture the sexual experiences and attitudes of these individuals in a comprehensive way. The OASIS utilises the perceptions of heterosexual men and women in mid and later life to highlight factors that are relevant to their sexuality and sexual experiences. In doing so, it provides researchers and clinicians with a new tool for examining sexuality in mid and later life that recognises the uniqueness of the sexual experiences of these cohorts. The process used to develop the OASIS conforms with stages two and three of best practice guidelines for scale development (Boateng et al., 2018) and builds on previous work that aligns with the first stage of best practice guidelines for scale

135

development (Macleod et al., in press; Macleod & McCabe, 2019). In study one, CI provided evidence that respondents were interpreting content of items of the OASIS as was intended by the research team, that items were considered relevant to how heterosexual men and women aged 45 years and over conceptualise their sexuality, and that the rating scale categories assigned to each category were easily understood and relevant to the items. In study two, Rasch modelling was used to produce a final measurement scale with 78 psychometrically sound items in nine subscales, each of which captures a separate sub-topic of sexuality relevant to heterosexual adults in mid and later life. By using Rasch modelling to refine each subscale, problematic items were identified and removed, thus increasing the likelihood of the OASIS performing well in future psychometric evaluations. In study three, the factorial structure of the OASIS was examined through the application of CFA, and the convergent and discriminant validity of the OASIS subscales were assessed through the examination of correlations with existing validated scales that measured related constructs. A critical difference between the OASIS and existing later life sexuality measures is the breadth of topics assessed. The OASIS provides researchers with a tool designed specifically for use with older populations that can measure a broad range of topics related to sexuality. Further, because the OASIS subscales are independent of each other, researchers and clinicians can administer all subscales or select only those subscales that are most relevant to their purposes, thus minimising responder burden. For example, the Senior Adult Sexuality Scale (SASS; E. Weinstein, 2013) measures sexual attitudes, sexual interest, and a range of sexual activities for older adults, and each subscale can be used individually or collectively. Similarly, the OASIS can also assess the topics captured within the SASS and can be used as a whole or in parts. However, the OASIS goes beyond the scope of topics included in the SASS and includes subscales that measure elements of partner compatibility, intimacy, pleasure, sexual urges, barriers to intimacy and sexual desire, and a range of factors associated with sexual desire and sexual expression. Many of the topics reflected in the OASIS have previously been investigated using unvalidated questionnaire item(s), measurement scales validated with younger populations, or a combination of both validated and unvalidated items (e.g., Matthias et al., 1997; Minichiello et al., 2004; Momtaz et al., 2014; Penhollow et al., 2009). For example, data collected by the Personal Assessment of Intimacy in Relationships (PAIR) Inventory (Schaefer & Olson, 1981) captures similar topics to those captured

136

by the IPA, IPF, PC, and SU subscales. However, the PAIR Inventory captures only a limited range of sexuality aspects and related factors, and the sample used to validate the PAIR Inventory only included individuals aged 21-60 years (Schaefer & Olson, 1981). The absence of any individuals in later life (i.e., 65+ years) in this sample may mean that the information obtained using the PAIR Inventory is not as reflective of the experiences of older adults as it is for younger adults. Each OASIS subscale reflects the areas that are relevant to how respondents experience sexuality and the factors associated with these experiences in mid and later life. This will help ensure that future sexuality and ageing research focuses on the topics relevant to older populations, and provides an alternative to the current highly- medicalised approach to understanding sexuality that focuses on sexual function and sexual activity frequency. In doing so, the OASIS is better-able to reflect the increasing importance of intimacy and relationships to the experience of sexuality as people age (Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; McCarthy et al., 2013). Scores on each OASIS subscale show how applicable the subscale topic is to how respondents experience sexuality and the different factors related to these experiences. This approach provides respondents with a way of identifying the areas that are most relevant to how they perceive their sexuality and sexual experiences and does not assume that 'sexuality' is synonymous with 'sexual activity'. Similarly, the OASIS allows for the possibility that the experience of sexuality is different at different stages of life, and that not everyone is focused on the attitudes and behaviours typical of younger populations. In contrast to many existing sexuality measurement scales, the OASIS does not ask respondents about sexual activity frequency or specific sexual functioning, except in terms of whether issues with sexual functioning act as barriers to sexual expression and/or intimacy. In doing so, the OASIS challenges the idea that sexual functioning and/or sexual intercourse are the central focus of sexuality for those in mid and later life. 5.7.1 Limitations Based on participant feedback from a previous study, exact age of participants was not collected. As such, only information about whether participants were in midlife (aged 45-64 years) or in later life (aged 65 years and over). As such, it was not possible to calculate mean age of samples and sub-samples, and this information should be examined in future studies. This study was designed to produce a sexuality

137

measurement scale for heterosexual men and women in mid and later life. No investigation of differences between participants with and without clinically diagnosed conditions was made in this study. As such, a further study that examined the response of clinical and non-clinical populations to this scale could identify whether the scale can distinguish between these groups. Arguments may be made against combining males and females, as well as mid and later life cohorts for this study. However, the inclusion of men and women, as well as a broad range of ages in this study, helps increase the heterogeneity of the sample, thus improving the generalisability of results to the broader population. While every effort was made to match the OASIS subscales with existing sexuality measurement no existing measurement scale was found that captured the topics included in the determinants of sexual desire OASIS subscale. As such, the convergent validity of this scale remains to be explored in future studies. Participants in this study were predominantly Australian, despite the study being promoted internationally. As such, further validation of the OASIS for ethnically and culturally diverse populations is still required. Similarly, LGBTQIA+ populations were not included in this study because the experiences of the LGBTQIA+ community were not reflected in study that the OASIS was developed from (see Macleod et al., in press). Thus, to reflect the sexual experiences of the LGBTQIA+ older populations accurately, future research is needed that examines the relevance of the OASIS for an exclusively LGBTQIA+ older sample. 5.8 Conclusions The OASIS is a new measure of sexuality for heterosexual men and women in mid and later life that has empirical evidence to support its structural and construct validity and reliability. This scale that has been developed directly from a new approach to understanding the experience of sexuality and the factors associated with these experiences for heterosexual men and women aged 45 years and above. By targeting only adults aged 45 years and above, data can be collected that reflect the experience of sexuality for adults in mid and later life, without assuming that this cohort seeks to maintain or return to the sexual experiences of their younger years. This scale provides an alternative means of understanding sexuality and intimacy beyond the approach often used in sexuality and ageing research that emphasises sexual functioning and sexual activity frequency, thereby creating new opportunities for how sexuality in mid and later life is understood.

138

Chapter 6. Study 3: Similarities and differences in the experience of sexuality for men and women in mid and later life

139

6.1 Chapter guide Macleod, A. & McCabe, M. P. (in preparation [2]). Similarities and differences in the experience of sexuality for men and women in mid and later life. Journal of Sex Research.

This chapter builds on work from earlier chapters to examine sex and age differences for men and women in mid and later life. The novel contribution made by this chapter is the broad examination of sexuality for heterosexual men and women in mid and later life using the new approach to understanding the experience of sexuality and associated factors, and the related measurement scale developed in this thesis (Chapters 4 and 5, respectively). This analysis provides valuable information about the way that men and women experience sexuality and the factors associated with these experiences currently. It uses MANOVA to identify differences between how cohorts score on each of the nine OASIS subscales based on sex and age group membership. The current chapter is presented in the style of a published manuscript, while still conforming to the style used in this thesis: abstract (Section 6.2), introduction (Section 6.3), methods (Section 6.4), results (Section 6.5), discussion (Section 6.6) and conclusions and future directions (Section 6.7). New information relating to areas of similarity and difference between men and women in mid and later life are presented, further expanding the current understanding of sexuality for heterosexual men and women in mid and later life. Ethics approval and related documentation for this study can be located in Appendix E.

140

6.2 Abstract Background: Knowledge describing how the experience of sexuality differs between men and women in mid and later life is limited in scope and provides conflicting information. Areas of similarity and difference between these cohorts were examined using a new approach to understanding the experience of sexuality for heterosexual adults aged 45 years and over, and a recently developed sexuality measurement scale. Methods: Three hundred and ninety-nine heterosexual men and women aged 45 years and over (47% men, 53% women; 55% midlife, 45% later life) completed the Over 45s Adult Sexuality and Intimacy Scale (OASIS). Two factorial MANOVAs were used to examine whether significant differences were present between groups based on age cohort membership (midlife/later life), biological sex (men/women) or an interaction between age cohort membership and biological sex on scores produced on the OASIS subscales. Data were examined separately for subscales measuring the experience of sexuality, and for subscales measuring factors associated with the experience of sexuality. Results: Sexual expression and sexual urges were more relevant to the experience of sexuality for men than women. Partner compatibility was more relevant to the experience of sexuality for women than men. Sexual urges were more relevant to the experience of sexuality for those in midlife than those in later life. Determinants of sexual desire were more relevant to the experience of sexuality for those in midlife than those in later life. No significant differences were found between groups for scales that assessed intimacy and pleasure, determinants of sexual expression, or barriers to intimacy or sexual expression. Conclusions: The experience of sexuality for men and women in mid and later life is more similar than previous research has indicated. Approaches to understanding sexuality in mid and later life that focus on sexual functioning and sexual activity frequency do not adequately capture the significant overlap in how these cohorts experience their sexuality and the factors that influence these experiences. Further research is needed that explores similarities and differences between these cohorts across a broader range of variables, including sexual orientation, relationship status, health status, and partner variables such as partner health status.

Keywords: sexuality, ageing, intimacy, cohort differences

141

6.3 Introduction Sexuality in mid and later life is a complex but understudied area of research (Baldissera et al., 2012; Heywood et al., 2018). Current approaches to the study of sexuality and ageing primarily focus on sexual functioning and sexual activity frequency and are predominantly developed from knowledge about sexuality during younger adulthood (McCarthy et al., 2013; McHugh & Interligi, 2015). However, there is limited evidence that older adults aspire to a 'youthful ideal' in terms of their sexuality, sexual functioning, or sexual experience (Deacon et al., 1995; Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; Gott & Hinchliff, 2003b; McCarthy et al., 2013; Sandberg, 2013). This failure of current approaches to incorporate the perspectives of older adults into conceptualisations of sexuality across the later life stages means that existing research findings are unlikely to reflect the sexual experiences of older people fully. Despite these limitations, several important differences have been identified in terms of the experience of sexuality for men and women, and for those in mid and later life. Thus far, research examining the sexuality of men and women in mid and later life has focused primarily on the association of biological sex and/or age with an individual's sexual behaviours and the factors associated with these behaviours. In line with this, differences between men and women have been found across a range of sexuality measures. For example, men report a higher amount of sexual pleasure, sexual desire and sexual interest than women (Arias-Castillo et al., 2009; Carvalho & Nobre, 2010; Johnson, 1996; Knodel & Chayovan, 2001). Men also report higher masturbation frequencies, higher rates of sexual participation, and place greater importance on sexual activity than do women (Arias-Castillo et al., 2009; Johnson, 1996; O. Kim & Jeon, 2013; Papaharitou et al., 2008). However, the impact of sex on other aspects of sexuality is less clear. For example, men frequently report a higher amount of sexual activity than do women (Arias-Castillo et al., 2009; Beckman et al., 2014; Bortz & Wallace, 1999; Dello Buono et al., 1998; O. Kim & Jeon, 2013; Momtaz et al., 2014; Papaharitou et al., 2008; Trudel et al., 2013; Wang et al., 2008), yet some studies suggest that these differences are not significant (Padoani et al., 2000; Papaharitou et al., 2008; Shkolnik & Iecovich, 2013). Several studies have also identified significant differences between mid and later life cohorts. For example, sexual functioning typically decreases with age (Connaughton & McCabe, 2015), adults in midlife report higher frequencies of sexual

142

activity than those in later life (Bortz & Wallace, 1999; Bortz et al., 1999; Papaharitou et al., 2008), and sexual activity is considered more important for midlife cohorts than later life cohorts (Müller et al., 2014). Midlife cohorts also report a higher amount of sexual desire, sexual interest, and arousal than do later life cohorts (Carvalho & Nobre, 2010; Dello Buono et al., 1998; Johnson, 1996; Knodel & Chayovan, 2001; Padoani et al., 2000; Papaharitou et al., 2008; Thompson et al., 2011; Trudel, 2002). However, the higher levels of sexual desire, sexual interest and arousal reported by midlife cohorts do not appear to extend to differences between mid and later life cohorts in terms of the desired frequency of sex (Iveniuk, O'Muircheartaigh, & Cagney, 2016). Inconsistency is also present in terms of differences between men and women and age differences relating to factors associated with intimacy and affection. For example, women report a greater importance of intimacy across mid and later life than do men (Johnson, 1996). Likewise, affection has been shown to become increasingly important to individuals as age increases (Müller et al., 2014). However, other research has indicated that older men consider affection to be more important than older women, and that older men consider a love attachment necessary for sex moreso than older women (Iveniuk et al., 2016; Moore, 2010; Müller et al., 2014). Current approaches to understanding the experience of sexuality for adults in mid and later life are further complicated by the limited investigation of topics that older adults consider relevant to their experience of sexuality (Macleod & McCabe, 2019). As a result, the differences between individuals based on their sex and age cohort membership for many areas of sexuality remain unknown. However, identifying sex and age cohort differences across a more diverse range of sexual experiences and the factors associated with these experiences will require a step away from the dominant approach used in sexuality and ageing research that centres on sexual function and sexual activity frequency. By examining areas of similarity and difference between men and women in mid and later life more broadly, our understanding of the unique experience of sexuality and the factors associated with these experiences across the later stages of life can be expanded. This study uses a new approach to understanding sexuality across mid and later life (Macleod et al., in press) and the associated measurement scale (Macleod & McCabe, in preparation [1]) developed to investigate whether significant differences are present in how heterosexual men and women aged 45 years and above experience

143

sexuality and the factors associated with these experiences based on biological sex (men/women) and age cohort membership (midlife/later life). This new approach was developed utilising the results of a group concept mapping study that used the thoughts, perspectives, and experiences of 40 heterosexual men and women aged 45 years and above to identify the areas they considered relevant to their sexuality (Macleod et al., in press). Based on the descriptions and experiences provided within the group concept mapping study, the experience of sexuality for heterosexual men and women in mid and later life was re-conceptualised to reflect several major sub-constructs. Three themes reflect the experience of sexuality and include: aspects of intimacy and pleasure, sexual expression, and sexual urges. Six themes reflect the factors related to the experience of sexuality and include: partner compatibility, factors associated with intimacy and pleasure, determinants of sexual desire, determinants of sexual expression, barriers to intimacy and barriers to sexual expression. Applying this novel approach to understanding sexuality across mid and later life creates a framework for understanding sexuality in mid and later life that does not rely on perspectives that emphasise sexual functioning and sexual activity frequency, or that is based on the medicalisation of the ageing process. This paper aims to determine areas of similarity and difference between men and women, and between those in mid and later life, in terms of their current experience of sexuality and the factors associated with these experiences. Data collected using the Over 45s Adult Sexuality and Intimacy Scale (OASIS; Macleod & McCabe, in preparation [1]) are examined using MANOVA to answer the following research questions: 1. Are there significant differences between men and women on the OASIS subscales? 2. Are there significant differences between middle-aged and older people on the OASIS subscales? 3. Are there interactions between biological sex and life stage on the OASIS subscales? Based on existing research, a main effect for sex is hypothesised for subscales that measure sexual expressions and sexual urges, whereby men are predicted to score higher than women on the sexual expressions and sexual urges subscales. No predictions have been made in relation to higher scores by women on any subscales,

144

because of conflicts in the existing literature. A positive main effect for age cohort membership (whereby it is predicted that those in later life will score higher than those in midlife) is also predicted for OASIS subscales measuring aspects of intimacy and pleasure and the factors associated with intimacy and pleasure, and for subscales that measure barriers to sexual expressions. Negative main effects (whereby it is predicted that those in midlife will score higher than those in later life) are predicted for subscales that measure sexual expression, sexual urges and determinants of sexual expression. No specific hypotheses were developed relating to main effects for respondent’s sex and age cohort membership on scores produced on other OASIS subscales. Interaction effects between a respondent’s sex and age cohort membership were not predicted for any of the OASIS subscales, given the lack of consensus in the existing literature on other measures of the constructs evaluated in the OASIS. 6.4 Method 6.4.1 Participants A convenience sample of 399 adults in mid and later life (47% men, 53% women) participated in this study. A total of 220 participants were aged between 45 and 64 years (42% men, 68% women), and 179 participants were aged 65 years or over (53% men, 47% women). Further demographics are outlined in Table 6.1. Eligible participants were heterosexual, aged 45 years or over, could read English competently, lived independently within the community, and were legally able to provide consent. Individuals were not excluded based on their current relationship status, the presence of one or more medically diagnosed health or sexual health conditions, or their country of residence. Recruitment was targeted to sample approximately equal numbers of individuals across four subgroups: men in midlife, men in later life, women in midlife, and women in later life. 6.4.2 Materials The OASIS is a sexuality measurement scale designed to assess a wide range of topics relevant to the sexuality and sexual experiences of heterosexual adults aged 45 years and over. Psychometric evaluations have shown that the scale has good internal consistency (α = .95) and good to excellent test-retest reliability (ICC = .85- .90, p<.001; Macleod & McCabe, in preparation [1]). The OASIS contains 78 items distributed across nine subscales that describe the experience of sexuality and the factors related to these experiences for heterosexual adults in mid and later life.

145

Three subscales measure the experience of sexuality in mid and later life. These subscales are the intimacy and pleasure - aspects (IPA), sexual expression (SE), and sexual urges (SU) subscales. The IPA subscale contains 14 items and provides an assessment of how applicable a range of affectionate, intimate, and some sexual behaviours are on the respondent’s experience of sexuality. This subscale emphasises the type and/or quality of the intimate relationship. Topics included in the IPA subscale include trust, understanding, attraction, playfulness, anticipation, emotion, communication, energy, sharing, and other intimate behaviours. The SE subscale contains seven items and provides an assessment of how applicable specific sexual acts or partnered sexual activities are to how a respondent engages with and/or expresses their sexuality currently. The SU subscale contains seven items and measures how applicable statements relating to sexual desire, physical sexual urges, and other forms of sexual expression are to an individual’s experience of sexuality currently. Respondents are asked to indicate how well each statement applies to their current feelings about sexuality and intimacy on a 5-point Likert-type scale (from 1 = ‘does not apply at all’ to 5 = ‘completely applies’). Each of the nine subscales is scored independently to produce an average applicability rating for the subscale construct. The average applicability rating for each subscale is calculated by adding the number value assigned to the nominated response scale category (ranging from 1- 5) for each item in the subscale. The summed total for each subscale is then divided by the total number of items in the subscale. The final (average) score for the subscale is interpreted using the same five response categories assigned to items on the Likert- type scale. High scores indicate a greater relevance of the subscale construct to an individual’s current experience of sexuality. Six subscales measure factors associated with sexuality in mid and later life. These subscales are the intimacy and pleasure - factors (IPF), partner compatibility (PC), determinants of sexual desire (DSD), determinants of sexual expression (DSE), barriers to intimacy (BI), and barriers to sexual expression (BSE) subscales. The IPF subscale contains eight items and assesses how applicable a range of factors related to affectionate, intimate, and some sexual behaviours are to the experience of sexuality for respondents, with an emphasis on the type and/or quality of the intimate relationship. The PC subscale contains seven items and examines how applicable factors associated with feeling comfortable with and cared for by a romantic partner

146

are to the experience of sexuality for the respondent. Topics assessed within this subscale include communication between partners, acceptance, and mutual respect between partners. The DSD subscale contains seven items and measures how applicable factors associated with the experience or onset of sexual desire are to the experience of sexuality for respondents. The DSE subscale contains nine items and assesses how applicable factors related to sexual expression are to the experience of sexuality for respondents. Topics in this subscale include the way that respondents think about and express themselves sexually, their sexual standards, and the ways they engage sexually with others. The BI subscale contains eight items and assesses how applicable a range of factors that can impinge upon a person's willingness to engage with another person on an intimate level are to the experience of sexuality for respondents. Finally, the BSE subscale contains 11 items and assesses how applicable a range of medical, and/or social factors that function as barriers to sexual expression are on the experience of sexuality for the respondent. 6.4.3 Procedure Ethics approval for the study was received from the Human Research Ethics Committee prior to recruitment. Recruitment involved the international distribution of information about the study via seniors’ associations, media outlets (including radio, online publications, and newsletters), social media, and word-of-mouth promotion by participants, including the promotion of a prize draw for a gift card worth $250AUD to increase participation rates. Participants could choose to take part online using the survey tool Qualtrics XM (Qualtrics, 2018), or offline via post-pack. Participants taking part offline were sent a hard-copy of the survey to an address nominated by the participant. The post-pack included survey instructions, a printed copy of the survey, a prize draw entry form, a postage-paid return envelope for the completed survey, and a separate postage-paid envelope for the prize draw entry. Prize draw data were stored in a separate file to preserve response anonymity. All online and offline survey responses were consolidated in a Microsoft Excel spreadsheet and cleaned before analysis. Incomplete survey responses were removed from the data file, and all remaining responses were checked for correct coding. The cleaned data files were then uploaded into IBM SPSS Version 25 (IBM Corporation, 2017) for analysis. 6.4.4 Data analysis MANOVA was used to examine differences between men and women and between those in mid and later life across the nine OASIS subscales. The data

147

included two dichotomous categorical IVs (sex/age cohort membership), and nine continuous DVs (OASIS subscales). Groups were larger than the minimum recommended 20 cases per independent variable (Ho, 2013). Independence of observations was confirmed for each group, but group sizes were not equal. Data were examined for univariate and multivariate outliers using Mahalanobis distances, and six cases with Mahalanobis distances above the critical Mahalanobis distance (d.f. = 6, critical c2 = 22.46) were removed from the dataset. An assessment of correlations between variables indicated that two subscales correlated with other OASIS subscales at a level that compromised analysis using a single MANOVA. This issue was resolved by splitting the subscales into those that measured experiences related to sexuality in mid and later life, and those that measured factors associated with the experience of sexuality in mid and later life. Likewise, the high correlation between the PC and IPF subscales (r = .87) was anticipated because the presence of intimacy within a relationship implies that a reasonable degree of partner compatibility will also be present. Thus, the data were confirmed as non-collinear. After splitting the data into subscales relating to the experience of sexuality and subscales relating to factors associated with the experience of sexuality, covariance matrices were re-evaluated. Separating the data into two groups resulted in correlations between variables being compatible with MANOVA assumptions. All correlations between subscale totals were within accepted guidelines (Garson, 2012) except for the correlation between the BI subscale with the DSE, IPF, and PC subscales (r = .16, .10, and .13, respectively), and the BSE and IPF and PC subscales (r = .06 and .02, respectively). These low correlations were anticipated because they reflect the difference between factors associated with the experience of sexuality and factors that inhibit the experience of sexuality. Two 2x2-factorial MANOVAs were conducted to examine whether biological sex (men/women), age cohort membership (midlife/later life), or an interaction between sex and age cohort membership had a statistically significant effect on the score on each of the nine OASIS subscales. The first MANOVA examined the difference between groups based on sex and age cohort membership of respondents across the three subscales that measure aspects related to the experience of sexuality. The second MANOVA examined the difference between groups based on sex and age

148

cohort membership of respondents for the six subscales that measure factors associated with sexuality. 6.5 Results 6.5.1 Subscales measuring aspects related to the experience of sexuality MANOVA produced significant main effects for sex, F(3,393) = 19.67, p<.001, η2 = 0.13 and age cohort membership, F(3,393) = 5.94, p<.001, η2 = 0.04 for subscales that measured aspects related to the experience of sexuality. Biological sex and age cohort membership accounted for 13% and 4% of the variance in the canonically derived dependent variable across these subscales, respectively. The interaction effect for sex and age cohort membership was not significant for the three subscales, indicating that there is no interaction between biological sex and age cohort membership for how participants scored on the subscales that reflect the experience of sexuality for heterosexual men and women in mid and later life. Univariate results showed that a significant effect was present between men and women on the SE and SU subscales, with men scoring higher than women on these subscales (Table 6.2 – Aspects subset). A significant effect was also present between individuals in mid and later life for the SU subscale (Table 6.2 – Aspects subset), with midlife individuals scoring higher than later life individuals on this subscale. No significant effect for biological sex or age cohort membership was found for the IPA subscale, nor was one found for age cohort membership on the SE subscale. Table 6.3 (Aspects subset) provides a summary of identified similarities and differences between groups based on sex and age cohort membership. Table 6.4 (Aspects subset) displays mean scores and standard deviations of each OASIS subscale for groups based on sex and age cohort membership. 6.5.2 Subscales measuring factors associated with the experience of sexuality MANOVA produced significant main effects for sex, F(6,390) = 3.80, p<.001, η2 = 0.06 and age cohort membership, F(6,390) = 5.71, p<.001, η2 = 0.08 for subscales that measure factors associated with the experience of sexuality. Biological sex and age cohort membership accounted for 8% and 6% of the variance in the canonically derived dependent variable across these subscales, respectively. No significant interaction effect was found for sex and age cohort membership on these subscales. Univariate results showed that a significant effect was present between men and women on the PC subscale (Table 6.2 – Factors subset), with women scoring

149

Table 6.1 Demographic profile of participants

150

Table 6.2 Significant univariate effects for sex and age cohort membership across OASIS subscales

Subscales IV F d.f. Error d.f. p η2 Sexual urges Age cohort 15.37 3.00 395.00 0.00 0.04 Sex 26.98 3.00 395.00 0.00 0.06 Aspects Sexual expression Age cohort - - - - - Sex 35.03 3.00 395.00 0.00 0.08 Partner compatibility Age cohort - - - - - Sex 8.00 3.00 395.00 0.01 0.02 Factors Determinants of sexual desire Age cohort 11.55 3.00 395.00 0.00 0.03 Sex - - - - -

151

Table 6.3 Summary of predicted and obtained main effects identified for each OASIS subscale, based on age cohort membership, sex, and age cohort membership and sex combined

Subscale Age cohort Sex Age cohort*Sex Intimacy and pleasure - Aspects ML = LL M = F = Aspects Sexual expression ML = LL M > F = Sexual urges ML > LL M > F = Intimacy and pleasure - Factors ML = LL M = F = Partner compatibility ML = LL M < F = Determinants of sexual desire ML > LL M = F = Factors Determinants of sexual expression ML = LL M = F = Barriers to intimacy ML = LL M = F = Barriers to sexual expression ML = LL M = F = *ML = midlife; LL = later life; M = male; F = female

152

Table 6.4 Mean and standard deviation of scores for age cohort membership and sex groups for each OASIS subscale

Midlife (45-64 years) Later life (65+ years) Total Subscale Sex Mean s.d. Mean s.d. Mean s.d. Male 3.79 0.66 3.86 0.63 3.82 0.64 Intimacy and pleasure - Aspects Female 3.80 0.58 3.70 0.86 3.76 0.70 Total 3.80 0.61 3.78 0.75 3.79 0.68 Male 3.42 0.71 3.21 0.79 3.31 0.76 Sexual expression Female 2.90 0.69 2.82 0.87 2.87 0.76

Aspects Total 3.12 0.74 3.02 0.85 3.08 0.79 Male 3.17 0.76 2.89 0.82 3.03 0.80 Sexual urges Female 2.79 0.73 2.44 0.84 2.65 0.79 Total 2.95 0.76 2.68 0.86 2.83 0.82 Male 3.85 0.67 3.93 0.59 3.89 0.63 Intimacy and pleasure - Factors Female 4.05 0.48 3.95 0.74 4.01 0.60 Total 3.97 0.58 3.94 0.66 3.95 0.62 Male 4.00 0.66 4.03 0.61 4.02 0.64 Partner compatibility Female 4.28 0.45 4.10 0.73 4.21 0.59 Total 4.16 0.57 4.06 0.67 4.12 0.62 Male 3.50 0.70 3.33 0.69 3.41 0.70 Determinants of sexual desire Female 3.56 0.51 3.27 0.83 3.45 0.67 Total 3.54 0.60 3.30 0.76 3.43 0.68 Male 3.28 0.61 3.21 0.67 3.24 0.64 Factors Factors Determinants of sexual expression Female 3.14 0.60 3.05 0.67 3.11 0.63 Total 3.20 0.61 3.13 0.68 3.17 0.64 Male 2.40 0.77 2.36 0.66 2.38 0.71 Barriers to intimacy Female 2.58 0.65 2.32 0.71 2.47 0.69 Total 2.50 0.71 2.34 0.68 2.43 0.70 Male 2.53 0.69 2.54 0.72 2.54 0.70 Barriers to sexual expression Female 2.50 0.58 2.28 0.66 2.41 0.62 Total 2.51 0.63 2.42 0.70 2.47 0.66

153

higher on the PC subscale than men. A significant effect was also found between age cohorts for the DSD subscale (see Table 6.2– Factors subset), with individuals in midlife scoring higher on the DSD subscale than those in later life. No significant effect for age cohort membership was found for other subscales that measure factors associated with the experience of sexuality, nor were any significant effects found for biological sex on all other subscales measuring these associated factors. Table 6.3 (Factors subset) identifies areas of similarity and difference between groups based on sex and age cohort membership. Table 6.4 (Factors subset) shows mean scores and standard deviations for each of the OASIS subscales across groups based on sex and age cohort membership. 6.6 Discussion The current study shows that men and women in mid and later life experience their sexuality more similarly than previously thought. Results related to the experience of sexuality for men and women in mid and later life will be discussed first, followed by a discussion of the factors associated with sexuality for these cohorts. The discussion will then examine these results using the new approach to understanding sexuality for heterosexual men and women in mid and later life developed in a prior study (Macleod et al., in press). 6.6.1 Understanding the experience of sexuality and ageing 6.6.1.1 Biological sex and age differences. In line with the original hypotheses, both sexual expressions and sexual urges were more relevant to the experience of sexuality for men than they were for women. This is supported by research showing that men are more likely than women to consider sexual intercourse and the frequency of sexual intercourse as being personally important (Arias-Castillo et al., 2009; O. Kim & Jeon, 2013; Müller et al., 2014), and that men report experiencing sexual desire and/or interest more often than women (Arias-Castillo et al., 2009; Johnson, 1996; Knodel & Chayovan, 2001). Sexual urges were also found to be more relevant to the experience of sexuality for individuals in midlife than they were for those in later life. This result corresponds with research indicating that those in midlife report a higher amount of sexual desire and/or interest than do those in later life (Carvalho & Nobre, 2010; Dello Buono et al., 1998; Johnson, 1996; Knodel & Chayovan, 2001; Padoani et al., 2000; Papaharitou et al., 2008; Trudel, 2002). One explanation for this result is that a change in focus from specific sexual urges to a more generalised desire for physical

154

contact and intimacy may occur as individuals age. This shift in focus may be related to an increase in the incidence of sexual dysfunction as individuals age (Corona et al., 2013; DeLamater, 2012; Lammerink et al., 2017). However, research examining optimal sexual experiences suggest that adjustments to sexual preferences are made by older adults to focus on activities that are less impacted by sexual dysfunction than more traditionally defined sexual urges such as the desire for penetrative intercourse (Ménard et al., 2014). As such, the level of relevance that is placed on sexual urges to an individual’s experience of sexuality may depend on whether they consider sexual urges to be specifically related to sexual activities such as sexual intercourse, or whether these urges are perceived to include the urge for more affectionate behaviours. 6.6.1.2 Similarities across sexes and age cohorts. No differences between men and women were found in terms of the relevance of aspects of intimacy and pleasure on their current experience of sexuality. This result corresponds with existing research that suggests that as they age, men place increasing importance on intimacy within a relationship (Moore, 2010; Sandberg, 2013). The similarity in the relevance of aspects of intimacy and pleasure on the experience of sexuality for men and women is present despite noted differences in the relevance of sexual expression and sexual urges to the experience of sexuality for these cohorts. This result suggests that an increase in the relevance of aspects of intimacy and pleasure for men does not necessarily indicate that this increased importance negates the relevance of sexual expressions or sexual urges on the experience of sexuality for men. As such, the increased relevance of intimacy and pleasure on men’s experience of sexuality may indicate that their experience of sexuality does not necessarily shift focus, so much as it becomes more complex over time. No differences between individuals in midlife and later life were found in terms of the relevance of aspects of intimacy and pleasure or sexual expression on current experiences of sexuality. These results contradict the original hypotheses, in which those in later life were expected to consider aspects of intimacy and pleasure more relevant to their current experience of sexuality and consider sexual expression to be less relevant to their current experience of sexuality than their midlife counterparts. These results also conflict with research showing that intercourse frequency decreases with age (Connaughton & McCabe, 2015) and that that older

155

people place a higher priority on emotional intimacy than sexual intimacy (Hurd Clarke, 2006; Lodge & Umberson, 2012; Sandberg, 2013). However, sexual expressions are not limited to penile-vaginal intercourse, and the similarity across age cohorts in terms of the relevance of sexual expressions to the experience of sexuality may reflect changes in how individuals define sexual expression as they age. Individuals across mid and later life cohorts did not differ in terms of the relevance of aspects of intimacy and pleasure or sexual expression. These results support research showing that differences in the experience of sexuality are present across age groups (Sandberg, 2013) and that individuals who adjust to the presence of sexual dysfunction within their sexual relationships find new ways to achieve sexual gratification that do not rely on penile-vaginal intercourse (Ménard et al., 2014). Further, research also shows that mutually-supportive partnerships help limit the negative impact of barriers to sexual expression on relationships (Hinchliff & Gott, 2004). Thus, the presence of mutually supportive partnerships may play a role in how individuals choose to express themselves sexually. 6.6.2 Factors associated with sexuality and ageing 6.6.2.1 Biological sex and age differences. Partner compatibility was considered more relevant to the experience of sexuality for women than it was for men. This result aligns with research that shows that feeling special or valued by their partner is an important part of the experience of sexual desire for women (Birnbaum et al., 2016). The greater relevance of partner compatibility to women as compared to men indicates that despite an increased relevance of partner compatibility to the experience of sexuality for men as they age (Sandberg, 2013), women still consider partner compatibility more relevant to their sexuality than men do overall. Factors associated with the experience of sexual desire were considered more relevant for those in midlife than for those in later life. The greater relevance of factors associated with the experience of sexual desire for individuals in midlife may be at least partially explained by the higher likelihood of competing priorities interfering with the ability for those in midlife to express their sexuality (Beckman et al., 2014; Penhollow et al., 2009). For example, individuals in midlife are more likely to feel that work and family commitments influence the amount of time and energy they have for expressing themselves sexually, as compared to their later life

156

counterparts who are more likely to be retired and have adult children that live outside the family home. 6.6.2.2 Similarities across sexes and age cohorts. No significant difference was found between men and women in terms of factors associated with intimacy and pleasure or the relevance that barriers to intimacy or barriers to sexual expression had to their experience of sexuality. Research indicates that the sex lives of older adults are strongly influenced by their partners, regardless of whether they are male or female (Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; Ménard et al., 2014; Sandberg, 2013). This lack of difference in the relevance of factors associated with intimacy and pleasure, barriers to intimacy and barriers to sexual expression for men and women may therefore be a reflection of the ability of individuals to adjust to changing sexual function (Ménard et al., 2014). Alternatively, the lack of difference between cohorts on the barriers to sexual expression and barriers to intimacy subscales may indicate the presence of mutually-supportive relationships within the current sample (Hinchliff & Gott, 2004), in that they help limit the negative impact of barriers to sexual experiences. No significant differences were found between those in midlife and those in later life in terms of the relevance of barriers to intimacy or barriers to sexual expression on their experience of sexuality. Most research examining differences between age cohorts in terms of barriers to sexual expression focus on differences in sexual functioning between cohorts. The correlation between age and erectile dysfunction is well established in the literature (Bortz et al., 1999), and midlife cohorts are more concerned with sexual dysfunction than their older counterparts (Lodge & Umberson, 2012). However, the results of the present study suggest that a decrease in sexual functioning may be less relevant to how individuals experience their sexuality across mid and later life than previously assumed. This decreased emphasis on sexual functioning may be driven by a change in focus from quantity to quality in terms of sexual expression (Deacon et al., 1995; Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; McCarthy et al., 2013). Such a change may mean individuals focus on positive aspects of their sexuality to a greater extent than reductions to their sexual functioning. Further, barriers to sexual expression extend beyond the presence of sexual dysfunction (Foley, 2015), and barriers related to sexual expression and sexual dysfunction may differ in terms of their relevance to individuals across age cohorts. This difference in relevance may, therefore, mask

157

potential differences in the relevance of sexual functioning to adults in mid and later life. 6.6.3 Limitations The current study examined an exclusively heterosexual sample, and additional research is required that looks at the similarities and differences in how LGBTQIA+ populations in mid and later life perceive their sexuality. Convenience sampling was used in this study, resulting in a healthy, largely Australian sample with few diagnoses of sexual dysfunction. As such, there may be unexplored differences in terms of how those with diverse cultural backgrounds or various levels of health and/or sexual functioning perceive their sexuality. Additionally, only adults aged 45 years and above were included, so no comments can be made regarding potential similarities and differences in how adults in mid and later life perceive their sexuality as compared with younger cohorts. Finally, the small sample size used in this study places constraints on the generalisability of the obtained results. Ideally, these results would benefit from further investigation using a larger sample. 6.7 Conclusions and future directions The experience of sexuality for heterosexual men and women in mid and later life is more similar than previously thought and raises questions about assumptions made using a more restricted approach to how the different components of sexuality across mid and later life are understood and measured. However, further research is needed that explores similarities and differences between men and women in mid and later life using more culturally diverse populations, and across a broader range of sexual orientations. Additional research is also warranted that looks more closely at similarities and differences between clinical and non-clinical samples with regards to sexual functioning.

158

Chapter 7. General discussion

159

7.1 General discussion The overarching aim of this thesis was to develop a novel approach to understanding and measuring sexuality for community-dwelling heterosexual adults in mid and later life. This overall aim was operationalised into three sub-aims: 1. Develop a new way of understanding sexuality in mid and later life that reflects the lived experience of sexuality for community-dwelling heterosexual adults in mid and later life; 2. Operationalise the experience of sexuality and the factors associated with these experiences identified in (1) to produce a valid and reliable measurement scale to evaluate these topics; and 3. Utilise the new approach developed in (1) and the associated measurement scale developed in (2) to examine differences for men and women in mid and later life. At the beginning of this thesis, two systematic reviews examined how sexuality is currently conceptualised for older adults and investigated the scope and availability of valid, reliable sexuality measures designed for older populations to highlight the gaps present in existing research that the first aim was designed to address (Chapters 2 and 3; Macleod & McCabe, accepted, 2019). These reviews provided justification for reconstructing how sexuality in mid and later life is conceptualised and measured to reflect the experience of sexuality more accurately for older adults than what is captured within existing approaches. In the first of three studies, group concept mapping was used to examine the experience of sexuality and the factors associated with these experiences for heterosexual adults aged 45 years and over (Chapter 4; Macleod et al., in press). This approach provided a quantitative mechanism for identifying major sub-constructs within the overall experience of sexuality for heterosexual men and women in mid and later life. Thematic analysis then provided further insight into how this cohort perceived their current experience of sexuality to differ from their experience of sexuality when younger. At the conclusion of this study, a new approach to understanding sexuality for heterosexual men and women aged 45 years and over was developed. This new approach expanded existing conceptualisations of sexuality to include a range of topics identified by participants as relevant to their current experience of sexuality that have not been adequately reflected in existing approaches.

160

In the second study, cognitive interviewing and Rasch modelling were used to operationalise the new approach for understanding sexuality in mid and later life generated in the first study. The psychometric properties of the resultant scale (the OASIS) were then assessed, demonstrating the reliability and validity of the OASIS and its’ subscales as a measure of sexuality for heterosexual men and women in mid and later life (Chapter 5; Macleod & McCabe, in preparation [1]). The third and final study investigated whether significant differences in the experience of sexuality and the factors associated with these experiences were present between men and women, or between those in mid and later life (Chapter 6; Macleod & McCabe, in preparation [2]). The first of two 2x2 Factorial MANOVAs investigated similarities and differences in the experience of sexuality across aspects of intimacy and pleasure, sexual expression, and sexual urges for men and women, and for individuals in mid and later life. The second MANOVA examined similarities and differences between men and women, and between individuals in mid and later life for factors associated with the experience of sexuality including partner compatibility, factors associated with intimacy and pleasure, factors that influence sexual desire and sexual expression, and barriers to intimacy and sexual expression. At the conclusion of this study, cohorts were shown to be more similar than previously thought, although some differences between cohorts were present in terms of their experiences of sexual expression, sexual urges, partner compatibility, and determinants of sexual desire. The next section outlines the structure used in the current chapter to discuss these results more broadly. 7.2 Chapter guide The current chapter integrates the findings from the two systematic reviews and three studies completed in this thesis, and explores their relevance to the field of sexuality and ageing. The chapter will first discuss the newly developed approach to understanding the experience of sexuality for heterosexual men and women in mid and later life and the factors related to these experiences (Section 7.3). This is followed by a discussion of the development of the OASIS, and the contribution its development makes to the field of sexuality and ageing (Section 7.4). An exploration of similarities and differences between heterosexual men and women in mid and later life is then presented (Section 7.5), followed by an interpretation of the results within Levinson’s normative-crisis model of lifespan development (Section 7.6). This is followed by a discussion of the broader implications of the work conducted in this

161

thesis (Section 7.7). Lastly, limitations and recommendations for future research are discussed (Sections 7.8 and 7.9) before conclusions are presented (Section 7.10). 7.3 A novel approach to understanding sexuality for heterosexual men and women in mid and later life In Chapters 1-3, several gaps were identified between how older adults describe their experience of sexuality and the way that it is conceptualised and measured in the broader sexuality and ageing literature. Consequently, to address the overall aim of the thesis, additional research was needed that investigated how heterosexual men and women in mid and later life describe their current experience of sexuality and the factors related to these experiences. This process involved the application of group concept mapping and thematic analysis to develop an innovative approach to understanding sexuality across the later life stages that more accurately reflected the perspectives of heterosexual men and women in mid and later life than those reflected in existing approaches. At the conclusion of this process, a new approach was developed that can accommodate changes in sexual priorities associated with unique life events while still being representative of the ‘general’ experiences of heterosexual men and women in mid and later life, thereby addressing the first aim of the thesis and contributing to the overall aim of the thesis. Thus, the new approach challenges the longstanding research focus on sexual frequency and sexual functioning for older cohorts and simultaneously supports research that indicates that as people age, the quality of their sexual experiences becomes more important than the frequency with which these experiences occur (Deacon et al., 1995; Fileborn, Thorpe, Hawkes, Minichiello, & Pitts, 2015; McCarthy et al., 2013). Many of the topics reflected in existing conceptualisations of sexuality are present in the new approach, including sexual urges, sexual expression, and various physiological factors associated with the ability to engage in sexual activities. However, new and important information about how these topics interconnect and the relative importance of these topics for individuals in mid and later life have also been identified. In doing so, this new approach broadens our understanding of the experience of sexuality for mid and later life in a way that does not rely on our understanding of sexuality in early adulthood, and that provides a framework for investigating areas of similarity and difference in the experience of sexuality for men and women in mid and later life that reflect the lived experience of sexuality for these cohorts.

162

Despite the presence of many topics already present in existing approaches, several significant differences were identified in how these topics were being conceptualised by men and women in mid and later life as compared to existing conceptualisations of sexuality in mid and later life. Specifically, the current research provides evidence that discussions of barriers to sexuality would better reflect the experiences of older cohorts if they were split into two sub-categories: barriers to intimacy and barriers to sexual expression. This split provides clearer acknowledgement that sexual problems are not only a product of sexual dysfunction or other medical impediments, and that relationship problems can also restrict how older adults engage with their sexuality. As such, this alternative interpretation provides a more effective mechanism for recognising sexual problems beyond those that are considered medical problems that require chemical intervention to achieve positive sexual outcomes (Bradway & Beard, 2015; Gott, 2004, 2006; Phillipson, 2013). A primary benefit of the new approach developed in this thesis over existing approaches is the removal of sexual functioning as the centre-point of the experience of sexuality for those in mid and later life. Instead, no single area is emphasised over others, allowing interpersonal aspects to be recognised as important components to the experience of sexuality in mid and later life, while still recognising that sexual functioning and sexual expression continue to be part of the broader experience of sexuality throughout the lifespan. Further, this new approach provides an interpretation of sexuality that is compatible with evidence that the presence of sexual dysfunction is not always considered an impediment to a satisfying sex life (Ménard et al., 2014). By developing this new approach using the actual perspectives of adults in mid and later life, one of the major hurdles to understanding sexuality in the later stages of life has been removed; namely, the assumption that the experience of sexuality in early adulthood is the gold standard with which to compare the experience of sexuality across all adult life stages (Bouman & Kleinplatz, 2014; Sanders et al., 2013). The benefits of such a shift away from this assumption are further supported by the results of the thematic analysis in the first study that indicate most adults in mid and later life feel that their experience of sexuality has changed over time (Macleod et al., in press). Thus, the development of an evidence-based approach to understanding the lived experience of sexuality for older adults makes it

163

possible to adjust the focus of research and clinical practice onto topics that older cohorts have explicitly identified as relevant and important to their actual experiences of sexuality across mid and later life. This adjustment is further facilitated by the operationalisation of the new approach with the development of the OASIS. 7.4 Development of the OASIS The OASIS was developed as a way to operationalise the experience of sexuality and the factors associated with these experiences as identified by heterosexual adults aged 45 years and over (Chapter 4; Macleod et al., in press). In doing so, it addresses the second aim of the thesis and contributes to the overall aim of the thesis. This process involved using best practice scale development guidelines (Boateng et al., 2018) to produce a valid and reliable measurement scale that quantifies the lived experience of sexuality for community-dwelling heterosexual adults in mid and later life. Draft items were developed using statements provided by heterosexual men and women aged 45 years and over that described their experience of sexuality, intimacy, and desire from the first study (Chapter 4; Macleod et al., in press). These items were then refined using cognitive interviewing before Rasch modelling was used to identify the most psychometrically robust items for inclusion in the final version of the scale. This process resulted in the development of a sexuality measurement scale containing 78-items distributed across nine subscales that captures the experience of sexuality and the factors associated with these experiences for heterosexual men and women aged 45 years and above (Macleod & McCabe, in preparation [1]). Confirmatory factor analysis supported the use of a nine-subscale structure for the OASIS, and psychometric evaluations showed that the OASIS has good internal consistency (α between .74 and .93) and good to excellent test-retest reliability (ICC between .85 and .94). Moderate to strong correlations (i.e., Pearson's r = 0.3 - 0.5; Heale & Twycross, 2015) between individual OASIS subscales with other relevant measurement scales such as the PAIR inventory (Schaefer & Olson, 1981), SASS (E. Weinstein, 2013), MoSEIC (Worthington et al., 2008), and ASCS (Cowden & Koch, 1995) then demonstrated that the OASIS is a reasonable alternative to sexuality measures currently used in later life sexuality research. Previously, collecting data to examine sexuality in mid and later life involved the use of unvalidated survey items as well as cherry-picked subscales from other sexuality scales, many of which have not

164

been validated for use with older populations. Thus, the development of the OASIS provides researchers with a more efficient and reliable measurement tool to collect data on a broad range of topics related to the experience of sexuality for individuals in mid and later life than has previously been available. The development of the OASIS provides researchers and clinicians with a robust, psychometrically sound sexuality measurement scale that has been designed both for and with the target cohort using an approach that aligns with best practice guidelines for scale development (Boateng et al., 2018). The development of the OASIS creates new opportunities to collect data on a more diverse range of topics related to the experience of sexuality for individuals in mid and later life than has previously been available. In particular, the nine-subscale structure of the OASIS provides researchers and clinicians with the flexibility to examine specific sub- constructs of the experience of sexuality for heterosexual men and women in mid and later life or to conduct a more extensive examination of sexuality for these cohorts overall. In addition to the flexibility provided by the creation of subscales that can be used as stand-alone measures, the delineation between subscales that measure the experience of sexuality (intimacy and pleasure – aspects, sexual expression, and sexual urges subscales) and those that measure the factors associated with the experience of sexuality (intimacy and pleasure – factors, partner compatibility, determinants of sexual desire, determinants of sexual expression, barriers to intimacy, and barriers to sexual expression subscales) also creates unique opportunities to examine individual topics related to the experience of sexuality, or the factors associated with these experiences, including areas of similarity and difference between cohorts across multiple variables. 7.5 Areas of similarity and difference in the experience of sexuality and associated factors across the later life stages After developing and then operationalising a new approach to understanding sexuality across mid and later life in Chapters 4 and 5, data collected with the OASIS were analysed using two 2x2 Factorial MANOVAs to address the third thesis aim. These analyses examined whether any significant differences were present in terms of how men and women in mid and later life experienced sexuality across their separate cohorts based on biological sex (male/female) and age group membership (midlife/later life). Descriptions of areas of similarity and differences between cohorts

165

are provided in the following sub-sections. Similarities and differences between men and women are discussed first, before discussing areas of similarity and difference between those in mid and later life. 7.5.1 Areas of similarity and difference between men and women Overall, few significant differences were identified between men and women in terms of their reported experience of sexuality or the factors associated with these experiences. Men and women scored similarly in terms of the relevance that aspects of intimacy and pleasure, factors related to intimacy and pleasure, determinants of sexual desire, determinants of sexual expression, barriers to intimacy, and barriers to sexual expression had on their current experience of sexuality (Macleod & McCabe, in preparation [2]). Previous research has failed to agree whether an individual’s sex is related to how important they feel that intimacy and affection are to their experience of sexuality in mid and later life. For example, Johnson (1996) identified a greater importance of intimacy for older women as compared to men, but other research suggests that affection and love are of greater importance to older men than older women (Iveniuk et al., 2016; Moore, 2010; Müller et al., 2014). The results of this thesis suggest that these inconsistent results (i.e., Johnson, 1996; Iveniuk et al., 2016; Moore, 2010; Müller et al., 2014) may reflect differences in how specific terms such as intimacy and affection are interpreted, rather than genuine differences in the importance of intimacy and affection between older men and women. The absence of significant differences across six of the nine OASIS subscales aligns with research that shows that men place increasing importance on intimacy within a relationship as they age (Moore, 2010; Sandberg, 2013). The increased relevance of intimacy for men, however, does not come at the cost of decreasing importance in other areas of sexuality. As such, the experience of sexuality for older men does not necessarily change, per se, but becomes increasingly complex over time. Similarities between men and women across these domains may also reflect an equal ability for both men and women to adjust their sexual priorities to accommodate changing sexual functioning (whether their own or that of their partner) as a result of age (Ménard et al., 2014). Research has shown that the sex lives of both older men and older women are directly influenced by their partners (Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; Ménard et al., 2014; Sandberg, 2013). Further, the presence of mutually-supportive relationships has been shown to decrease the

166

negative impact of barriers to sexuality (Hinchliff & Gott, 2004), which may explain why no significant differences were present between men and women on subscales measuring barriers to sexuality. The absence of significant differences between men and women across most OASIS subscale constructs may also indicate that sex differences play a lesser role in the experience of sexuality for individuals in the later stages of life than other external factors, at least insofar as they influence intimacy and pleasure, determinants of sexual desire and sexual expression, or barriers to intimacy and sexual expression. For example, both men and women in the group concept mapping study indicated that the moral and privacy standards imposed on them by society influenced how and when they felt they were able to express themselves sexually (Chapter 4; Macleod et al., in press). This suggests that societal attitudes portraying the sexuality of older adults as non-existent or inappropriate (Bouman & Kleinplatz, 2014; Taylor & Gosney, 2011) may play an important role in how older adults choose to engage with their sexuality in a way that reduces the influence that gender or biological sex has on these experiences. Differences in scores on the OASIS were found between men and women on subscales measuring sexual expression, sexual urges, and partner compatibility. These results confirm research that shows men consider sexual expression and sexual urges more relevant to the experience of sexuality than women (Arias-Castillo et al., 2009; Johnson, 1996; O. Kim & Jeon, 2013; Knodel & Chayovan, 2001; Müller et al., 2014) and that women consider partner compatibility to be more relevant to the experience of sexuality than men (Birnbaum et al., 2016). However, individuals in the group concept mapping study did not explicitly identify gender roles as something that they were conscious of playing an integral part in their experience of sexuality (Chapter 4; Macleod et al., in press). This could mean that societal attitudes that rely on stereotypes that portray sexual differences related to gender roles as ‘innate’ biological differences may prevent older adults from recognising a gender influence on their experiences, even when one is present. For example, the widely popular book “Men are from Mars; Women are from Venus” (Gray, 1992) promoted the idea that men are driven by their biological urges to seek out sex and women are inherently nurturing and seek out an emotional connection. As a result, this may have shaped societal attitudes in such a way that older adults have been socialised to believe that sexuality is innately different for men

167

and women, and that these differences are biological in nature (Gray, 1992, 2011). As a consequence, these misguided attitudes fail to acknowledge that not all of these ‘sex differences’ are associated with biological sex, and that some experiences are actually driven by social constructs related to gender roles (Carothers & Reis, 2013; Petersen & Hyde, 2011). 7.5.2 Areas of similarity and difference between individuals in mid and later life As with cohort differences based on sex, few significant differences were identified between those in mid and later life in terms of their reported experience of sexuality or the factors associated with these experiences. Areas of similarity between mid and later life cohorts were found for partner compatibility, sexual expressions and their related determinants, the experience of intimacy and pleasure and the factors associated with these experiences, and barriers to intimacy and sexual expression. These results appear to contradict research that shows intercourse frequency declines with age (Connaughton & McCabe, 2015). However, the absence of significant differences between those in mid and later life for sexual expressions are more likely to reflect an expansion of the term sexual activity to include sexual activities beyond that of penile-vaginal intercourse, rather than a genuine contradiction of the well-documented decrease in sexual intercourse frequency as people age (Addis et al. 2006; Bortz et al., 1999; Corona et al., 2013). The absence of significant differences between mid and later life cohorts for domains that reflect intimacy, affection, and partner compatibility likewise contradicts research that shows that older adults value emotional intimacy over sexual intimacy (Hurd Clarke, 2006; Lodge & Umberson, 2012; Sandberg, 2013). Again, this may be more reflective of differences in how terms are being defined by those in mid and later life, with elements of both emotional intimacy and sexual intimacy reflected across the subscales that measure partner compatibility, and intimacy and pleasure. The similarities between midlife and later life cohorts in terms of the relevance of barriers to intimacy and sexual expression suggests that those in mid and later life are equally concerned with impediments to their sexuality and sexual expression. If so, this result would indicate that older adults are less concerned with decreased sexual functioning than previously thought, or that these concerns begin much earlier than previously assumed.

168

Differences in scores on the OASIS between those in midlife and those in later life were present only on subscales measuring sexual urges and determinants of sexual desire. This result indicates that individuals in midlife consider sexual urges and determinants of sexual desire to be more relevant to their experience of sexuality than individuals in later life. These cohort differences may relate to two previously identified ideas: that rates of sexual dysfunction increase as individuals age (Beckman et al., 2008; Bortz et al., 1999), and that the presence of sexual dysfunction within relationships can result in adjustments to how individuals conceptualise sexuality and sexual activity (e.g., Gott, 2006; Ménard et al., 2014). In other words, adjusting to the presence of sexual dysfunction (or a reduced level of sexual functioning) that can occur with increasing age may also involve adjusting one’s sexual priorities to be simultaneously less reflective of the physiological experience of sexuality and more reflective of the components of sexuality that relate to emotionally connecting with a sexual partner. Thus, determinants of sexual desire and sexual urges remain an important component of sexuality in later life generally but become secondary to other interpersonal measures in the overall experience of sexuality. Within a medical model of sexuality, changes in the experience of sexuality are primarily attributed to age-related changes such as as hormonal changes or age- related health problems for an individual or their partner (e.g., Beckman et al., 2014; DeLamater et al., 2008; Hyde et al., 2010; Padoani et al., 2000). However, results from the thematic analysis in the first study indicate that participants attributed changes in the experience of sexuality over time to changes in perspective, changes to relationship dynamics, the impact of environmental changes, and behavioural changes that occur as a result of changing perspectives and priorities. These changes were described by individuals across both mid and later life, and align with research that attributes the ability to engage in ‘optimal’ sex in later life with overcoming learned negative attitudes about sexuality and ageing, embracing change, and seeking out opportunities for personal growth independently and with a long-term partner (Ménard et al., 2014). Thus, the new approach developed in this thesis expands the focus of the current medical model to include the diverse range of continuing developmental processes that influence the experience of sexuality throughout the life course.

169

7.5.3 Summary of the experience of sexuality for community-dwelling heterosexual adults in mid and later life Overall, the data indicated that there are more similarities than differences between cohorts, both in terms of sex and age group membership, and that interpersonal dynamics were the main focus of sexuality for both men and women in mid and later life. The emphasis on interpersonal dynamics over sexual urges and sexual expression supports research that suggests the experience of sexuality for older adults is not adequately reflected in current research (Deacon et al., 1995; Fileborn, Thorpe, Hawkes, Minichiello, Pitts, et al., 2015; McCarthy et al., 2013). Further, these results question the relevance of using an approach that frames the experience of sexuality for younger cohorts as the gold standard used to measure sexuality throughout the adult lifespan (Sanders et al., 2013). As such, these results indicate that the current focus in sexuality and ageing research on sexual activity frequency and sexual functioning may be specious, and a significant shift in focus away from the ‘youthful’ gold standard is required for research to align with the sexual experiences of older adults. The following section outlines how these results fit within the normative-crisis model of lifespan development. 7.6 Theoretical alignment of results to the normative-crisis model Within a normative-crisis model of lifespan development, the experience of sexuality for older individuals is expected to be different from the experience of sexuality for younger individuals. These differences are proposed to occur as a result of developmental differences across the separate stages of adulthood. Specifically, Levinson’s normative-crisis model allows the experience of sexuality for adults to be compartmentalised into three distinct life stages: early adulthood, middle adulthood, and later adulthood (Hoffnung et al., 2016). This split allows for additional nuances to be examined across adulthood that align with significant life events typically associated with the different life stages. For example, early adulthood is often a time of personal discovery and development and frequently involves increased sexual interest and the development of intimate relationships (Zimmer-Gembeck & Petherick, 2006). Middle adulthood often includes the formation of long-term romantic relationships and lifestyle adjustments related to work and family commitments (Lachman, 2004). Late adulthood often sees individuals reducing the number of external commitments, including children moving out of the family home, or retiring from full-time employment (J. E. Kim & Moen,

170

2002). Thus, applying Levinson’s ‘seasons of adult life’ developmental approach to how adult sexuality is understood makes it possible to acknowledge and account for the continued development that occurs as a result of various life experiences across adulthood. Under Levinson’s normative-crisis model of lifespan development, the different stages of adulthood are influenced by significant relationships present within each life stage (Hoffnung et al., 2016). In line with this, research has shown that as people age, they become more selective of the types of relationships they maintain, and that older adults prioritise relationships that involve close bonds over other, less intimate connections (Antonucci, Akiyama, & Takahashi, 2004; Carstensen, 1992). The noted increased relevance of the interpersonal components of sexuality in this thesis may, therefore, be a reflection of the increasing focus on relationships that individuals find emotionally rewarding more generally as they age (Carstensen, Fung, & Charles, 2003; Carstensen, Gross, & Fung, 1998; Lindley, Harper, & Sellen, 2008). As such, an increased focus on the interpersonal components of sexuality by adults in mid and later life has direct implications for how health professionals engage with older adults in terms of their sexuality and sexual health needs. These points are explained in the next section. 7.7 Implications The novel approach outlined in this thesis is a direct reflection of the way that heterosexual men and women in mid and later life have described their experience of sexuality and the factors related to these experiences, including reflections of the impact that physiological changes that occur as a result of age have on their experience of sexuality. As such, it provides a framework for understanding sexuality that incorporates the importance of partner compatibility, intimacy, and affection in how heterosexual men and women in mid and later life experience their sexuality while still acknowledging the relevance of sexual urges and sexual expression to these experiences. The OASIS operationalises this new understanding to provide researchers and clinicians with a psychometrically robust measurement scale that reflects the everyday experience of sexuality for adults in mid and later life without assuming that this cohort seeks to maintain or return to the sexual experiences of their younger years. The OASIS can, therefore, facilitate the development of a comprehensive picture of the ‘average’ experience of sexuality for older adults across numerous domains.

171

Re-framing the experience of sexuality in mid and later life to emphasise the importance of interpersonal dynamics challenges the use of sexual activity frequency as the ‘gold standard’ measure of sexuality for these cohorts. Indeed, these results also raise questions about whether frameworks for understanding sexuality that find their roots in measures of sexual activity frequency accurately reflect the broader experience of sexuality for adults of any age. However, adjusting the way that sexuality in mid and later life is understood (let alone how sexuality is conceptualised throughout the lifespan) will require a significant paradigm shift away from the medical model that frames changes in sexual function associated with the ageing process as inherently negative and requiring medical intervention (Bradway & Beard, 2015; Malatesta, 2007). Under the current medical paradigm, research has primarily focused on measures of sexual functioning, sexual activity frequency, and their association with physiological changes that occur as part of the ageing process. While the physiological ageing process has the potential to impact on an individual’s ability to engage in activities like sexual intercourse (among others), the overall experience of sexuality encompasses more than the ability to physically participate in a range of sexual behaviours (Bouman & Kleinplatz, 2014; DeLamater, 2012; Hinchliff & Gott, 2004; Hurd Clarke, 2006; Ménard et al., 2014). By measuring the current experience of older adults, changes to the experience of sexuality that occur due to the ageing process are accounted for in how an individual responds to items on the OASIS. Thus, the OASIS provides researchers and health professionals with a sexuality measurement scale that reflects the general experience of sexuality and the factors related to these experiences for those in mid and later life that can also accommodate the heterogeneity present within older populations in terms of their functional age. The current thesis delivers a clear message for health professionals and researchers to be mindful of the potential for changed perceptions of sexuality in clients aged 45 years and over. These changed perceptions can now be quantified using the OASIS, thereby creating new opportunities for collecting information about how relevant an individual considers each of the sub-constructs reflected in the nine subscales to their own experience of sexuality. The focus on the current experience of sexuality for respondents also presents opportunities to accommodate changing perspectives over time as a result of developmental changes through the life stages. This information can then help clinicians tailor treatment plans to the individual needs

172

of each client. Further, the OASIS can also be used to examine whether therapeutic outcomes include changes to the sexual priorities of older adults. Thus, using the OASIS in a clinical setting can help ensure interventions focus on areas that matter to an individual, limiting the amount of time spent on topics that do not hold much relevance to an individual’s current experience of sexuality and increasing the likelihood of achieving positive outcomes for clients. 7.8 Limitations The absence of a consistent conceptualisation for sexuality in mid and later life meant that some subjective interpretation was needed to determine the suitability of relevant papers to include in the two systematic reviews (Chapters 2 and 3), and to determine how topics investigated in each study aligned with the WHO definition of sexuality. While every effort was made to ensure that these interpretations were as objective as possible, other explanations may also be relevant. The decision to focus exclusively on heterosexual adults aged 45 years and above also meant that some papers were excluded from the two systematic reviews. The decision to apply stringent age cut-offs was made to ensure that the reviews were an accurate reflection of the experiences of adults in mid and later life and were not influenced by the experiences of younger cohorts. Likewise, the decision to focus exclusively on heterosexual adults was made in deference to several noted differences in the experience of sexuality for LGBTQIA+ groups (Berger, 1982; Eliason & Schope, 2001). These decisions mean that comments cannot be made regarding the direct relevance of this work to LGBTQIA+ populations of any age, or to heterosexual adults in early adulthood without further research to determine if the new approach and measurement scale are also relevant to these populations. Both time and funding constraints prohibited the use of probability sampling, and as such, convenience sampling was used across all three studies. The use of convenience sampling and the possibility of self-selection bias in participants may have implications for the generalisability of this research. Given the limitations associated with convenience sampling, efforts were made to ensure the surveys were as non-threatening as possible to make participation more likely. Similarly, efforts to recruit similar numbers of participants in mid and later life, and men and women were also made. Despite these efforts, recruitment across all studies produced a predominantly Australian sample that was competent in reading and writing English. Recruitment occurred at five different time points across all studies. In the group

173

concept mapping study, recruitment occurred separately for phase one and phase two. During the development of the OASIS, recruitment occurred at three separate time points, and data collected during the third sub-study in Chapter 5 was also used in Chapter 6. Broadly, the study sample in Chapter 4 had a slightly larger proportion of heterosexual adults in later life (57% compared to 51% of the study sample), and substantially more diagnosed health conditions (80% compared to 33% of the study sample). The distribution of participants across relationship categories was notably similar (26%, 63%, and 11% for single, partnered and dating but not living together groups respectively as compared to 24%, 69%, and 6%). However, there was a slightly larger proportion of respondents that were dating but not living with a partner in the Chapter 4 sample (specifically, 11% as compared to 6%). No information is available to determine whether self-selection bias across the studies led to a subsection of the target population with similar ideas about sexuality throughout the lifespan taking part in the studies. These limitations highlight that there may be unexplored differences in how those with diverse cultural backgrounds, who are less proficient in English, or who speak languages other than English perceive their sexuality compared to the samples used in this research. Feedback received early on in the first study highlighted that participants felt more secure providing information about their age group (midlife/later life) instead of providing their year of birth. This meant that participant mean age could not be calculated for any study. Although recruitment efforts were extensive and included the promotion of research studies across a range of media, participation rates were lower than anticipated. Each study achieved the minimum acceptable sample size for the analysis type used within the study (Ho, 2013; Linacre, 1994; Peterson et al., 2017; Rosas & Kane, 2012; Trochim, 1989, 1989; Van Zile-Tamsen, 2017). Larger sample sizes may have provided a more robust result, allowing for greater confidence in the generalisability of findings. In line with this, smaller sample sizes than anticipated in Studies 1 and 2 meant that age cohort and/or sex differences were not explored in these studies, and potential cohort differences may have been masked. However, the results of Study 3 (Chapter 6) suggest that these differences, if present, are minimal. Thus, the sampling issues identified in Studies 1 and 2 (Chapters 4 and 5) are likely to have only a limited impact on the generalisability of results to the broader population.

174

Finally, each study used samples recruited from the general population and results did not include a separate analysis of research questions for individuals with clinically diagnosed medical conditions. As such, the results in this thesis should not be assumed to reflect the experiences and perspectives of clinical populations. In addition, the OASIS has been designed as an information-gathering tool for understanding the broad experience of sexuality for heterosexual men and women in mid and later life but has not been designed for use as a diagnostic tool. The OASIS can therefore provide valuable information about how an individual experiences sexuality within the context of their own life, helping clinicians tailor treatment plans to reflect the unique perspectives of the individual, but will not necessarily provide information that could contribute to the diagnosis of sexual problems. 7.9 Future research Despite the expanded understanding of sexuality for heterosexual men and women in mid and later life provided by the new approach and measurement scale developed in this thesis, it is crucial to recognise that there are many variables that remain unexplored. As noted in the limitations, further research is needed that examines the relevance of both the new approach to understanding the experience of sexuality in mid and later life and the OASIS across culturally and ethnically diverse populations, clinical populations, and across a broader range of sexual orientations. Future research could also explore whether using sexual activity frequency as the ‘gold standard’ measure of sexuality truly reflects the experience of sexuality for adults at any age. Research by Kleinplatz et al. (2009) has previously highlighted that the dominant themes in how ‘optimal sex’ is described by adults of all ages, backgrounds and orientations were almost universal. As such, a re-assessment of how sexuality is being conceptualised more broadly throughout the lifespan and without emphasis placed on sexual functioning or sexual activity frequency is warranted. Additional research investigating whether a person’s perceived experience of age influences how they understand and experience their sexuality may also be beneficial. The research within this thesis only examines the current experience of sexuality for individuals, without framing the discussion within the constructs of sexuality ‘during midlife’ or ‘during later life’. Thus, the results of this thesis may already reflect views of sexuality outside of the construct of ‘age’. However, it would be prudent to verify this supposition with further research that examines whether

175

differences are present between groups that do or do not identify as being within their chronologically-defined age cohort. Future research could also explore the relevance of the timing-of-events model of lifespan development in understanding the experience of sexuality across the life- course in place of the normative-crisis model. An advantage of the timing-of-events model over the normative-crisis model is its ability to account for non-normative changes that occur in an individual's life (Hoffnung et al., 2016); while the normative- crisis model can account for the average experience of adult development, individuals rarely conform to averages. In an increasingly diverse society, less rigidity is present in the 'expected age' of significant life events; people are marrying and having children later in life, midlife career changes are increasingly common, and divorce and re-marriage are no longer as taboo as they once were. As such, using a timing-of- events model to frame discussions about sexuality throughout the lifespan may provide greater flexibility for contextualising differences in the experience of sexuality as a result of different life experiences than the normative-crisis provides. The potential value of a sexuality measurement scale that reflects a broader range of topics than existing scales cannot be understated. The OASIS provides new opportunities to explore the experience of heterosexual men and women in mid and later life across a vast array of sexual variables. For example, further research exploring similarities and differences between cohorts based on a range of demographic profiles would provide valuable information about if and how sexuality differs across population groups. Using the OASIS in conjunction with existing measures of sexuality that focus on sexual behaviours would also help identify whether areas of relevance to an individual’s experience of sexuality are related to the frequency of various sexual behaviours, or to the level of sexual function possible for that individual. Doing so would provide additional information to help improve how we engage with older adults in terms of their sexuality and sexual health. 7.10 Conclusions The novel approach to understanding the experience of sexuality for heterosexual men and women in mid and later life presented in this thesis helps to bridge conceptual gaps identified within existing sexuality and ageing research. Specifically, the proposed conceptualisation provides a detailed description of sexuality for heterosexual cohorts in mid and later life that is reflective of the specific experience of sexuality and the factors associated with these experiences for these

176

individuals, as opposed to one that is based on the experiences and perspectives of younger adults. As such, this thesis presents an alternative approach to understanding sexuality for heterosexual men and women in mid and later life that does not over- emphasise sexual functioning or sexual activity frequency, nor does it portray the sexual ideals of those in early adulthood as universal across all adult life stages. Beyond the added benefit that the new approach to understanding the experience of sexuality for heterosexual men and women in mid and later life has for the broader field of sexuality and ageing, the development of the OASIS gives researchers and clinicians a new sexuality measurement scale that captures a wider range of topics than other existing later life sexuality measurement scales. Further, the OASIS provides a tool for quantifying the different topics captured in the new approach to understanding the experience of sexuality for heterosexual adults in mid and later life, thereby providing researchers and clinicians with a measurement tool that directly reflects the everyday experiences of these cohorts. Such a tool is invaluable if further progress is to be made in understanding the lived experiences of heterosexual men and women aged 45 years and over. This process has already begun with an exploration of differences between men and women in mid and later life based on sex and age group membership (Macleod & McCabe, in preparation [2]). Combined, the development of this new approach and the OASIS makes room for a significant paradigm shift in terms of how sexuality in mid and later life is understood so that the real-life experiences of older adults are better-reflected. In doing so, the sexual potential of older adults can be more effectively incorporated into perspectives of sexuality across all life stages, rather than framing the experience of sexuality in mid and later life as a continuous effort to maintain (or reproduce) the sexual experiences and sexual functioning of early adulthood. Such a shift will also mean that sexual decline is no longer framed as an inevitable outcome of the ageing process, thereby creating new opportunities to improve the way that we engage with adults in midlife and beyond in terms of their sexuality and sexual health needs. This, in turn, will help empower older adults to engage with their sexuality in their own way, and to define for themselves what it means to have a fulfilling sex life.

177

References Aboderin, I. (2014). Sexual and reproductive health and rights of older men and women: Addressing a policy blind spot. Reproductive Health Matters, 22(44), 185-190. doi: 10.1016/S0968-8080(14)44814-6 Addis, I. B., Van Den Eeden, S. K., Wassel-Fyr, C. L., Vittinghoff, E., Brown, J. S., Thom, D. H., & Reproductive Risk Factors for Incontinence Study at Kaiser (RRISK) Study Group. (2006). Sexual activity and function in middle-aged and older women. Obstetrics and Gynecology, 107(4), 755-764. doi: http://dx.doi.org/10.1097/01.AOG.0000202398.27428.e2 Allen, M. E. (1987). A holistic view of sexuality and the aged. Holistic nursing practice, 1(4), 76-83. doi: http://dx.doi.org/10.1097/00004650-198708000- 00011 Althof, S. E. (2010). What's new in sex therapy. Journal of Sexual Medicine, 7(1), 5- 13. Retrieved from: http://www.cmshsf.com/docs/Whats%20New%20JSM.pdf Antonucci, T., Akiyama, H., & Takahashi, K. (2004). Attachment and close relationships across the life span. Attachment and Human Development, 6(4), 353-370. doi: http://dx.doi.org/10.1080/1461673042000303136 Arias-Castillo, L., Ceballos-Osorio, J., Ochoa, J. J., & Reyes-Ortiz, C. A. (2009). Correlates of sexuality in men and women aged 52-90 years attending a university medical health service in Colombia. Journal of Sexual Medicine, 6(11), 3008-3018. doi: 10.1111/j.1743-6109.2009.01488.x Attride-Stirling, J. (2001). Thematic networks: An analytic tool for qualitative research. Qualitative Research, 1(3), 385-405. doi: http://dx.doi.org/10.1177/146879410100100307 Australian Bureau of Statistics. (2014a). Age Standard. (Cat. No. 1200.0.55.006, Version 1.7. ). Retrieved from http://www.abs.gov.au/ausstats/[email protected]/ Lookup/1200.0.55.006mainþfeatures32014,%20Version%201.7. Australian Bureau of Statistics. (2014b). Australia’s demographic challenges version 1.7. (Cat. No. 1200.0.55.006. ). Canberra (AU) Retrieved from http://demographics.treasury.gov.au/content/_download/australias_demograph ic_challenges/australias_demographic_challenges.pdf. Bach, L. E., Mortimer, J. A., VandeWeerd, C., & Corvin, J. (2013). The association of physical and mental health with sexual activity in older adults in a retirement community. Journal of Sexual Medicine, 10(11), 2671-2678. doi: 10.1111/jsm.12308 Bacon, C. G., Mittleman, M. A., Kawachi, I., Giovannucci, E., Glasser, D. B., & Rimm, E. B. (2003). Sexual function in men older than 50 years of age: Results from the health professionals follow-up study. Annals of Internal Medicine, 139(3), 161-168. doi: http://dx.doi.org/10.7326/0003-4819-139-3- 200308050-00005 Baldissera, V. D., Bueno, S. M., & Hoga, L. A. (2012). Improvement of older women's sexuality through emancipatory education. Health Care for Women International, 33(10), 956-972. doi: 10.1080/07399332.2012.684986 Bancroft, J. H. J. (2007). Sex and aging. The New England Journal of Medicine, 357(8). doi: http://dx.doi.org/10.1056/NEJMe078137 Barber, H. R. (1996). Sexuality and the art of arousal in the geriatric woman. Clinical Obstetrics and Gynecology, 39(4), 970-973. http://dx.doi.org/10.1097/00003081-199612000-00025

178

Bates-Jensen, B. M. (1989). Sexuality and the elderly. Journal of Enterostomal Therapy, 16, 158-163. Retrieved from https://insights.ovid.com/enterostomal- therapy/jent Bauer, M., Haesler, E., & Fetherstonhaugh, D. (2015). Let's talk about sex: Older people's views on the recognition of sexuality and sexual health in the healthcare setting. Health Expectations, 19(6), 1237-1250. doi: 10.1111/hex.12418 Bauer, M., McAuliffe, L., & Nay, R. (2007). Sexuality, health care and the older person: An overview of the literature. International Journal of Older People Nursing, 2(1), 63-68. doi: http://dx.doi.org/10.1111/j.1748-3743.2007.00051.x Beatty, P. C., & Willis, G. B. (2007). Research synthesis: The practice of cognitive interviewing. Public Opinion Quarterly, 71(2), 287-311. doi: 10.1093/poq/nfm006 Beckman, N., Waern, M., Gustafson, D., & Skoog, I. (2008). Secular trends in self reported sexual activity and satisfaction in Swedish 70 year olds: Cross sectional survey of four populations, 1971-2001. BMJ: British Medical Journal, 337, a279. doi: 10.1136/bmj.a279 Beckman, N., Waern, M., Ostling, S., Sundh, V., & Skoog, I. (2014). Determinants of sexual activity in four birth cohorts of Swedish 70-year-olds examined 1971- 2001. Journal of Sexual Medicine, 11(2), 401-410. doi: 10.1111/jsm.12381 Bell, S., Reissing, E. D., Henry, L. A., & VanZuylen, H. (2017). Sexual activity after 60: A systematic review of associated factors. Sexual Medicine Reviews, 5(1), 52-80. doi: http://dx.doi.org/10.1016/j.sxmr.2016.03.001 Benbow, S. M., & Beeston, D. (2012). Sexuality, aging, and dementia. International Psychogeriatrics, 24(7), 1026-1033. doi: http://dx.doi.org/10.1017/S1041610212000257 Berger, R. M. (1982). The unseen minority: Older gays and lesbians. Social Work, 27(3), 236-242. doi: http://dx.doi.org/10.1093/sw/27.3.236 Bergeron, C. D., Goltz, H. H., Szucs, L. E., Reyes, J. V., Wilson, K. L., Ory, M. G., & Smith, M. L. (2017). Exploring sexual behaviors and health communication among older women. Health Care for Women International, 38(12), 1356- 1372. doi: 10.1080/07399332.2017.1329308 Bergström‐Walan, M. B., & Nielsen, H. H. (1990). Sexual expression among 60–80‐ year‐old men and women: A sample from Stockholm, Sweden. Journal of Sex Research, 27(2), 289-295. doi: 10.1080/00224499009551558 Beutel, M. E., Burghardt, J., Tibubos, A. N., Klein, E. M., Schmutzer, G., & Brahler, E. (2018). Declining sexual activity and desire in men: Findings from representative German surveys, 2005 and 2016. Journal of Sexual Medicine, 15(5), 750-756. doi: 10.1016/j.jsxm.2018.03.010 Bevilacqua, G., Leite, M. T., Hildebrandt, L. M., & do Carmo Jahn, A. (2013). Sexuality in the perception and experience of elderly women members of a living group. Acta Scientiarum. Health Sciences, 35(1), 29-35. doi: http://dx.doi.org/10.4025/actascihealthsci.v35i1.10700 Birnbaum, G. E., Reis, H. T., Mizrahi, M., Kanat-Maymon, Y., Sass, O., & Granovski-Milner, C. (2016). Intimately connected: The importance of partner responsiveness for experiencing sexual desire. Journal of Personality and Social Psychology, 111(4), 530. doi: http://dx.doi.org/10.1037/pspi0000069 Boateng, G. O., Neilands, T. B., Frongillo, E. A., Melgar-Quiñonez, H. R., & Young, S. L. (2018). Best practices for developing and validating scales for health, social, and behavioral research: A primer. Frontiers in Public Health, 6. doi: http://dx.doi.org/10.3389/fpubh.2018.00149

179

Bond, T., & Fox, C. (2015). Applying the Rasch model: Fundamental measurement in the human sciences. New York and London: Routledge, Taylor & Francis Group. Bortz, W. M., & Wallace, D. H. (1999). Physical fitness, aging, and sexuality. Western Journal of Medicine, 170(3). Retrieved from https://www.ncbi.nlm.nih.gov/pmc/journals/183/ Bortz, W. M., Wallace, D. H., & Wiley, D. (1999). Sexual function in 1,202 aging males: differentiating aspects. Journals of Gerontology Series A: Biomedical Sciences and Medical Sciences, 54(5), M237-M241. doi: http://dx.doi.org/10.1093/gerona/54.5.M237 Bouman, W. P., & Kleinplatz, P. J. (2014). Moving towards understanding greater diversity and fluidity of sexual expression of older people. Sexual and Relationship Therapy, 30(1), 1-3. doi: 10.1080/14681994.2015.990192 Bradway, K. E., & Beard, R. L. (2015). “Don’t be trying to box folks in.” Older women’s sexuality. Affilia, 30(4), 504-518. Retrieved from https://journals.sagepub.com/loi/aff Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77-101. doi: http://dx.doi.org/10.1191/1478088706qp063oa Bretschneider, J. G., & McCoy, N. L. (1988). Sexual interest and behavior in healthy 80- to 102-year-olds. Archives of Sexual Behavior, 17(2). doi: http://dx.doi.org/10.1007/BF01542662 Burns, A., & Leonard, R. (2005). Chapters of our lives: Life narratives of midlife and older Australian women. Sex Roles, 52(5-6), 269-277. doi: 10.1007/s11199- 005-2671-8 Carmichael, C. L., Reis, H. T., & Duberstein, P. R. (2015). In your 20s it's quantity, in your 30s it's quality: The prognostic value of social activity across 30 years of adulthood. Psychology and Aging, 30(1), 95-105. doi: 10.1037/pag0000014 Carothers, B. J., & Reis, H. T. (2013). Men and women are from Earth: Examining the latent structure of gender. Journal of Personality and Social Psychology, 104(2), 385. doi: http://dx.doi.org/10.1037/a0030437 Carpenter, L., & DeLamater, J. (Eds.). (2012). Sex for life: From virginity to Viagra, how sexuality changes throughout our lives. New York: New York University Press. Carstensen, L. L. (1992). Social and emotional patterns in adulthood: Support for socioemotional selectivity theory. Psychology and Aging, 7(3), 331. doi: http://dx.doi.org/10.1037/0882-7974.7.3.331 Carstensen, L. L., Fung, H. H., & Charles, S. T. (2003). Socioemotional selectivity theory and the regulation of emotion in the second half of life. Motivation and Emotion, 27(2), 103-123. Retrieved from https://link.springer.com/journal/11031 Carstensen, L. L., Gross, J. J., & Fung, H. H. (1998). The social context of emotional experience. In K. W. Schaie & M. P. Lawton (Eds.), Annual review of gerontology and geriatrics: Vol. 17. Annual review of gerontology and geriatrics, Vol. 17. Focus on emotion and adult development (p. 325–352). Springer Publishing Co. Carvalheira, A., Traeen, B., & Stulhofer, A. (2014). Correlates of men's sexual interest: A cross-cultural study. Journal of Sexual Medicine, 11(1), 154-164. doi: 10.1111/jsm.12345

180

Carvalho, J., & Nobre, P. (2010). Gender issues and sexual desire: The role of emotional and relationship variables. Journal of Sexual Medicine, 7(7), 2469- 2478. doi: 10.1111/j.1743-6109.2009.01689.x Catania, J. A., & White, C. B. (1982). Sexuality in an aged sample: Cognitive determinants of masturbation. Archives of Sexual Behavior, 11(3). doi: http://dx.doi.org/10.1007/BF01544992 Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 9(2), 233-255. doi: http://dx.doi.org/10.1207/S15328007SEM0902_5 Choi, K. B., Jang, S. H., Lee, M. Y., & Kim, K. H. (2011). Sexual life and self-esteem in married elderly. Archives of Gerontology and Geriatrics, 53(1), e17-20. doi: 10.1016/j.archger.2010.08.011 Christenson, C. V., & Gagnon, J. H. (1965). Sexual behavior in a group of older women. Journals of Gerontology, 20(3), 351-356. Clauser, B., & Linacre, J. (1999). Relating Cronbach and Rasch reliabilities. Rasch Measurement Transactions, 13(2), 696. Retrieved from https://www.rasch.org/rmt/index.htm Collins, D. (2003). Pretesting survey instruments: An overview of cognitive methods. Quality of Life Research, 12(3), 229-238. Retrieved from https://www.springer.com/journal/11136 Connaughton, C., & McCabe, M. (2015). Sexuality and Aging. Encyclopedia of Geropsychology, 1, 1-10. doi: 10.1007/978-981-287-080-3_5-1 Corona, G., Rastrelli, G., Maseroli, E., Forti, G., & Maggi, M. (2013). Sexual function of the ageing male. Best Practice and Research: Clinical Endocrinology and Metabolism, 27(4), 581-601. doi: 10.1016/j.beem.2013.05.007 Cowden, C. R., & Koch, P. B. (1995). Attitudes related to sexual concerns: Gender and orientation comparisons. Journal of Sex Education and Therapy, 21(2), 78-87. doi: http://dx.doi.org/10.1080/01614576.1995.11074139 Craig, L., & Sawrikar, P. (2009). Work and family: How does the (gender) balance change as children grow? Gender, Work and Organization, 16(6), 684-709. doi: 10.1111/j.1468-0432.2009.00481.x Critical Appraisal Skills Programme (2015). CASP (Qualitative Research) Checklist. [online] Available at: https://casp-uk.net/casp-tools-checklists/. Accessed: 10 May 2016. Crose, R., & Drake, L. K. (1993). Older women's sexuality. Clinical Gerontologist, 12(4), 51-56. doi: 10.1300/J018v12n04_04 Deacon, S., Minichiello, V., & Plummer, D. (1995). Sexuality and older people: Revisiting the assumptions. Educational Gerontology, 21(5), 497-513. doi: 10.1080/0360127950210509 DeLamater, J. (2012). Sexual expression in later life: A review and synthesis. Journal of Sex Research, 49(2-3), 125-141. doi: 10.1080/00224499.2011.603168 DeLamater, J., Hyde, J. S., & Fong, M. C. (2008). Sexual satisfaction in the seventh decade of life. Journal of Sex and Marital Therapy, 34(5), 439-454. doi: 10.1080/00926230802156251 DeLamater, J., & Koepsel, E. (2014). Relationships and sexual expression in later life: A biopsychosocial perspective. Sexual and Relationship Therapy, 30(1), 37- 59. doi: 10.1080/14681994.2014.939506 DeLamater, J., & Moorman, S. M. (2007). Sexual behavior in later life. Journal of Aging and Health, 19(6), 921-945. doi: 10.1177/0898264307308342

181

Dello Buono, M., Zaghi, P. C., Padoani, W., Scocco, P., Urciuoli, O., Pauro, P., & De Leo, D. (1998). Sexual feelings and sexual life in an Italian sample of 335 elderly 65 to 106-year-olds. Archives of Gerontology and Geriatrics, 6, 155- 162. doi: http://dx.doi.org/10.1016/S0167-4943(98)80023-X Drench, M. E., & Losee, R. H. (1996). Sexuality and sexual capacities of elderly people. Rehabilitation Nursing, 21(3). doi: http://dx.doi.org/10.1002/j.2048- 7940.1996.tb01689.x Drennan, J. (2002). Cognitive interviewing: Verbal data in the design and pretesting of questionnaires. Journal of Advanced Nursing, 42(1), 57-63. doi: http://dx.doi.org/10.1046/j.1365-2648.2003.02579.x Drummond, J. D., Brotman, S., Silverman, M., Sussman, T., Orzeck, P., Barylak, L., & Wallach, I. (2013). The impact of caregiving. Affilia, 28(4), 415-428. doi: 10.1177/0886109913504154 Eddy, J. M., St. Pierre, R. W., & Alles, W. F. (1982). Life span development: Intervention implications for concepts of aging. The Journal of School Health, 52(9), 259-263. doi: http://dx.doi.org/10.1111/j.1746-1561.1982.tb04039.x Eliason, M. J., & Schope, R. (2001). Does “don't ask don't tell” apply to health care? Lesbian, gay, and bisexual people's disclosure to health care providers. Journal of the Gay and Lesbian Medical Association, 5(4), 125-134. doi: https://doi.org/10.1023/A:1014257910462 Erikson, E. H., & Erikson, J. M. (1998). The life cycle completed (extended version). New York. NY: WW Norton & Co, United States. Fileborn, B., Thorpe, R., Hawkes, G., Minichiello, V., & Pitts, M. (2015). Sex and the (older) single girl: Experiences of sex and dating in later life. Journal of Aging Studies, 33, 67-75. doi: 10.1016/j.jaging.2015.02.002 Fileborn, B., Thorpe, R., Hawkes, G., Minichiello, V., Pitts, M., & Dune, T. (2015). Sex, desire and pleasure: Considering the experiences of older Australian women. Sexual and Relationship Therapy, 30(1), 117-130. doi: 10.1080/14681994.2014.936722 Fingerman, K. L., Berg, C., Smith, J., & Antonucci, T. C. (Eds.). (2011). Handbook of life-span development. New York: Springer Publishing Company. Fisher Jr, W. P., & Stenner, A. J. (2011). Metrology for the social, behavioral, and economic sciences. Social, Behavioral, and Economic Sciences White Paper Series. Available at: http://www.nsf.gov/sbe/sbe_2020/submission_detail.cfm Flynn, T. J., & Gow, A. J. (2015). Examining associations between sexual behaviours and quality of life in older adults. Age and Ageing, 44(5), 823-828. doi: 10.1093/ageing/afv083 Foley, S. (2015). Older adults and sexual health: A review of current literature. Current Sexual Health Reports, 7(2), 70-79. doi: http://dx.doi.org/10.1007/s11930-015-0046-x Freak-Poli, R., Kirkman, M., De Castro Lima, G., Direk, N., Franco, O. H., & Tiemeier, H. (2017). Sexual activity and physical tenderness in older adults: Cross-sectional prevalence and associated characteristics. Journal of Sexual Medicine, 14(7), 918-927. doi: 10.1016/j.jsxm.2017.05.010 Gagnon, M., Hébert, R., Leclerc, G., & Lefrançois, R. (2002). Development and validation of a sexual actualization measuring instrument for the elderly. Educational Gerontology, 28(7), 621-633. doi: 10.1080/03601270290099804 Galinsky, A. M., McClintock, M. K., & Waite, L. J. (2014). Sexuality and physical contact in National Social Life, Health, and Aging Project Wave 2. Journals of

182

Gerontology. Series B: Psychological Sciences and Social Sciences, 69 Suppl 2, S83-98. doi: 10.1093/geronb/gbu072 Garrett, D. (2014). Psychosocial barriers to sexual intimacy for older people. British Journal of Nursing, 23(6), 327-331. Retrieved from http://ezproxy.acu.edu.au/login?url=https://search.ebscohost.com/login.aspx?d irect=true&db=mdc&AN=24690929&site=ehost-live Garson, G. D. (2012). Testing statistical assumptions. Asheboro, NC: Statistical Associates Publishing. Gewirtz-Meydan, A., Levkovich, I., Mock, M., Gur, U., & Ayalon, L. (2019). Promoting a discussion on later life sexuality: Lessons from sexologist physicians. Sexual and Relationship Therapy, 34(2), 193-210. doi: http://dx.doi.org/10.1080/14681994.2018.1437258 Giambra, L. M., & Singer, J. L. (1998). Sexual daydreaming scale of the imaginal process inventory. In C. M. Davis, W. L. Yarber, R. Bauserman, G.Schreer, & S. L. Davis (Eds.), Handbook of sexuality-related measures, (pp. 234-235). Thousand Oaks, California: Sage Publications. Ginsberg, T. B., Pomerantz, S. C., & Kramer-Feeley, V. (2005). Sexuality in older adults: Behaviours and preferences. Age and Ageing, 34(5), 475-480. doi: 10.1093/ageing/afi143 Gonçalves, R., & Merighi, M. A. B. (2009). Reflections on sexuality during the climacteric. Revista Latino-Americana de Enfermagem, 17(2), 160-166. doi: 10.1590/s0104-11692009000200004 Gott, M. (2004). Sexuality, sexual health and ageing. Berkshire, England: McGraw- Hill Education (UK). Gott, M. (2006). Sexual health and the new ageing. Age and Ageing, 35(2), 106-107. doi: 10.1093/ageing/afj050 Gott, M., & Hinchliff, S. (2003). Barriers to seeking treatment for sexual problems in primary care: A qualitative study with older people. Family Practice, 20(6). doi: 10.1093/fampra/cmg612 Gott, M., & Hinchliff, S. (2003). How important is sex in later life? The views of older people. Social Science and Medicine, 56(8), 1617-1628. doi: 10.1016/s0277-9536(02)00180-6 Gray, J. (1992). Men are from Mars, women are from Venus. USA: HarperCollins Publishers. Gray, J. (2011). Mars and Venus in the bedroom: A guide to lasting romance and passion. United Kingdom: Random House. Guan, J. (2004). Correlates of spouse relationship with sexual attitude, interest, and activity among Chinese elderly. Sexuality and Culture, 8(1), 104-131. doi: http://dx.doi.org/10.1007/s12119-004-1008-z Gurian, B. S. (1986). The myth of the aged as asexual: Countertransferrance issues in therapy. Hospital and Community Psychiatry, 37(4). doi: http://dx.doi.org/10.1176/ps.37.4.345 Haesler, E., Bauer, M., & Fetherstonhaugh, D. (2016). Sexuality, sexual health and older people: A systematic review of research on the knowledge and attitudes of health professionals. Nurse Education Today, 40, 57-71. doi: 10.1016/j.nedt.2016.02.012 Hammarberg, K., Kirkman, M., & de Lacey, S. (2016). Qualitative research methods: When to use them and how to judge them. Human Reproduction, 31(3), 498- 501. doi: 10.1093/humrep/dev334

183

Harper, J. M., Schaalje, B. G., & Sandberg, J. G. (2000). Daily hassles, intimacy, and marital quality in later life marriages. American Journal of Family Therapy, 28(1), 1-18. doi: http://dx.doi.org/10.1080/019261800261770 Heale, R., & Twycross, A. (2015). Validity and reliability in quantitative studies. Evidence-Based Nursing, 18(3), 66-67. doi: http://dx.doi.org/10.1136/eb- 2015-102129 Helgason, A. R., Adolfsson, J., Dickman, P., Arver, S., Fredrikson, M., Gothberg, M., & Steineck, G. (1996). Sexual desire, erection, orgasm and ejaculatory functions and their importance to elderly Swedish men: A population-based study. Age and Ageing, 25(4), 285-291. doi: http://dx.doi.org/10.1093/ageing/25.4.285 Henry, J., & McNab, W. (2003). Forever young. Gerontology and Geriatrics Education, 23(4), 57-74. doi: 10.1300/J021v23n04_06 Herbenick, D., Reece, M., Schick, V., Sanders, S. A., Dodge, B., & Fortenberry, J. D. (2010). Sexual behavior in the United States: Results from a national probability sample of men and women ages 14-94. Journal of Sexual Medicine, 7 Suppl 5, 255-265. doi: 10.1111/j.1743-6109.2010.02012.x Herron, M. J., & Herron, W. G. (1999). Elder sexuality. Journal of Social Distress and the Homeless, 8(3), 167-176. doi: http://dx.doi.org/10.1023/A:1021316809953 Heywood, W., Lyons, A., Fileborn, B., Hinchliff, S., Minichiello, V., Malta, S., . . . Dow, B. (2018). Sexual satisfaction among older Australian heterosexual men and women: Findings from the Sex, Age & Me study. Journal of Sex and Marital Therapy, 44(3), 295-307. doi: 10.1080/0092623X.2017.1366959 Hinchliff, S., & Gott, M. (2004). Intimacy, commitment, and adaptation: Sexual relationships within long-term marriages. Journal of Social and Personal Relationships, 21(5), 595-609. doi: http://dx.doi.org/10.1177/0265407504045889 Hinchliff, S., & Gott, M. (2008). Challenging social myths and stereotypes of women and aging: Heterosexual women talk about sex. Journal of Women and Aging, 20(1-2), 65-81. doi: 10.1300/J074v20n01_06 Ho, R. (2013). Handbook of univariate and multivariate data analysis with IBM SPSS. Florida, USA: Chapman and Hall/CRC. Hoffnung, M., Seifert, K., Hine, A., Ward, L., Pause, C., & Swabey, K. (2016). Lifespan development (Third Australasian ed.). Stafford, Queensland: Wiley. Holzapfel, S. (1994). Aging and sexuality. Canadian Family Physician, 40, 748. doi: Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2380103/ Hu, L. t., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1-55. doi: http://dx.doi.org/10.1080/10705519909540118 Huang, A. J., Subak, L. L., Thom, D. H., Van Den Eeden, S. K., Ragins, A. I., Kuppermann, M., . . . Brown, J. S. (2009). Sexual function and aging in racially and ethnically diverse women. Journal of the American Geriatrics Society, 57(8), 1362-1368. doi: 10.1111/j.1532-5415.2009.02353.x Hurd Clarke, L. (2006). Older women and sexuality: Experiences in marital relationships across the life course. Canadian Journal on Aging, 25(02), 129. doi: 10.1353/cja.2006.0034 Hyde, Z., Flicker, L., Hankey, G. J., Almeida, O. P., McCaul, K. A., Chubb, S. P., & Yeap, B. B. (2010). Prevalence of sexual activity and associated factors in

184

men aged 75 to 95 years: A cohort study. Annals of Internal Medicine, 153(11), 693-702. doi: http://dx.doi.org/10.7326/0003-4819-153-11- 201012070-00002 IBM Corporation. (2013). IBM SPSS statistics for Windows (Version 23). Armonk, NY: IBM Corporation. IBM Corporation. (2017). IBM SPSS statistics for Windows (Version 25). Armonk, NY: IBM Corporation. Iveniuk, J., O'Muircheartaigh, C., & Cagney, K. A. (2016). Religious influence on older Americans' sexual lives: A nationally-representative profile. Archives of Sexual Behavior, 45(1), 121-131. doi: 10.1007/s10508-015-0534-0 Johnson, B. K. (1996). Older adults and sexuality: A multidimensional perspective. Journal of Gerontological Nursing, 22(2), 6-9. doi: http://dx.doi.org/10.3928/0098-9134-19960201-06 Jung, A., & Schill, W. B. (2004). Male sexuality with advancing age. European Journal of Obstetrics, Gynecology, and Reproductive Biology, 113(2), 123- 125. doi: 10.1016/j.ejogrb.2003.10.007 Kaas, M. J., Rousseau, G. K., & Haworth Continuing Features Submission. (2008). Geriatric sexual conformity. Clinical Gerontologist, 2(1), 31-44. doi: 10.1300/J018v02n01_03 Kalra, G., Subramanyam, A., & Pinto, C. (2011). Sexuality: Desire, activity and intimacy in the elderly. Indian Journal of Psychiatry, 53(4), 300-306. doi: 10.4103/0019-5545.91902 Kaplan, H. S. (1990). Sex, intimacy, and the aging process. Journal of the American Academy of Psychoanalysis, 18(2), 185-205. doi: http://dx.doi.org/10.1521/jaap.1.1990.18.2.185 Karraker, A., & DeLamater, J. (2013). Past-year sexual inactivity among older married persons and their partners. Journal of Marriage and Family, 75(1), 142-163. doi: 10.1111/j.1741-3737.2012.01034.x Karraker, A., Delamater, J., & Schwartz, C. R. (2011). Sexual frequency decline from midlife to later life. Journals of Gerontology. Series B: Psychological Sciences and Social Sciences, 66(4), 502-512. doi: 10.1093/geronb/gbr058 Kazer, M. W., Grossman, S., Kerins, G., Kris, A., & Tocchi, C. (2013). Validity and reliability of the Geriatric Sexuality Inventory. Journal of Gerontological Nursing, 39(11), 38-45. doi: http://dx.doi.org/10.3928/00989134-20130916-03 Kellett, J. M. (1991). Sexuality of the elderly. Sexual and Marital Therapy, 6(2), 147- 155. doi: 10.1080/02674659108406531 Kennedy, G. J., Martinez, M. M., & Garo, N. (2010). Sex and mental health in old age. Primary Psychiatry, 17(1), 22-30. Retrieved from http://www.primarypsychiatry.com/ Kessel, B. (2001). Sexuality in the older person. Age And Ageing, 30(2), 121-124. Retrieved from http://ezproxy.acu.edu.au/login?url=https://search.ebscohost.com/login.aspx?d irect=true&db=mdc&AN=11395341&site=ehost-live Killinger, K. A., Boura, J. A., & Diokno, A. C. (2014). Exploring factors associated with sexual activity in community-dwelling older adults. Research in Gerontological Nursing, 7(6), 256-263. doi: 10.3928/19404921-20141006-01 Kim, J. E., & Moen, P. (2002). Retirement transitions, gender, and psychological well-being: A life-course, ecological model. The Journals of Gerontology: Series B, 57(3), P212-P222. doi: 10.1093/geronb/57.3.P212

185

Kim, J. S., & Kang, S. (2015). A study on body image, sexual quality of life, depression, and quality of life in middle-aged adults. Asian Nursing Research, 9(2), 96-103. doi: http://dx.doi.org/10.1016/j.anr.2014.12.001 Kim, O., & Jeon, H. O. (2013). Gender differences in factors influencing sexual satisfaction in Korean older adults. Archives of Gerontology and Geriatrics, 56(2), 321-326. doi: 10.1016/j.archger.2012.10.009 Kingsberg, S. A. (2002). The impact of aging on sexual function in women and their partners. Archives of Sexual Behavior, 31(5), 431-437. doi: 10.1023/a:1019844209233 Kinsey, A.C., Pomeroy, W.B., & Martin, C.E. (1948). Sexual Behavior in the Human Male. Philadelphia: W.B. Kinsey, A.C., Pomeroy, W.B., Martin, C.E., & Gebhard, P.H. (1953). Sexual Behavior in the Human Female Philadelphia: W.B. Saunders. Kleinplatz, P. J., Ménard, A. D., Paquet, M. P., Paradis, N., Campbell, M., Zuccarino, D., & Mehak, L. (2009). The components of optimal sexuality: A portrait of “great sex”. Canadian Journal of Human Sexuality, 18(1-2), 1-13. Retrieved from: https://www.researchgate.net/publication/232545283_The_components_ of_optimal_sexuality_A_portrait_of_great_sex Knodel, J., & Chayovan, N. (2001). Sexual activity among older Thais: The influence of age, gender and health. Journal of Cross-Cultural Gerontology, 16(2), 173- 200. doi: 10.1023/a:1010608226594 Koo, T. K., & Li, M. Y. (2016). A guideline of selecting and reporting intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine, 15(2), 155-163. doi: http://dx.doi.org/10.1016/j.jcm.2016.02.012 Koren, C. (2014). Together and apart: A typology of re-partnering in old age. International Psychogeriatrics, 26(8), 1327-1350. doi: http://dx.doi.org/10.1017/S1041610214000738 Kruskal, J. B. (1964). Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis. Psychometrika, 29(1), 1-27. doi: http://dx.doi.org/10.1007/BF02289565 Lachman, M. E. (2004). Development in midlife. Annual Review of Psychology, 55, 305-331. doi: 10.1146/annurev.psych.55.090902.141521 Lagana, L., & Maciel, M. (2010). Sexual desire among Mexican-American older women: A qualitative study. Culture, Health and Sexuality, 12(6), 705-719. doi: 10.1080/13691058.2010.482673 Lammerink, E. A. G., de Bock, G. H., Pascal, A., van Beek, A. P., van den Bergh, A. C. M., Sattler, M. G. A., & Mourits, M. J. E. (2017). A survey of female sexual functioning in the general Dutch population. Journal of Sexual Medicine, 14(7), 937-949. doi: 10.1016/j.jsxm.2017.04.676 Langer-Most, O., & Langer, N. (2010). Aging and sexuality: How much do gynecologists know and care? Journal of Women and Aging, 22(4), 283-289. doi: 10.1080/08952841.2010.518882 Langer, N. (2009). Late life love and intimacy. Educational Gerontology, 35(8), 752- 764. doi: 10.1080/03601270802708459 Laumann, E. O., & Waite, L. J. (2008). Sexual dysfunction among older adults: Prevalence and risk factors from a nationally representative US probability sample of men and women 57–85 years of age. The Journal of Sexual Medicine, 5(10), 2300-2311. doi: http://dx.doi.org/10.1111/j.1743- 6109.2008.00974.x

186

Lehmiller, J. J. (2017). The psychology of human sexuality: 2nd edition. New York: John Wiley & Sons. Levinson, D. J. (1986). A conception of adult development. American Psychologist, 41(1), 3. doi: http://dx.doi.org/10.1037/0003-066X.41.1.3 Linacre, J. (1994). Sample size and item calibration stability. Rasch Measurement Transactions, 7, 328. Linacre, J. (2009). WINSTEPS (Version 3.70. 0.3). [Computer Software]. Chicago: Winsteps. com. Lindau, S. T., Schumm, L. P., Laumann, E. O., Levinson, W., O'muircheartaigh, C. A., & Waite, L. J. (2007). A study of sexuality and health among older adults in the United States. New England Journal of Medicine, 357(8), 762-774. doi: http://dx.doi.org/10.1056/NEJMoa067423 Lindley, S. E., Harper, R., & Sellen, A. (2008, September). Designing for elders: Exploring the complexity of relationships in later life. In Proceedings of the 22nd British HCI Group Annual Conference on People and Computers: Culture, Creativity, Interaction-Volume 1 (pp. 77-86). British Computer Society. doi: http://dx.doi.org/10.14236/ewic/HCI2008.8 Linn, R. L. (2011). The standards for educational and psychological testing: Guidance in test development. In T. M. Haladyna & S. M. Downing (Eds.), Handbook of test development, (pp. 41-52) New York:Routledge. doi: http://dx.doi.org/10.4324/9780203874776.ch2 Lodge, A. C., & Umberson, D. (2012). All shook up: Sexuality of mid- to later life married couples. Journal of Marriage and Family, 74(3), 428-443. doi: 10.1111/j.1741-3737.2012.00969.x Loe, M. (2004). Sex and the senior woman: Pleasure and danger in the Viagra era. Sexualities, 7(3), 303-326. doi: 10.1177/1363460704044803 Loehr, J., Verma, S., & Seguin, R. (1997). Issues of sexuality in older women. Journal of Women's Health, 6(4), 451-457. doi: http://dx.doi.org/10.1089/jwh.1997.6.451 Lusti-Narasimhan, M., & Beard, J. R. (2013). Sexual health in older women. Bulletin of the World Health Organization, 91(9), 707-709. doi: 10.2471/BLT.13.119230 Macleod, A., Busija, L., & McCabe, M. (in press). Mapping the perceived sexuality of heterosexual men and women in mid-and later life: A mixed-methods study. Sexual Medicine. doi: http://dx.doi.org/10.1016/j.esxm.2019.10.001 Macleod, A., & McCabe, M. P. (accepted). Defining sexuality in later life: A systematic review Australasian Journal on Ageing(00), 1-10. doi: https://doi.org/10.1111/ajag.12741 Macleod, A., & McCabe, M. P. (2019). How well do measurement scales reflect the actual experience of sexuality in mid-life and beyond? Sexual Medicine Reviews, 7(1), 29-45. doi: 10.1016/j.sxmr.2018.04.004 Malakouti, S. K., Salehi, M., Nojomi, M., Zandi, T., & Eftekhar, M. (2012). Sexual functioning among the elderly population in Tehran, Iran. Journal of Sex and Marital Therapy, 38(4), 365-377. doi: 10.1080/0092623X.2011.628438 Malatesta, V. J. (2007). Sexual problems, women and aging: An overview. Journal of Women and Aging, 19(1-2), 139-154. doi: 10.1300/J074v19n01_09 Malloy, T. E. (2010). Bem sex role inventory. The Corsini Encyclopedia of Psychology, 1-1. doi: https://doi.org/10.1002/9780470479216.corpsy0123

187

Mansfield, P. K., Koch, P. B., & Voda, A. M. (2000). Midlife women's attributions for their sexual response changes. Health Care for Women International, 21(6), 543-559. doi: 10.1080/07399330050130322 Marcia, J. E. (1966). Development and validation of ego-identity status. Journal of Personality and Social Psychology, 3(5), 551. doi: http://dx.doi.org/10.1037/h0023281 Marshall, B. L. (2010). Science, medicine and virility surveillance: 'Sexy seniors' in the pharmaceutical imagination. Sociology of Health and Illness, 32(2), 211- 224. doi: 10.1111/j.1467-9566.2009.01211.x Marshall, B. L. (2012). Medicalization and the refashioning of age-related limits on sexuality. Journal of Sex Research, 49(4), 337-343. doi: 10.1080/00224499.2011.644597 Marsiglio, W., & Donnelly, D. (1991). Sexual relations in later life: A national study of married persons. Journal of Gerontology, 46(6), S338-S344. doi: http://dx.doi.org/10.1093/geronj/46.6.S338 Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50(4), 370. doi: http://dx.doi.org/10.1037/h0054346 Masters, W. H., & Johnson, V. E. (1966). Human sexual response. Boston: Little Brown. Masters, W. H., & Johnson, V. E. (1981). Sex and the aging process. Journal of the American Geriatrics Society, 29(9), 385-390. doi: http://dx.doi.org/10.1111/j.1532-5415.1981.tb02375.x Matthias, R. E., Lubben, J. E., Atchison, K. A., & Schweitzer, S. O. (1997). Sexual activity and satisfaction among very old adults: Results from a community- dwelling Medicare population survey. The Gerontologist, 37(1), 6-14. doi: http://dx.doi.org/10.1093/geront/37.1.6 Maul, A., Torres Irribarra, D., & Wilson, M. (2016). On the philosophical foundations of psychological measurement. Measurement, 79, 311-320. doi: 10.1016/j.measurement.2015.11.001 Mazo, G. Z., & Cardoso, F. L. (2011). Sexual satisfaction and correlates among elderly Brazilians. Archives of Gerontology and Geriatrics, 52(2), 223-227. doi: http://dx.doi.org/10.1016/j.archger.2010.03.024 McCarthy, B., Farr, E., & McDonald, D. (2013). Couple sexuality after 60. Journal of Family Psychotherapy, 24(1), 38-47. doi: 10.1080/08975353.2013.762867 McCracken, A. L. (1988). Sexual practice by elders: The forgotten aspect of functional health. Journal of Gerontological Nursing, 14(10), 13-18. doi: http://dx.doi.org/10.3928/0098-9134-19881001-05 McHugh, M. C., & Interligi, C. (2015). Sexuality and older women: Desirability and desire. In V. Muhlbauer, J. C. Chrisler, & F. L. Denmark (Eds.), Women and aging: An international,intersectional, power perspective (pp. 89-116). New York: Springer. McIntosh, D. (1981). Sexual attitudes in a group of older women. Issues in Mental Health Nursing, 3(1-2), 109-122. doi: 10.3109/01612848109140864 Ménard, A. D., Kleinplatz, P. J., Rosen, L., Lawless, S., Paradis, N., Campbell, M., & Huber, J. D. (2014). Individual and relational contributors to optimal sexual experiences in older men and women. Sexual and Relationship Therapy, 30(1), 78-93. doi: 10.1080/14681994.2014.931689 Mercer, N. (2002). Words and minds: How we use language to think together. London: Routledge. doi: http://dx.doi.org/10.4324/9780203464984

188

Minichiello, V., Plummer, D., & Loxton, D. (2004). Factors predicting sexual relationships in older people: An Australian study. Australasian Journal on Ageing, 23(3), 125-130. doi: http://dx.doi.org/10.1111/j.1741- 6612.2004.00018.x Momtaz, Y. A., Hamid, T. A., Ibrahim, R., & Akahbar, S. A. (2014). Racial and socioeconomic disparities in sexual activity among older married Malaysians. Archives of Gerontology and Geriatrics, 58(1), 51-55. doi: 10.1016/j.archger.2013.08.003 Moore, K. L. (2010). Sexuality and sense of self in later life: Japanese men's and women's reflections on sex and aging. Journal of Cross-Cultural Gerontology, 25(2), 149-163. doi: 10.1007/s10823-010-9115-9 Müller, B., Nienaber, C. A., Reis, O., Kropp, P., & Meyer, W. (2014). Sexuality and affection among elderly German men and women in long-term relationships: Results of a prospective population-based study. PloS One, 9(11), e111404- e111404. doi: 10.1371/journal.pone.0111404 Muruthi, B., McCoy, M., Chou, J., & Farnham, A. (2018). Sexual scripts and narrative therapy with older couples. American Journal of Family Therapy, 46(1), 81-95. doi: 10.1080/01926187.2018.1428129 Myers, W. A. (1985). Sexuality in the older individual. Journal of the American Academy of Psychoanalysis, 13(4), 511-520. doi: http://dx.doi.org/10.1521/jaap.1.1985.13.4.511 National Institutes of Health. NIH quality assessment tool for observational cohort and cross-sectional studies. Available from https://www.nhlbi.nih.gov/health- pro/guidelines/in-develop/cardiovascular-risk-reduction/tools/cohort. Accessed 10 May 2016 Newman, G., & Nichols, C. R. (1960). Sexual activities and attitudes in older persons. Journal of the American Medical Association, 173(1), 33-35. doi: http://dx.doi.org/10.1001/jama.1960.03020190035007 Nicolosi, A., Laumann, E. O., Glasser, D. B., Moreira Jr, E. D., Paik, A., & Gingell, C. (2004). Sexual behavior and sexual dysfunctions after age 40: The global study of sexual attitudes and behaviors. Urology, 64(5), 991-997. doi: http://dx.doi.org/10.1016/j.urology.2004.06.055 Nilsson, L. (1987). Sexuality in the elderly. Acta Obstetricia et Gynecologica Scandinavica, 66(sup140), 52-58. doi: http://dx.doi.org/10.3109/00016348709156479 Padoani, W., Dello Buono, M., Marietta, P., Scocco, P., Zaghi, P., & De Leo, D. (2000). Influence of cognitive status on the sexual life of 352 elderly Italians aged 65–105 years. Gerontology, 46(5), 258-265. doi: http://dx.doi.org/10.1159/000022169 Palacios-Cena, D., Carrasco-Garrido, P., Hernandez-Barrera, V., Alonso-Blanco, C., Jimenez-Garcia, R., & Fernandez-de-las-Penas, C. (2012). Sexual behaviors among older adults in Spain: Results from a population-based national sexual health survey. Journal of Sexual Medicine, 9(1), 121-129. doi: 10.1111/j.1743- 6109.2011.02511.x Papaharitou, S., Nakopoulou, E., Kirana, P., Giaglis, G., Moraitou, M., & Hatzichristou, D. (2008). Factors associated with sexuality in later life: An exploratory study in a group of Greek married older adults. Archives of Gerontology and Geriatrics, 46(2), 191-201. doi: 10.1016/j.archger.2007.03.008

189

Pendrill, L. (2013). Discrete ordinal and interval scaling and psychometrics. Paper presented at the 16th International Congress of Metrology. doi: http://dx.doi.org/10.1051/metrology/201314001 Penhollow, T. M., Young, M., & Denny, G. (2009). Predictors of quality of life, sexual intercourse, and sexual satisfaction among active older adults. American Journal of Health Education, 40(1), 14-22. doi: http://dx.doi.org/10.1080/19325037.2009.10599074 Petersen, J. L., & Hyde, J. S. (2011). Gender differences in sexual attitudes and behaviors: A review of meta-analytic results and large datasets. Journal of Sex Research, 48(2-3), 149-165. doi: http://dx.doi.org/10.1080/00224499.2011.551851 Peterson, C. H., Peterson, N. A., & Powell, K. G. (2017). Cognitive interviewing for item development: Validity evidence based on content and response processes. Measurement and Evaluation in Counseling and Development, 50(4), 217- 223. doi: http://dx.doi.org/10.1080/07481756.2017.1339564 Pfeiffer, E., Verwoerdt, A., & Wang, H.-S. (1969). The natural history of sexual behavior in a biologically advantaged group of aged individuals. Journal of Gerontology, 24(2), 193-198. doi: http://dx.doi.org/10.1093/geronj/24.2.193 Phillipson, C. (2013). Ageing. Cambridge, United Kingdom: John Wiley & Sons. Portonova, M., Young, E., & Newman, M. A. (1984). Elderly women's attitudes toward sexual activity among their peers. Health Care for Women International, 5(5-6), 289-298. doi: 10.1080/07399338409515661 Qualtrics. (2017). Qualtrics (Version 2016). Provo, Utah, USA. Available at http://www.qualtrics.com Qualtrics. (2018). Qualtrics XM (Version XM). Provo, Utah, USA. Available at: https://www.qualtrics.com Rahn, A. A., Bennett, C., Lykins, A., & Jones, T. (2018). Behind closed doors: Exploring ways to support partnered Baby Boomers' coupledom in residential aged care settings. (Doctoral thesis, University of New England, Sydney, Australia). Retrieved from https://hdl.handle.net/1959.11/26566 Ravanipour, M., Gharibi, T., & Gharibi, T. (2013). Elderly women’s views about sexual desire during old age: A qualitative study. Sexuality and Disability, 31(2), 179-188. doi: 10.1007/s11195-013-9295-7 Renshaw, D. C. (1983). Sex, intimacy, and the older woman. Women and Health, 8(4), 43-54. doi: 10.1300/J013v08n04_05 Rosas, S. R., & Kane, M. (2012). Quality and rigor of the concept mapping methodology: A pooled study analysis. Evaluation and Program Planning, 35(2), 236-245. doi: 10.1016/j.evalprogplan.2011.10.003 Rosen, C., Brown, J., Heiman, S., Leiblum, C., Meston, R., Shabsigh, D., . . . D'Agostino, R. (2000). The Female Sexual Function Index (FSFI): A multidimensional self-report instrument for the assessment of female sexual function. Journal of Sex and Marital Therapy, 26(2), 191-208. doi: 10.1080/009262300278597 Rosen, R. C., Riley, A., Wagner, G., Osterloh, I. H., Kirkpatrick, J., & Mishra, A. (1997). The International Index of Erectile Function (IIEF): A multidimensional scale for assessment of erectile dysfunction. Urology, 49(6), 822-830. doi: http://dx.doi.org/10.1016/S0090-4295(97)00238-0 Russell, P. (1998). Sexuality in the lives of older people. Nursing Standard (through 2013), 13(8), 49. doi: http://dx.doi.org/10.7748/ns1998.11.13.8.49.c2559

190

Sandberg, L. (2013). Just feeling a naked body close to you: Men, sexuality and intimacy in later life. Sexualities, 16(3-4), 261-282. doi: 10.1177/1363460713481726 Sandberg, L. (2016). In lust we trust? Masculinity and sexual desire in later life. Men and Masculinities, 19(2), 192-208. doi: 10.1177/1097184x15606948 Sanders, S. A., Herbenick, D., Reece, M., Schick, V., Mullinax, M., Dodge, B., & Fortenberry, J. D. (2013). The development and validation of a brief Quality of Sexual Experience (QSE) scale: Results from a nationally representative sample of men and women in the United States. Journal of Sexual Medicine, 10(10), 2409-2417. doi: 10.1111/jsm.12198 Schaefer, M. T., & Olson, D. H. (1981). Assessing intimacy: The PAIR inventory. Journal of Marital and Family Therapy, 7(1), 47-60. doi: http://dx.doi.org/10.1111/j.1752-0606.1981.tb01351.x Schulz, R., & Heckhausen, J. (1996). A life span model of successful aging. American Psychologist, 51(7), 702. doi: http://dx.doi.org/10.1037/0003-066X.51.7.702 Segraves, R., & Segraves, K. (1995). Human sexuality and aging. Journal of Sex Education and Therapy, 21(2), 88-102. doi: http://dx.doi.org/10.1080/01614576.1995.11074140 Sharpe, T. H. (2004). Introduction to sexuality in late life. The Family Journal, 12(2), 199-205. doi: 10.1177/0022167804264106 Shearer, M. R., & Shearer, M. L. (1977). Sexuality and sexual counseling in the elderly. Clinical Obstetrics and Gynecology, 20(1), 197-208. Retrieved from http://ezproxy.acu.edu.au/login?url=https://search.ebscohost.com/login.aspx?d irect=true&db=mdc&AN=837585&site=ehost-live Shkolnik, D., & Iecovich, E. (2013). Health, body image, gender, and migration status: Their relationship to sexuality in old age. International Psychogeriatrics, 25(10), 1717-1727. doi: 10.1017/S1041610213000604 Smith, K. R., & Moen, P. (1988). Passage through midlife: Women's changing family roles and economic well-being. The Sociological Quarterly, 29(4), 503-524. doi: http://dx.doi.org/10.1111/j.1533-8525.1988.tb01431.x Smith, L. J., Mulhall, J. P., Deveci, S., Monaghan, N., & Reid, M. C. (2007). Sex after seventy: A pilot study of sexual function in older persons. Journal of Sexual Medicine, 4(5), 1247-1253. doi:10.1111/j.1743-6109.2007.00568.x Steinke, E. E., & Bergen, M. B. (1986). Sexuality and aging. Journal of Gerontological Nursing, 12(6), 6-9. doi: http://dx.doi.org/10.3928/0098-9134- 19860601-04 Stimson, A., Wase, J. F., & Stimson, J. (1981). Sexuality and self-esteem among the aged. Research on Aging, 3(2), 228-239. doi: http://dx.doi.org/10.1177/016402758132008 Stulhofer, A., Hinchliff, S., Jurin, T., Hald, G. M., & Traeen, B. (2018). Successful aging and changes in sexual interest and enjoyment among older European men and women. Journal of Sexual Medicine, 15(10), 1393-1402. doi: 10.1016/j.jsxm.2018.08.011 Syme, M. (2014). The evolving concept of older adult sexual behavior and its benefits. Generations, 38(1), 35-41. Retrieved from https://www.asaging.org/generations-journal-american-society-aging Taylor, A., & Gosney, M. A. (2011). Sexuality in older age: Essential considerations for healthcare professionals. Age and Ageing, 40(5), 538-543. doi: 10.1093/ageing/afr049

191

Thompson, W. K., Charo, L., Vahia, I. V., Depp, C., Allison, M., & Jeste, D. V. (2011). Association between higher levels of sexual function, activity, and satisfaction and self-rated successful aging in older postmenopausal women. Journal of the American Geriatrics Society, 59(8), 1503-1508. doi: 10.1111/j.1532-5415.2011.03495.x Travis, S. S. (1987). Older adults' sexuality and remarriage. Journal of Gerontological Nursing, 13(6), 8-9. doi: http://dx.doi.org/10.3928/0098-9134-19870601-04 Trochim, W., & Kane, M. (2005). Concept mapping: An introduction to structured conceptualization in health care. International Journal for Quality in Health Care, 17(3), 187-191. doi: http://dx.doi.org/10.1093/intqhc/mzi038 Trochim, W. M. (1989). An introduction to concept mapping for planning and evaluation. Evaluation and Program Planning, 12(1), 1-16. doi: http://dx.doi.org/10.1016/0149-7189(89)90016-5 Trudel, G. (2002). Sexuality and marital life: Results of a survey. Journal of Sex and Marital Therapy, 28(3), 229-249. doi: http://dx.doi.org/10.1080/009262302760328271 Trudel, G., Dargis, L., Villeneuve, L., Cadieux, J., Boyer, R., & Préville, M. (2013). Marital, sexual and psychological functioning of older couples living at home: The results of a national survey using longitudinal methodology (Part I). Sexologies, 22(4), e101-e107. doi: 10.1016/j.sexol.2013.03.005 Trudel, G., Dargis, L., Villeneuve, L., Cadieux, J., Boyer, R., & Préville, M. (2014). Marital, sexual and psychological functioning of older couples living at home: The results of a national survey using longitudinal methodology (Part II). Sexologies, 23(2), e35-e48. doi: 10.1016/j.sexol.2013.03.007 Trudel, G., Turgeon, L., & Piché, L. (2010). Marital and sexual aspects of old age. Sexual and Relationship Therapy, 15(4), 381-406. doi: 10.1080/713697433 Twenge, J. M., Sherman, R. A., & Wells, B. E. (2017). Declines in sexual frequency among American adults, 1989-2014. Archives of Sexual Behavior, 46(8), 2389-2401. doi: 10.1007/s10508-017-0953-1 Umidi, S., Pini, M., Ferretti, M., Vergani, C., & Annoni, G. (2007). Affectivity and sexuality in the elderly: Often neglected aspects. Archives of Gerontology and Geriatrics, 44 Suppl 1, 413-417. doi: 10.1016/j.archger.2007.01.058 United Nations, Statistical Office. (1982). Provisional guidelines on standard international age classifications. New York: United Nations. United Nations, Department of Economic and Social Affairs, Population Division (2019). World Population Prospects 2019: Highlights (ST/ESA/SER. A/423). In: United Nations New York. Van Zile-Tamsen, C. (2017). Using Rasch analysis to inform rating scale development. Research in Higher Education, 58(8), 922-933. doi: 10.1007/s11162-017-9448-0 Wang, T.-F., Lu, C.-H., Chen, I. J., & Yu, S. (2008). Sexual knowledge, attitudes and activity of older people in Taipei, Taiwan. Journal of Clinical Nursing, 17(4), 443-450. doi: 10.1111/j.1365-2702.2007.02003.x Watson, W. K., Bell, N. J., & Stelle, C. (2010). Women narrate later life remarriage: Negotiating the cultural to create the personal. Journal of Aging Studies, 24(4), 302-312. doi: http://dx.doi.org/10.1016/j.jaging.2010.07.002 Watson, W. K., & Stelle, C. (2011). Dating for older women: Experiences and meanings of dating in later life. Journal of Women and Aging, 23(3), 263-275. doi: 10.1080/08952841.2011.587732

192

Watters, Y., & Boyd, T. V. (2009). Sexuality in later life: Opportunity for reflections for healthcare providers. Sexual and Relationship Therapy, 24(3-4), 307-315. doi: 10.1080/14681990903398047 Weinstein, E. (2013). Senior Adult Sexuality Scales. In T. D. Fisher, C. M. Davis, W. L. Yarber, & S. L. Davis (Eds.), Handbook of sexuality-related measures (Third ed.), (pp. 66-73). New York & London: Routledge. Weinstein, S., & Rosen, E. (1988). Senior adult sexuality in age segregated and age integrated communities. The International Journal of Aging and Human Development, 27(4), 261-270. doi: http://dx.doi.org/10.2190/06HW-J4DF- PJ2C-80M0 Weizman, R., & Hart, J. (1987). Sexual behavior in healthy married elderly men. Archives of Sexual Behavior, 16(1). doi: http://dx.doi.org/10.1007/BF01541840 White, C. B. (1982). A scale for the assessment of attitudes and knowledge regarding sexuality in the aged. Archives of Sexual Behavior, 11(6), 491-502. doi: http://dx.doi.org/10.1007/BF01542474 Willert, A., & Semans, M. (2000). Knowledge and attitudes about later life sexuality: What clinicians need to know about helping the elderly. Contemporary Family Therapy, 22(4). doi: https://doi.org/10.1023/A:1007896817570 Winn, R. L., & Newton, N. (1982). Sexuality in aging: A study of 106 cultures. Archives of Sexual Behavior, 11(4). doi: 10.1007/bf01541590 Wood, J. M., Mansfield, P. K., & Koch, P. B. (2007). Negotiating sexual agency: Postmenopausal women’s meaning and experience of sexual desire. Qualitative Health Research, 17(2), 189-2000. doi: 10.1177/1049732306297415 World Health Organization. (2006). Defining sexual health: report of a technical consultation on sexual health, 28-31 January 2002, Geneva. World Health Organization. Retrieved from https://www.who.int/reproductivehealth/publications/sexual_health/defining_s exual_health.pdf Worthington, R. L., Navarro, R. L., Savoy, H. B., & Hampton, D. (2008). Development, reliability, and validity of the Measure of Sexual Identity Exploration and Commitment (MoSIEC). Developmental Psychology, 44(1), 22. doi: http://dx.doi.org/10.1037/0012-1649.44.1.22 Wright, B. D., & Linacre, J. M. (1994). Reasonable mean-square fit values. Rasch Measurement Transactions. Retrieved from https://rasch.org/rmt/rmt83b.htm Wylie, K. R., Wood, A., & McManus, R. (2013). Sexuality and old age. Bundesgesundheitsblatt, Gesundheitsforschung, Gesundheitsschutz., 56(2), 223-230. doi: 10.1007/s00103-012-1602-4 Yan, E., Wu, A. M., Ho, P., & Pearson, V. (2011). Older Chinese men and women's experiences and understanding of sexuality. Culture, Health and Sexuality, 13(9), 983-999. doi: 10.1080/13691058.2011.605471 Youn, G. (2009). Marital and sexual conflicts in elderly Korean people. Journal of Sex and Marital Therapy, 35(3), 230-238. doi: http://dx.doi.org/10.1080/00926230802716369 Yun, O., Kim, M., & Chung, S. E. (2014). The sexuality experience of older widows in Korea. Qualitative Health Research, 24(4), 474-483. doi: 10.1177/1049732313518978 Zimmer-Gembeck, M. J., & Petherick, J. (2006). Intimacy dating goals and relationship satisfaction during adolescence and emerging adulthood: Identity

193

formation, age and sex as moderators. International Journal of Behavioral Development, 30(2), 167-177. doi: http://dx.doi.org/10.1177/0165025406063636

194

Appendix A: Systematic review search terms

195

Appendix B: Study 1 ethics approvals and related documentation

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

Appendix C: Study 2 ethics approvals and related documentation

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

Appendix D: Chapter 5 supplementary materials Table D1 Cognitive scorecard for all OASIS items

261

262

263

264

Table D2 Response category endorsement and fit assessments of OASIS subscale items

265

266

267

268

269

270

271

Table D3

Final version of the Over 45s Adult Sexuality and Intimacy Scale

272

273

274

Table D4 Standardised and unstandardized loadings (standard errors) for items in nine-factor confirmatory factor model of sexuality in mid and later life

275

276

277

278

279

280

Appendix E: Study 3 ethics approvals and related documentation

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309