Family caregiving experience and health status among Chinese in the United States
Item Type text; Dissertation-Reproduction (electronic)
Authors Hsueh, Kuei-Hsiang
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Download date 03/10/2021 20:50:45
Link to Item http://hdl.handle.net/10150/279876 INFORMATION TO USERS
This manuscript has been reproduced fix)m the microfilm master UMI films the text directly f^ the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.
In the unlikely event that the author dkl not send UMI a complete manuscript and there are missing pages, these will be noted. Also, unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sectk>ns with small overiaps.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6' x 9' black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.
ProQuest Infbrmatkm and Learning 300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA 800-521-0600
FAJVOLY CAREGIVING EXPERIENCE AND HEALTH STATUS
AMONG CHINESE IN THE UNITED STATES
By
Kuei-Hsiang Hsueh
Copyright © Kuei-Hsiang Hsueh 2001
A Dissertation Submitted to the Faculty of the
COLLEGE OF NURSING
In Partial ftilfillment of the Requirements For the Degree of
DOCTOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
200 t UMI Number: 3031388
Copyright2001 by Hsueh, Kuet-Hsiang
All rights reserved.
UMI(ft
UMI Microform 3031388 Copyright 2002 by Bell & Howell Informatfon and Leamfng Company. All rights reserved. This microfonm edition is protected against unauthorized copying under Title 17, United States Code.
Bell & Howell Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, MI48106-1346 2
THE UNIVERSITY OF ARIZONA ® GRADUATE COLLEGE
As nembers of the Final Examination Committee, we certify that we have
read the dissertation prepared by Kuet-Hstanq Hsueh
entitled Famfly caregiving experience and health status amonn
Chinese caregivers in the United States
and recoDBend that it be accepted as fulfilling the dissertation
requirement for the Degree of Doctor of Philosophy
9^z/<= ?^Linda R.. PhilJ4PS^Phil^ps^ ^
DrTsandra ^SiTwel^ ^ 0_v D?t/^^^
if 27/01 Dr. Lee Sechrest Date
f/C/ 7 /0 / Dr. mne V/oodtli Date
Date
Final approval and acceptance of this dissertation, is contingent upon the candidate's submission of the final copy of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under ny direction and reconmend that it be accepted as fulfilling the dissertation requirement.
V_»igsertation Director Dr. Linda R. Phillips Date 3
STATEMENT BY AUTHOR
This dissertation has been subniitted in partial Mfillment of requffements for an advance degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.
Brief quotations from, this dissertation are allowable without special permission, provided the accurate acknowledgement of sources is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the copyright holder.
Signed: ^ 4
ACKNOWLEDGEMNTS
With sincere appreciation, I acknowledge the support, understanding, encouragement, patience, and guidance of Dr. Linda PhilUps, my comnoittee chairperson.
She has been very generous with her time and assistance during this doctoral scholarship.
With special thanks, I acknowledge the encouragement, support and contribution of my committee members. Dr. Anne Woodtii, Dr. Sandra Cromwell and Dr. Lee B. Sechrest.
Dr. Sechrest was the centered source providing guidance and consultation in advance statistics - structural equation modeling. His support and encouragement strengthened my beliefs that I could handle the EQS program mdependently, which I did it.
There are a number of people I need to thank: Dr. Pamela Reed and Dr. Ruth
Eskesen granted me general scholarships that financially supported this project; Dr.
Beverly Rosenthal who was always the first person to read and edit my dissertation draft; finends, who helped me to validate the translation of instruments and referred qualified recruiting representatives for this study; recruiting representatives, who so willingly assisted in identifying and recruiting available subjects for this study; and specially to the subjects who agreed to participate in this study.
Finally, I offer special gratitude to my family back in Taiwan, espedally my parentSr BQng-Po Hsueh and Mei-Tai Chen, for theur unconditional love, encouragement and endless financial support that sustained me during this long journey of learning here. 5
TABLE OF CONTENTS Page UST OF FIGURE. 8 LISTOFTABLES 9 ABSTRACT II
CHAPTER I: INTRODUCTION
Introduction 13 Background and Significant of Problem. 15 Family Caregiving-A Universal Valuable Sodai Resource 15 Issues Related to Family Caregivmg withm and across Cultures 16 Individual Versus Group (Ethnic) Di^rences 18 Lack of Attendon to the Largest Growing Fastest Minority. 20 Statement of Problem. 21 Significance of the Study. 22 Theoretical Orientation 23 Sodal Exchange Theory. 23 Role Theory. 24 Stress-coping Theory. 26 Summary. 30 Theoretical Framework. 31 Beliefs and Values 32 Subjective Beliefs of Traditional Filial Obligation J4 Subjective Feelings about Role Requkements. 35 Caregiving Appraisal 36 Perceived Burdens. 37 Perceived Rewards 38 Use of Copmg. 39 Consequences of Family Caregiving. .40 Caregivers' Health Status .40 Hypotheses .41 Summary. .43
CHAPTER 2: LITERATURE REVIEW
Subjiective Beliefs of Traditional Filial Obhgation .45 Concept of Traditional Mai Piety in Chinese Culture. .46 Research Findings 53 Conceptualizatioa and OperationalLzation 56 6
TABLE OF CONTENTS - Continued Page Subjective Feelings about Role Requirements 58 Research Findings. 58 Conceptualization and Operationalization 6 L Perceived Burdens 62 Research Rndings 62 Conceptualization and Operationalization 63 Perceived Rewards .72 Research Findings ,11 Conceptualization and Operationalization J5 Use of Coping J5 Research Findings .75 Conceptualization and Operationalization J1 Caregiver's Health Status 83 Research Findings 83 Conceptualization and Operationalization 85
CHAPTER 3: METHODOLOGY
Research Design 88 Causal Model Testing. 89 Modified Cross-cultural Approach. .93 Hypothesized Causal Statements in the Conceptual Model ^9 Sampling Procedures. 104 Instrumentation 109 The Revised Yang's Filial Values Subscale (YFVS-R) 110 The Revised Yang's Filial Affection Subscale (YFAS-R)— 113 Poulshock and Deimling Burden Inventory (BI) 113 Personal-Sodal Cost Subscale (PSCS) 114 Physical-Emotional Cost Subscale (PECS) 116 Economical Cost Subscale (ECS) 116 Picot's Caregiver Rewards Subscale (PCRS) 116 Behavioral-Oriented Subscale (BOS) 117 Cognitive-Oriented Subscale (COS) 118 Physical Component Summary Subscale (PCSS) 119 Mental Component Summary Subscale (MCSS) 120 Process for cultural equivalence for English Version 6istrument. 121 Data Analysis Plan 123 Summary. 134 7
TABLE OF CONTENTS - Continued
Page CHAPTER 4: DESCRIPTION OF SAMPLE AND RESULTS
Description of the Sample 135 Characteristics of Caregivers 136 Characterfetics of Care Redpients 139 Kinship Between the Caregivers and the Care Redpients 144 Caregiving Involvement 144 Results Related to Subjective Feelings about Role Requirements 149 Psychonaetric Properties of Measures 154 Description of Study Variables 164 Estimation of Correlations 166 Descriptive Statistics 167 Results of Evaluation of Model Adequacy. 171 Results of the Origmal Model Testmg 172 Evaluation of Poor Fitting and Justification of Modification 173 Causal Hypothesced Paths for the Efet Revised Model 177 Causal Hypothesued Paths for the Second Revised Model 178 Results of the Revised Models 179 Summary. 188
CHAPTER 5: DISCUSSIONS, IMPUCATIONS, LIMITATIONS, RECOMMENDATIONS AND CONCLUSION
Brief Discussion of Sample Profile 190 Interpretation and Discussion of Models Testing 192 Findings in Related to Theoretical Orientations 204 Implications 206 Limitations and Recommendations .211 Conclusion 213 APPENDIX A, 215 APPENDDCB 218 APPENDDCC .220 APPENDDCD 225 REFERENCES 251 8
USTOFHGURES
Page Figure L. Conceptual Firainework 33 Figure 2, The Original Model 101 Figure 3. Measurement Models 102 Figure 4. The Fist Revised Model 175 Figure 5. The Second Revised Model 176 Figure 6. The First Revised Model with Standardized Path Effects 183 Figure 7. The Second Revised Model with Standardized Path Effects 186 9
UST OF TABLES
Page Table I. Chinese among Asian Population m the Study Sites 106
Table 2. Age, Length ofTime in the United States and Educational Level of the Caregivers 137
Table 3. Characteristics of the Caregivers „..137
Table 4. House Income and Types of Occupation of the Caregivers 138
Table 5. Geographic Location of the Caregivers 140
Table 6. Characteristics of the Care Recipients 141
Table 7. Summary ofHealth Problems among Care Redpients 142
Table 8. Types ofHealth Problems among Care Recipients. 143
Table 9. Kinship between the Caregivers and the Care Recipients 146
Table 10. Types of Assistance Provided by the Caregivers 147
Table IL Descriptive Statistics of the Subjective Feelings about Role Requriements 149
Table 12. Firequendes of the Subjective Feelings about Role Requriements 150
Table 13. Summary of Reliability Assessment for All Measures 156
Table 14. Summary of The Range of Factor Loadings for Att Measures 157
Table 15. Zero-Order Correlations of the Study Variables 165
Table 16. Descriptive Statistics of Study Variables 168
Table 17. Summary of Average Off-diagnoal Absolute Standardized Residual for Each. Model 180
Table 18. Summary of Goodness-of-Fit Indices for the Original and Revised Models—181 10
LIST OF TABLES - Continued
Page Table 19. Standardized Direct, Indirect and Total Effects for the First Revised Model 184
Table 20. Standardfeed Direct, Indirect and Total Effects for the Second Revised Model 187 11
ABSTRACT
It is estimated that by 2025 the Asian population in the United States will be over
40 million; it will represent nearly 10% of the total U.S population. Currently Chinese immigrants are the largest Asian group and constitute 22.6% of all Asian immigrants. As the Chinese immigrant population is growing steadily, the Chinese elderly population will increase in the next decade.
Historically, "Hsiao" (Mai piety in English), rooted in Confucianism, teaches
Chinese people to pay respect to living parents, provides rules for culturally defined intergenerational relationships as well as family caregiving and putting family needs above individual interests. Among the Chinese, family caregiving is perceived as a duty, obligation, responsibility, and cultural virtue.
The purpose of the study was to test a theoretical model specifying how the interrelationship of factors including subjective beliefs of traditional filial obligation, subjective feelings about role requurements, perceived burdens, perceived rewards and use of coping affect Chinese caregivers' health in the United States. The aim of this study was to understand how caregivers' cultural background and the appraisal of family caregivmg affect caregivers' wellbeing.
Data firom 137 caregivers were used to describe sample profile and for model testmg. The original model did not have a good fit with the data. Two revised models were developed. Results suggest that data fit better with physical health, as indicated by
CFI= .91 and the x" / df ratio = 2.9, than than mental health model, as indicated by CFI= 12
.90 and the XI df ratio = 3.5. Findings, in both revised models, suggest caregivers' beliefs of traditional filial obligation is a primary predictor for caregivers' physical health.
Caregivers' subjective beliefs of traditional filial obligation affect caregivers' physical health not only directly but also indirectly.
Implications for nursing research, practice, education and theory construction, and policy making were addressed. Limitations and recommendations for research design and interpretation of model fit were described. Further studies are needed to compare differences in cultural influences among different ethmc groups, identify effective coping style and test the effect of intervention caregivers' health. 13
CHAPTER!
INTRODUCTION
According to the U.S. Bureau of the Census (1991), Asian American is the third largest minority group in the United States. In the past decade, Asian immigrants have been the fastest growing minority in the Um'ted States, swelling from 3.8million in 1980 to
9.7 million in 1996 (Statistical Bulletin, 1997). This population increased more than two fold in just 14 years and still continues to grow rapidly. It is estimated that by 2025 the
Asian population will be over 40 million; it will represent nearly 10% of the total U.S population (Day, 1996). In addition, the number of Asian elderly, age 65 and over, has more than tripled from 1980 to 1995 (Young & Gu, 1995).
Among the Asian immigrants, the Chinese were the earliest significant Asian immigrants and arrived in the United States between 1849 and 1882 (Dudley, 1997).
Currently they are the largest Asian group and constitute 22.6% of all Asian immigrants
(U.S. census, 1990). The Chinese population, including the elderly, has grown steadily and increased in visibility in the Um'ted States in the last decade. As a result, the Chinese elderly population will become even more visible than at present in the next decade.
For most people, a longer life implies a greater chance of acquiring one or more chronic illnesses and becoming disabled and dependent. In general, when individuals have a severe illness, they expect theur spouses or family members to visit, show concern, provide physically demanding care and take over chores related to daily living. L4
Consequently, as the number of aged people with impairments and disabiUues increases, then: requirements for care and/or support from family members are likely to increase.
For several centuries, "Hsiao" (filial piety in English), rooted in Confiidanism, has been the dominant ethical influence in the daily lives of Chinese people. Filial piety teaches
Chinese people to pay back and pay respect to living parents and the spirit of ancestors, and to put family needs above individual interests (Analects, 1639). It provides rules for culturally defined intergenerational relationships as well as family caregiving relationships
(Ho, 1996). Filial piety plays an important role in motivating family caregiving for the elderly (Wu, Li & Chang, 1997), espedally intergenerational family caregiving (Ho, 1996).
Thus, as the elderly Chinese population in the U.S. increases, issues related to family caregiving will become evident. Health care providers need to address the health care needs of the increasing numbers of Chinese caregivers who provide care to their elderly dependents. Understanding cultural factors that influence health problems relevant to
Chinese family caregivers will present a challenge to health care providers in the years ahead.
There has been no research that specifically addresses the relationship of subjective beliefs and perception of, and responses to, family caregiving and their relationships to the health status of Chinese family caregivers in the United States. The purpose of this study was to examme the interrelationships and determine the causal relationship, if any, of factors that contribute to caregivers' health status. Specifically, the study attempted to examine the impact of caregivers' subjective environments on caregivers' health through the mediated effects of caregiving appraisal among Chinese caregivers in the United L5
Stales. The caregivers' subjective environments are the caregivers' subjective bebefs and values of fffiai obligation and subjective feeling about role requirements. Caregiving appraisal includes perceived burdens, perceived rewards and use of coping options.
Background and Significance of Problem
In this section, the meaning and value of a universal human experience, family caregiving, is addressed. Issues related to family caregiving within and across cultures are described. Finally issues related to Chinese family caregivers in the United States are also presented.
Familv Caregiving-A Universal Valuable Sodal Resource
Family caregiving is a universal human experience (Fry, 1996) that has been recognized as a meaningful family value and valuable social resource, especially for the dependent elderly (Baines & Oglesby, 1992; Boland & Sims, 1996; PhiUips et al., 1994,
1996). In the United States, nearly 80% of the health care for the elderly is provided by family members (Baines & Oglesby). The care for frail elderly is often assumed by female family members, either wives or daughters (Brody, 1981; Brower, 1987; Deimling &
Poulshocfc, 1984; Given, 1998; Lang & Brody, 1983; Moss, Moss, Moles, 1985; Shanas,
1979; Zarit et al, 1980). Family caregiving is perceived, as a gender role for women and meeting parents' basic needs is a filial responsibility for an adult (Schoor, 1960).
In Oriental societies, family caregiving is perceived as a duty, obligation and responsibility (Seltzer & Li, 1996; Freed, 1990). The Chinese view family caregiving as a familial expectation, a moral obligation as well as a cultural virtue (Ho, L996). Family caregivmg is the "natural" extension of a woman's family role and the "naturaT' 16 responsibility of an adult son's role (Ho). Historically, family caregiving among the
Chinese has been highly regarded as a virtue and sodal resource, and perceived as a life long filial responsibility (Analects, 1963).
Issues Related to Family Caregiving within and across Cultures
Although family caregiving is highly regarded in many cultures, it is not without problems. In 1963, Grad and Sainsbury first identified burden as an important impact felt by family members who cared for dependent relatives. Since then, caregiving burden has been recognized as a common human phenomenon in family caregiving. Findings from the caregiving -iterature indicate that most family caregivers experience physical, emotional, social, finandal and/or relationship difficulties with their care receivers and/or other family members at some point while providing care. Family caregiving has been described as a stressful situation that is perceived as demanding and burdensome (Brody, 1981; 1985,
Brower, 1987; Demling & Poulshock, 1984; Given, 1998; Horowitz, 1985; Lang &
Brody, 1983; Masciocci, L978; Moss, Moss, Moles, 1985; Shanas, 1979; Zarit et al,
1980).
Similarly, among Chinese, a prolonged caregiving obligation imposes immense stress on the family caregivers, as more constant care is needed (Ngan & Cheng, 1992).
Caregiving burden has also been recognized as a significant phenomenon in Chinese family caregiving (Chou, 1999; Liu, 1990; Ngan & Cheng, 1992; Yeh, L992). The overall burden expenences mclude physical, emotional, finandal and sodal strams.
Recent studies have also disclosed that family caregiving can have a negative impact on the caregiver's health (George & Gwyther, 1986; Wright, Clipp & George, 1993). The 17 higher the caregiving strains^ the poorer is the health status of the family caregivers (Pratt,
Wright, & Schmall, 1987). This result is supported by Hale et al. (1987) that caregivers complained of a poorer health status than did non-caregivers. Unfortunately, the family caregivers' health was often neglected by both the family caregivers themselves and health care providers (Fengler & Goodrich, 1979). Family caregivers have been described as the hidden victims of family caregiving (Richard & Holicky, 1996). As a result, both family caregiving strains and caregivers' health are critical issues for both family caregivers and health care providers.
While family caregiving can involve problems, not all caregivers experience these to the same degree. In the last 20 years, with increasing awareness of the family caregiving experience and caregivers' well-being, studies have shown family caregiving involves not only negative but also positive experiences (Chenoweth et al., 1986; Given &
Given, 1988; Jones, 1995; Lawton et al.,1989; Picot, 1995). The consequences of family caregiving vary widely from person to person. Some family caregivers report more burden or health deterioration than others, whereas some caregivers maintain the same health or adapt over a long period without feeling a burden (Montgomery, Gonyea, 1985;
Poulshock & Deimling, 1984). Moreover, positive experiences such as feeling satisfied
(Jones, 1995; Lawton etal., 1989), self-worth (Given & Given, 1991; Phillips, 1988) and rewards (Picot) have been also described in studies.
In Chinese society, family caregiving is perceived as a life long obligation
(Mancim", 1989). Studies mdicate both perceived demands and rewards co-occur during the journey of family caregiving (Jones, 1995; Shyu, Archbold, & ImJe, 1998; Wang, 18
Shyu,Yao, & Wu,1998). The process of family caregiving has been described as a process of finding a balance point for caregivers' competing needs (Shyu, Archbold, Sc. Imie, 1998;
Wang, Shyu,Yao, & Wu,1998). Once caregivers learn to find a balance point among the competing needs, family caregiving becomes a satisfaction. These results not only raise questions of why some caregivers experience more negative consequences than others, but also why some caregivers experience little or no burden, or even positive consequences, despite participation in very demanding care for a long period of time. Is the variation of
family caregiving experiences an individual or group (ethmc) difference, and why does it occur?
Individual Versus Group (Ethnic) Differences
Some studies suggest that an individual's subjective factors (i.e., beliefs and
attitudes, images and self-perceptions) related to family caregiving, rather than
demographic factors, are significant predictive factors in determining individual responses
to the caregiving role (Lawton et al., 1989; Phillips, 1988; Picot, 1991; Poulshock &
Deimling, 1984). Poulshock and Deimling believe the subjective views about providing
care are a vital filter in determining the perception and consequences of family caregiving.
Phillips suggests that attitudes and images toward the caregiving role determine family
caregiving commitments and performances. Jones and Lawton and colleagues (1989)
found that positive feelings, sense of mastery and sense of obligations toward the
caregiving role were significantly assodated with the caregivers' satisfactions with their
caregiving role. Those findings suggest that family caregiving is a very subjective 19 experience, and individual difference in caregiving experiences may result from different beliefs, attitudes, images and self-perceptions related to family caregiving.
There is also evidence that an individual's ethnic background has the potential to affect his/her subjective responses; family caregiving experiences vary for people of different cultures (Goodman, 1990; Jones, 1995; Mui, 1992). In a study comparing role strains between black and white family caregivers, Mui (1992) found that although a high prevalence of role strain was present in both groups, whites reported higher strains than blacks (MuO- In another study of 76 Chinese family caregivers of elders with Alzheimer's disease, Fuh, Wang, Liu and Wang (1999) report that Chinese caregivers viewed family caregiving as their destiny and tended to leam to tolerate their problems eventually.
Similar results were also found in a qualitative study done by Jones (1995). In Jones' study, the theme of satisfaction was identified as the core category, although some negative themes such as conflict and parental control were also described. Jones beheves that the sense of indebtedness, obligation, reciprocity and paying respect rooted in Asian culture enables and motivates Asian caregivers engaged in long-term caregivmg.
Coincidentally, this was also indicated in Goodman's (1990) study that Asian caregivers still held a strong sense of duty in fflial obligation, interpersonal redprodty and paying respect to elderly parents.
In addition, how individuals respond to a situation or event that is perceived as stressfiil is determined not only by a person's appraisal of the difBcuIt situation or events but also by their resources and capacities to handle the difEculty. In other words, individuals' resources and capacities play a mediating role in how they respond to stressfiil 20 situau'ons. This subjective response is called coping Q^azarus & Folkman, 1984). Coping is a process of psychological and/or behavioral protective responses induced by stressfiil environmental demands. Use of coping strategies is inspired by available personal or social resources such as skill to cope, religious background and knowledge based on cognitive appraisals (Lazarus and Folkman, 1984). The avaflability and mobilization of coping resources influences the caregiver's level of depression (Morrissey, Becker, &
Rubert, 1990). Coping serves as a potential mediator of a primary stressful situation and the caregivers' psychological wellbeing (Haley et al., 1987). Considerable evidence indicates that caregivers' burden (Chou, LaMontaqne & Hepworth, 1999; O'Brien, 1993) and well-being (Henkle, 1994; Neundorfer, 1991) are mediated by use of coping. It is likely that an individual's ethmc background has the potential to affect their subjective responses (e.g., perceptions, actions and reactions to family caregiving role) because an individual's cultural background involves a shared distinctive history, values, beUefs and practices (PhiUips et al., 1996; Picot, 1991).
Lack of Attention to the Largest Growing Fastest Minority
Although the number of elderly in the United States is predominately non-Asian
American, the number of Asian elderly wiH likely increase more in the future as the Asian population grows rapidly (Statistical BuDetin, 1997). Consequently the need for care and/or support from family members among Chinese elders is likely to increase. Since nearly 80% of adult Chinese immigrants are foreign-bom (U.S. Census, 1990), it is likely that cultural influences such as beliefs and/or practice of filial obligations will influence
Chinese caregivers' response to and consequences of family caregiving. Recently a few studies have focused on Chinese family caregivers (Chou, 1999; Goodman, 1990; Jones,
1995; 1996; Ngan & Cheng, 1994; Yeh, 1992). However, no literature specifically addresses the relationship of subjective beliefs, perception of and responses to family caregiving, and their relationships to the health status of Chinese family caregivers in the
United States. More attention needs be paid to issues related to family caregiving among the largest and fastest growing Asian population, the Chinese.
Statement of Problem
Family caregiving has been perceived as a valuable social resource. However, caregiver burden is a common human phenomenon in family caregiving that has been recognized as a negative predictor of the caregiver's health. Studies on family caregiving and caregiver burden have demonstrated that: there are inconsistencies in responses to and experiences of family caregiving among individuals and groups (e.g., different ethnic groups). Recent studies disclose that individuals' subjective beliefs and perceptions serve as important filters in influendng subjective responses to caregiving roles and afiecting the consequences of family caregiving. There has been no research that specifically addresses the relationship of subjective beliefs, perception of and responses to family caregiving, and their relationships to the health status of Chinese family caregivers in the United States.
Little attention has been paid to the healthcare needs of the largest and fastest growing
Asian population, the Chinese. There is a lack of empirical data for understanding family caregiving among Chinese caregivers in the United States. There is a need to explore the
Chinese caregiver's filial beliefs and their relationship to the perceptions to providing 22 assistance and responses to family caregiving role, and their effects on Chinese caregivers' health.
Significance of the Study
Nursing is defined as a human science, a practice-oriented discipline, a caring discipline and a health-oriented discipline (Meleis, 1997). As a human science, nursing is concerned with the experiences of human beings within health and illness (Meleis). Meleis also described nursing as an interpersonal process involving the personal care of individuals as they experience illness within a therapeutic context. Since these experiences are shaped by history, significant others, politics, social structures, gender, and culture, nurses should be concerned with how these perspectives shape actions and reactions of human beings. This study was designed to gain more detailed understanding of caregiving for a selected population and to provide important information about caregivers from a specific ethnic group. Findings in this study may be significant in examining caregivers practice in other ethnic groups as well.
Increasing health care professionals' awareness of the caregiving experience among the Chinese population could improve the quality of care and quality of life for both caregiver and care-receivers. Since health care professionals can act as mediators and facilitators for family caregiving, this study provides empuical data to guide practice and minimize the gap between nursing theory and nursing practice. Information obtained from this study can be used to advance nursing knowledge about providing culturally competent sensitive care to Chinese caregivers. Rndmgs also have important implications 23 for policy makers to develop health service programs to benefit Asian family caregivers and develop effective programs to assist family caregivers firom other cultures.
Theoretical Orientations
To guide this study, a conceptual firamework is proposed based, on three theoretical perspectives: social exchange theory, role theory and stress theory. The substantive theoretical backgrounds of the conceptual firamework are described.
Social Exchange Theorv
According to sodal exchange theory, individuals seek to use exchanges to maximize rewards and minimize costs through social interactions (Cook, 1987). To maintain stable social relationships as well as a reciprocal norm, individuals are expected to return help to those who have helped them (Gouldner, I960). All sodal interactions
(Zabieksi, 1986) and interpersonal relationships (Tiden, 1986) are characterized by acts of giving and receiving, but the degree of reciprodty within relationships varies firom relationship to relationship (Rook, 1987).
Carruth (1996) suggests that the relationship between family members reflects a lifelong pattern of exchanges. Family caregiving is described as a redprocal relationship that is perceived as having costs as well as rewards (Rossi & Rossi, 1990). The rules concerning exchanges among family members are different fi-om those governing other social interactions (Clark & Mills, 1979). The exchanges among family members are voluntary or involuntary depending upon the family members' perceptions of each other's important needs (Robson & Thumher, 1979). Family members tend to contribute to each other's well-bemg and to derive benefits firom others in the family network over long 24
periods of time (Carruth, 1996). Exchanges among family members do not need to be the same, nor do they have to be given immediately (Rossi & Rossi). Usually, types of exchanges in caregiving include instrumental, financial and affective exchanges among family members (Rossi & Rossi, 1990). Some alternative exchanges may be developed among family members when a family member is no longer able to engage in the customary exchange relationship (Carruth, 1996). For example, whea an elder family
member is unable to help with household chores, some alternative types of exchange such
as expressing affection, warmth, caring, concern, Idndness, and appreciation to other
family members may be used to maintain reciprocal exchanges (Nolan &. Grant, 1989;
Qureshi, 1986; Ward, 1985). Once these alternative exchanges have been recognized as
contributions to perceived self-worth, reciprocal exchanges are perceived as fair or
balanced (Carruth, 1996). To maintain stable relationships among family members, the
acknowledged balance of sodal exchanges by family members who are involved in the
relationships is the key (Carruth, 1996). Social exchange theory provides a global
explanation of family interactions and helps explain why some caregivers experience little
or no negative consequences, despite participation in physically demanding care. If
caregivers perceive reciprocity as balanced over time, their perception of demands or
rewards is altered. However, social exchange theory does not explain variation within and
across cultures.
Rote Theory
Rote theory has been applied by recent researchers to cover the gap in sodal
exchange theory in understanding the process of role acquisition m family caregiving (Bull, 25
1990; Pepin, 1992; Phillips, 1988). Researchers suggest that family caregiving is an acquired role in response to changes in a family member's health (Bull; Pepin; & PhiDips).
Family caregiving is viewed as a process of role taking initiated by individuals who image being in the caregiving roles (Schumacher, 1995).
Based on role theory, an individual's role behaviors are highly determined by their prior socialization (Cook, 1987), which is the result of self-expectations in association with particular positional norms (Ktryker&Statham, 1985; Schumacher, 1995). An individual's role behaviors tend to blend self-expectations with social expectations in response to situated ongoing interactions (Ktryker & Statham). Shared beliefs and customs about family caregiving modify family caregiving role transitions within particular ethnic and cultural groups (Pearlin & Zarit, 1993). As a result of self-expectancy to meet moral and social obligations, an individual's actions and reactions are formed in the images of the self in fulfilling a role responsibility in relation to a particular social position. This approach draws attention to the idea that the responses and consequences of family caregiving vary in individuals who hold different beliefs and self-expectations toward their caregiving roles. For example, for Asians (e.g. the Chinese), family caregiving has been emphasized as a lifelong family duty, cultural obh'gation and responsibility (Seltzer & Li,
1996; Freed, 1990). It is likely that individual role expectations are shaped by the Chinese culture, which may influence the subjective responses and the consequences.
Role theory does provide an explanation of how an individual's subjective expectations and prior sodaiization are assodated with the commitments and consequences of family caregiving. Role theory, however, does not account for the 26 important influence of situational factors, which nmy impact the process of the caregiver role acquisition, which is addressed in stress-coping theory.
Stress-coping Theorv
Stress-coping theory assumes that both personal and situational factors play important roles in influencing an individual's perception, actions and reacuons to a stressful situation (Lazarus & Folkman, L980). It implies that both personal factors (i.e.,
beliefs and commitments) and predictable environmental factors (i.e., resources) influence
individual or group differences in perceiving and responding to stressfiil situations
(Lazarus & Folkman, 1984). This prenuse fills the gap in social exchange theory and role
theory.
According to Lazarus and Folkman (1984), there are four essential components in stress-coping theory: stress (stimuli), cogm'tive appraisal, mediators (i.e., coping) and consequences (i.e., responses or wellbeing). Stress is an inevitable result of living and it is
"a relationship between the person and the environment that is appraised by the person as
taxing or exceeding his or her resources and endangering his or her well-being (Lazarus &
Folkman, 1984, p.2I)." Stress is described as a phenomenon, which represents anything
and everything that could cause distress (Lazarus, 1966). While certain environmental
demands produce stress, different experiential events produce different behavioral and
physiological effects (Lazarus, 84). Although different experiential events or different
degrees of environmental demands may result in different degrees of emotional responses
and impacts on the physical and psychological well-being of an individual, the variation of
individual responses to comparable situations is determmed by an individual's cognitive 27 appraisal (Lazarus &FoIkman, 1980). In other words, individual sensitivities and vulnerabilities, and interpretations of stressful situations detennine what kind of reaction and to what degree they respond (Lazarus). For example, under comparable conditions, while one family caregiver may feel benefit, another may respond with anger, and another may respond with anxiety or guilt.
Cogm'tive appraisal refers to "mental judgment, discriminauon, and choice of activity based largely on past experience" (Grinker & Spiegel, 1945, p.l22). It functions as a mental judgment subsystem to regulate an individual's response to the stress (Lazarus
& Folkman, 1984). Cogm'tive appraisal involves two forms; primary appraisal and secondary appraisal. The first is related to unique changing relationships influenced by personal distinctive characteristics (values, commitments, styles of perceiving and thinking). The second is related to how those characteristics are associated with prediction and interpretation of encountered environmental stressor (Lazarus & Folkman).
In primary appraisal, the following question naight be raised: " Am I in trouble or being benefited from, and m what way?" and if in trouble," Can anything be done about it, and what is it?"(Lazarus & Folkman). Primary appraisal plays a crudal role in affecting an individual's interpretation and reaction to the stressor (demands). For instance, an individual's negative feelings could be a result of either inappropriate interpretations or appropriate responses to stressful situations.
Secondary appraisal uivolves the evaluation of primary appraisal and whether the stressor is interpreted as a. threat or a challenge. In secondary appraisal, questions raise such as "What coping options should I use and are they applicable?" In other word. 28 secondary appraisal involves assessment and determining use of coping options (Lazarus
& Folkman, 1984). The process of cognitive appraisal functions as a buffer system intervening between the encounter and the reactions, and the factors that affect the nature of the mediated relationships (Lazarus &FoUanan),
Within the process of cognitive appraisal, coping intertwines with primary and secondary appraisal functioning as an integratal part of the buffering system to mediate the encounter and the reaction (Lawton, 1989). Coping has been defined as thoughts and actions an individual uses to manage the stressful encounter of daily living (Lazarus &
Folkman, L984). Both cognitive appraisals and coping are the key factors in mediating or moderating the impact and consequences of stress (Lazarus and Flokman, 1980).
The stress-coping model (Lazarus & Folkman, 1984) has been used to study the process of family caregiving and to understand the individual differences in response to family caregiving, and the needs of family caregivers (Haley et al., 1987,1996; Lawton, 1989;
Pearlin et al., 1990). In 1987, Haley and colleagues proposed four components: stimuli
(e.g., caregiving demands), cognitive appraisal, mediators (e.g., coping) and response
(e.g., manifestation of stress) to develop their model, which is similar to that denoted by
Lazarus and Fo Demands (1984).
In Haley and colleagues' model, the individual's buffering process, which focuses on evaluating and using environmental resources, is believed to be an important predictor for the responses and consequences of family caregiving. It is suggested that the higher the level of perceived stress, according to cognitive appraisals, the greater the negative experiences, which include physical problems, feeling of strams, social isolation, and 29 financial difficulties (Haley et al.^ 1996). This is supported by Poulshock and Deimling's
(1984) findings that an individual's subjective views about the caregiving context play important roles in influencing the emotional response of caregivers, although situational factors (e.g., level of impairments or caregiving involvement) have a high potential to create a stressful situation. In another word, an individual's perceptions, actions, reactions, which determine the consequences of family caregiving, depend on how personal factors (e.g., the subjective beliefs, self-expectations, commitments and ability to cope) intertwine with situational factors (e.g., available resources to cope successfully)
(Haley et a.; Poulshock & Deinding). Coping, as an integratal part of the individual's buffering process, is also regarded as a significant mediating factor influencing the outcome of family caregiving (Haley et al.).
While Haley and associates' (1987) model emphasizes the caregivers' cognitive appraisal and use of coping, another modified stress-coping model, which focuses on the background and context of the stress and the mediators of stress was developed by Pearlin and colleagues (1990). Pearlin and colleagues believe that the background and context of the stress, stressor, mediator of stress and outcome of stress are essential components for understanding the process of family caregiving experience. They suggest that life events such as family caregiving involve ccumulative demands of care over time, which may devour the caregiver's physical energy and/or psychological, sodal and economic resources. Consequently these may interfere with the caregiver's social role and handicap personal wellbeing (Pearlin et al., 1990). The caregiver's demographic characteristics including age, gender, education, occupation, economic situation, length of caregivmg and 30 caregiving history play important roles affecting the outcome of family caregiving (Pearlin et al.). It is suggested that the background and context of stress as reflected by these
variables should be included in family caregiving studies (Pearlin et al.).
To enhance understanding of process and experiences of Chinese family caregivers for this
study, propositions from social exchange theory, role theory and stress-coping theory
were included in the conceptual underpinnings of this study. Premises from social
exchange theory and role theory were applied to denote the family caregiving
relationships. Several propositions derived from the works of Poulshock and Deimling
(1984) and Haley and colleagues were used to postulate relationships among selected
study variables for the theoretical framework, particularly those related to caregiver
characteristics of attributed subjective views toward family caregiving. Other indicators,
level of impairment and the relationship to the care receiver proposed in Poulshock and
Deimling's paradigm and the caregiver's demographic characteristics and length of
caregiving suggested by Pearlin and colleagues were used as demographic information in
this study.
Summary
Sodal exchange theory describes the structure and process of the interaction
between the caregiver and the care receiver. Role theory explicates the process and
influence of the caregiving role on the family caregiver. Stress-coping theory provides an
overall framework to integrate sodal exchange and role theory into the theoretical
orientation in this study to depict the structure and process of family caregivrag and its
impact on the caregivers. Each theory emphasizes the cognitive elements m maintaining good family caregiving relationships with little or no negative response, despite participation in a lot of physically demanding care. Each theory contributes to analyzing the structure, process and influence of family careigving. An incorporation of these three theoretical perspectives provided a complete conceptual framework: to enhance the explanatory power of this study.
Theoretical Framework
The theoretical framework for this study was derived from three theoretical perspectives: social exchange theory, role theory and stress theory. The structure of the framework was modified from the caregiving burden model proposed by Poulshoch and
Deimling (1984) and the stress-coping model developed by Haley and colleagues (1996).
The relationships of variables proposed within the framework were based on propositions from the following works (Haley et al., 1987; BQein, 1989; Lazarus & Folkman, 1984;
Phillips et al., 1988, 1996; Pico, 1991; Poulshock & Deimling, 1984; Rossi & Rossi,
1990). The main proposition for the framework was from Poulshock and Deimling's
(1984) and Phillips and colleagues* (1996) works that individuals' feelings about the family caregiving role result from highly individualized responses to the overall caregiving role and the specific caregiving context. Individual responses such as perceived demands and rewards are determmed by an individual's subjective filter (i.e., beliefs, values, attitudes and judgements of family caregiving). In addition, use of coping has shown mediating effects on family caregiving (Haley et al., 1996). Use of coping is determined by predicable environmental situations such as resources OLazarus & Folkman, 1984).
Use of coping is determined by available resources and influenced by mdividual beliefs. 32 which are shaped by ethnic background (Picot, 1991). Since individual beliefs, values and experiences are formed by social structure and prior socialization (e.g., history, significant others and culture) (Meleis, 1997), cultural influences are important causal antecedent factors, which shape the perceptions, actions and reactions of family caregivers. There is likely an association between the meaning of family caregiving perceived by family caregivers and their behaviors (Picot, 1988).
Figure 1 is the conceptual model for this research specifying the constructs and concepts and their relationships. On the top of the figure, three categories (antecedents, responses and outcomes) are used to indicate the process of family caregiving experiences.
Three constructs, (1) beliefs and values, (2) caregiving appraisal and (3) consequence of caregiving, were proposed to demonstrate the process and influence of caregivers' subjective beliefs and values on their responses to and the outcome of family caregiving.
Six concepts (variables) relevant to family caregiving were proposed. The circles represent latent variables and the squares mean manifest variables. The six variables are
(1) subjective behefs and values in filial obUgation, (2) subjective feelings about role requirements, (3) perceived burdens, (4) perceived rewards, (5) use of copmg options, and
(6) health status. For conceptual clarity, the defimuons and relationships among variables are discussed as foQows.
Beliefs and Values
The construct of subjective beliefs or values is defined as a worldview that specifies the form of conduct of a person's particular interests, which are assodated with
the anticipation of a special status to those who fiilfill the desied behavioral patterns Antecedent Responses Outcomes
Construct level Beliefs and Values Caregiving Appraisal Consequence of caregiving
Concept level Subjcciiyc Perccivcd Beliefs of Burdens Tradillonal Filial Obligation Use of Coping
Health Status
Subjective Perceived Feelings about Rewards Role ReonirementK
Figure 1 Conceptual Framework
LO 34
(Goldschmidt, L971). Subjects' predominant beliefs or values are shaped by past socialization and present experiences, guide everyday speech and conduct and make for who they are and how they should respond (Peck, 1990). For example, caregiving is initiated by identifying the family caregiving role, which is determined by the individuals' subjective beliefs and values (Phillips, 1988). The individuals' subjective beliefs and values not only determine the process of role identity but also provide the avenue to enable an individual to engage in reciprocal activities (Phillips). Therefore, subjective beliefs and values have a crucial role in influencing the caregiver's perceptions, actions and reactions to family caregiving. Since individuals of various cultural backgrounds may have different beliefs and values, which affect group norms and behaviors (Rosen, 1988), it is necessary to examine the family caregiving experience from a cultural perspective. For the purpose of this study, the construct of subjective beliefs is reflected by two concepts: (I) subjective beliefs and values of traditional filial obligation and (2) subjective feelings about role requirements. The two concepts are described as follows.
Subjective Beliefs of Traditional Filial Obligation
The concept of subjective beliefs of traditional filial obligation is defined as the degree to which the subject ascribes to values about traditional fiUal obligation in Chinese culture. According to the literature review, cultural values make a difference in the acceptance of family caregiving by caregivers (Nydegger, 1986). The traditions and values of the society are nnportant factors affecting the willmgness and ability of caregivers to assume long-term responsibility for dependent elder family members (Pearlui
& Zarit, 1993). For the Chinese, Ho (1996) suggests that traditional cultural values such 35 as fUial obh'gations are the primary driving forces in directing and motivating family caregiving. Therefore, filial obbgation rooted in Chinese culture is assumed to be an unportant influence on the subject's subjective perceptions and responses toward the family caregiving role in this study. In this study, subjective beliefs and values regarding traditional filial obligation was used to explore how it affects family caregiving experiences among Chinese caregivers, and to test its predictive relationship with other study variables.
Subjective Feelings about Role Requirements
The concept of subjective perceptions about role requirements is defined as the subjects' affective or emotional responses about fiilfilhng family caregiving responsibilities in this study. Poulshock andDeimling (1984) believe that the individuals' subjective feelings about role requirements influence the caregivers' appraisal of the family caregiving responsibiUties. They indicate that caregivers' caregiving experiences depend on their individualized perceptions and responses to specific caregiving contexts such as tasks assodated with specific impairments of the care receivers. The impact of family caregiving on caregivers in similar situations is highly individualized (Poulshock &
Deimling). They suggest that the stressfiilness of fulfilling family caregiving responsibilities has a significant influence on caregivers' responses^ especially perceived burdens (Poulshock &. Deimling).
According to Poulshock and Deimling, the subjective feelmgs about providing assistance concerns the degree to which the caregivers react emotionally toward doing specific tasks related to physical dependence or mental incapacity of the care receiver. 36
Specifically, the concern is the degree to which caregiving tasks are viewed as troublesome, difficult or taxing. In Poulshockand Deimling's study, the subjective feelings about providing assistance was proposed as an important antecedent factor in determining perceived burdens. In this study, subjective feelings about role requirements is consistent with Poulshock and Deimling and was viewed as an antecedent factor for testing its predictive relationship with other study variables.
Caregiving Appraisal
The construct caregiving appraisal is defined as the degree to which the subject evaluates caregiving interactions as burdensome or rewarding, and responds through coping strategies. Appraisal may mediate relationships between the construct of beliefs and values and the construct of consequences of caregiving. In this process, all three theoretical views are involved. For example, according to sodal exchange theory, within a caregiving interaction, both the caregiver and care receiver attempt to maximize rewards and to minimize costs through reciprocal exchanges. When there is balance in reciprocal exchanges, perceived rewards are present. In contrast, perceived demands occur when reciprocal exchanges are perceived as unbalanced.
From the role theory perspective, it is believed that during a cargivmg interaction the caregiver has a moral and social obhgation directing the caregiving role that is determined by self-expectation and social expectation (Hem, L989). This has been seen among the Chinese. For Chinese, the concept of filial piety has been mstaUed in adult children- Caring for dependent elders is highly regarded as an unquestioned course among the Chinese (Yu et al., 1993). Caring for parents and paying respect to elders is a social 37 and moral obligation for Chinese. When the expectation of caregiving role is congruent
with the actual role performance, one would experience sense of rewards or glory; otherwise, one may experience burdensome or suffering (Klein). Both perceived burdens and rewards are the results of cognitive appraisal and can co-occur. Both play important
roles in predicting the consequences of family caregiving (Haley et al., 1987).
From the stress-coping perspective, when the perceived burdens or demands exceed the
perceived rewards, coping is induced (Haley et al, 1996). Coping is an important factor in
moderating stress and sustaining health protective behaviors (Haley et al.). Use of coping
has been viewed as an important driving force in the ability to handle stressful situations
successfully (Lazarus & Folkman, 1984). It enhances the caregiver's ability to manage
family caregiving and mediates the consequences of family caregiving (Haley et al, 1996).
Thus, the construct of caregiving appraisal is reflected by three concepts: perceived
burdens, perceived rewards and use of coping options. In the following section, a brief
description for each concept is given.
Perceived Burdens
Perceived burdens refer to the degree to which the subject perceives actual and
potentially negative impacts on his/her own life in response to the family caregiving role.
Burdens include physical, emotional, sodal, financial, and interpersonal relationship
difificulties (Given, 199l;Poutshock;andDeimIing, 1984). Caregivers'negative emotional
responses to family caregiving are determined by how they feel about providing assistance
with tasks related to the specific needs of the elder care receivers as negative demands
(Poulshock and Deimling). Caregiver's negative perceptions about providing family 38 caregiving have been recognized as a negative predictor of the consequences of family caregiving. In this study, the concept of perceived burdens was assumed to be a key negative predictor of the consequence of caregiving.
Perceived Rewards
Perceived rewards are concerned with the degree to which the caregiver perceives
"positive subjective feeh'ngs and objective changes in his/her own life" (Picot, 1991).
Literature shows that there is a significant relationship between perceived rewards and the use of certain coping strategies such as positive thinking and problem analysis (Picot).
Positive reactions to family caregiving have been connected to the caregivers' positive sense of obh'gations (Klein, 1989). This has been supported by Jones (1996) who found that traditional filial obligation motivates Asian caregivers to engage in long-term family caregiving. Finally, traditional filial obligation enables them to perceive a feeling of being satisfied. Perceived rewards were also associated with caregivers' religiosity (Picot,
1991). En a cross-cultural study of black family caregivers, Picot (1988) found that those who had feelings of satisfaction about providing family caregiving believed they would receive a heavenly reward for providing care to the care receiver. In addition, it is suggested that perceived rewards have potential effects on the consequences of family caregiving, for example, psychological well-being (Picot, L995). In this study, the concept of perceived rewards was identified key positive predictor of the consequence of caregiving. 39
Use of Coping
Use of coping is defined as the extent to which, the subject uses behavioral and/or cognitive reactions in seeking avaflable resources for managing stressful situations
(Folkman, 1988). According to FoDcman and Lazarus (1980), the degree to which caregivers employ coping as well as use type of coping options affects caregivuig burden.
Considerable literature indicates that coping is associated with caregivers' burden or well- being (Chou, LaMontaqne and Hepworth, 1999; Haley et al., 1987; Henkle, 1994; Hull,
1992; Irvin& Acton, 1996; Lazarus & Folkman, 1984; Neundorfer, 1991; O'Brien, 1993),
However, findings are inconsistent among those studies. While Chou, LaMontaqne &
Hepworth (1999) found that use of effective coping directly affected burdens in a sample of 150 Chinese caregivers in Taiwan, others indicate that coping is a mediator influencing
burden (Hull, 1995; Irvin& Acton, 1996; Lazarus & Folkman, 1984; O'Brien, 1993) and
health (Haley etal., 1987; Henkle, 1994; Neundorfer, 1991). In a longitudinal, qualitative study of 14 hospice families caring for their dying relative at home, Hull reported that use of effective coping was associated with burden. O'Brien indicated that use of coping was
significantly positively related to caregiving burden in a sample of 20 spouse caregivers of
multiple sclerosis. Henkle reported coping was a mediator influencing caregivers' well-
being in a study of family caregivers of homebound older adults. Neundorfer discovered
that caregivers' health was affected by use of coping in a study of 60 spouse caregivers of
persons with dementia. In this study, the concept of use of coping is proposed as a
mediator to test its efect on caregivers' health status and its relationships with other study
variables. 40
Consequences of Family Caregiving
The construct of consequences of caregiving is defined as the outcome of family caregiving. The outcome of family caregiving was evaluated by measuring the overall health status of the caregivers. According to the World Health Organization [WHO],
1974), health is a state of complete physical, mental and sodal wellbeing and not only the absence of disease and infirmity (p.l). Health is also defined as a state of optimum capacity of an individual for the effective roles and tasks for which she/he has been socialized"(Parsons, 1981, p,69).
Caregivers' Health Status
Caregivers' health status refers to the degree to which the subjects perceive their
own state of complete physical, mental and social well-being. Health status enables
him/her to perform daily roles, tasks and activities (Parsons, L981; WHO, 1974). George
and Gwyther (1986) believe the well-being of caregivers is a prime indicator in identifying
family caregivers' needs. They indicate it is hard for caregivers to maintain their own
health while caring for their ill family members. Recent studies have shown that the
caregivers' health status is significantly associated with caregiver burden (Maserang, 1992;
McDonald, L995; Mui,. 1995). Researchers believe family caregiving is causally tied to the
caregivers' health (Phillips & Rempusheski, 1986; Cameron, L990). Maserang reported
that the more burdens the caregivers feel, the greater depression the caregivers experience.
Therefore, caregivers' health status becomes a major concern for health care providers.
Kirkley,^ Griffin, McLintock and Ng (1998) suggest an individual's general health status is
the most important outcome measurement for health care, because it is a significant 41 indicator for assessing the benefits of treatment (Maciejewski, 1997) and evaluating the consequences of stressftil encounters such as family caregiving (Karen& Deborah, 1995).
In addition, studies show the caregivers* weE-being is related to use of coping strategies
(Jivanjee, 1994; Ray & Ritchie (1993). Use of coping not only mediates the feelings of burden but also influences the caregivers' physical and psychological health (Fink, 1992).
However, while Jivanjee believes effective coping options can enhance caregivers' well- being, Aldwin and Revenson (1987) indicate that there is no clear consensus regarding the effect of use of coping options. Therefore, the concept of health status of the family caregiver was defined as outcome variables and proposed to test its relationship with other variables in this study.
Hypotheses
The main purpose of this study was to test a theoretical model speci^g how the interrelationship of factors including subjective beliefs and values of traditional fiUal obUgation, subjective feelings about role requirements, perceived burdens, perceived rewards and use of coping option affect the health of Chinese caregivers living in the
United States. The aim of this study was to identify causal relationships among factors that contribute to caregivers' health status. Specifically, five hypothesized causal paths were proposed to examine the impact of subjective beliefs of filial obligation on caregivers' health through the mediated effects of perceived burden, perceived reward and use of coping options among Chinese caregivers in the United States. In the following hypotheses, direct and indirect relationships between variables are addressed. 42
Direct Relationships
Hi. Caregivers' health is directly positively affected by subjective beliefs of filial obligation, use of coping and perceived rewards, and directly negatively impacted by perceived burdens.
H2. Use of coping is directly positively influenced by perceived burdens.
H3. Perceived burdens are directly affected by subjective beliefs of filial obligation.
BI4. Perceived rewards are positively influenced by subjective beliefs of filial obUgation and negatively affected by subjective feelings about role requirements.
H5. Subjective feelings about role requirements are negatively influenced by subjective beliefs of fiUal obb'gation.
Indirect Relatioships
HI. Caregivers' health is indirectly affected by subjective beliefs and values of traditional filial obligation through the mediated effect of perceived burden.
H2. Caregivers' health is indirectly affected by subjective beliefs of traditional filial obligation through the mediated effects of perceived burden and use of coping.
EI3. Caregivers' health is indirectly affected by subjective beliefs and values of traditional filial obligation through the mediated effect of perceived rewards.
EI4. Caregivers' health is mdirectly affected by subjective beliefs and values of traditional filial obligation through the mediated effects of perceived rewards and subjective feelings about role requirements. 43
Summary
Asian Americans are the third largest minority group in the United States. Among the Asian immigrants, Chinese are the largest Asian population, consisting of 22.6 of all
Asian immigrants (U.S., Census, 1990). With the Asian elderly population dramatically growing in the Um'ted States, the requirement for providing family caregiving is increasing. This requirement is even more critical among Chinese, since, historically, the concept of filial piety, rooted in Confiiciamsm, has been dominant. As a result, family caregiving is a culturally imperative and a family obhgation.
Although family caregiving is regarded as a meaningful family value and a valuable social resource, the burden produced by caring for dependent elders were documented in the literature in the Um'ted States. Issues and factors relevant to family caregiving have also been explored firom many different perspectives to understand the process of caregiving in the United States and across cultures. However, little attention has been paid to specifying how factors such as the caregiver's cultural background affect the appraisals and consequences of family caregiving.
A theoretical model derived from sodal exchange theory, role theory and stress- coping theory was proposed to understand how the caregiver's cultural background affects the appraisals and consequences of family caregiving among Chmese Caregivers m the Um'ted States. Within the model, six concepts (variables) were proposed-
Interrelationships among the six variables including subjective beliefs and values of traditional filial obhgation, subjective feelings about role requirements, perceived burdens. 44 perceived rewards and use of coping affect on the caregivers' health were specified and described. Hypothesized causal relationships were defined and addressed. 45
CHAPTER2
LITERATURE REVIEW
In this chapter, a review of literature about key variables is presented. The key variables are: subjective beliefs of traditional glial obligation, subjective feelings about role requirements, perceived burdens, perceived rewards, use of coping and caregivers' health status. Demographic factors related to the characteristics of caregivers and care receivers, and caregiving situation are addressed.
Subjective Beliefs of Traditional Filial Obligation
According to Peck (1990), individuals' beliefs and/or values are formed within cultural contexts such as family and society, and are shaped by past socialization and present experiences (Peck, 1990). Individuals' beUefs and/or values affect group norms and behaviors (Rosen, 1988). For example, role requirements, which families use to deal with threats to life, health, wealth, or cherished relationships (Turner, 1970), often depend upon cultural norms and values (Brown, 1990). Although role requurements highly depend on cultural norms and. values, role strains or conflict may occur, when individuals are not ready for, or do not desire to assume a role assignment (Brown, 1990). Therefore, to understand why people do what they do, it is unportant to understand the meaning of relevant cultural norms and values, how they influence people and how people perceived them.
Since fflial is a strong culturally prescribed belief and value in Chinese culture, to accurately describe the caregiving experience, it is necessary to explore the meanmg of 46 traditional filial obligation in the Chinese culture (e.g., how Glial piety influences Chinese family caregiving? and how Chinese caregivers perceive the caregiving role?). Sung
(1995) suggests that understanding the concept and value of filial piety from a historical view is essential to validly measure the concept of fih'al obUgation. In the following sections, the social-cultural background of traditional filial obligation in Chinese culture and current studies on family caregiving experiences among Chinese are described.
In this section, information relevant to the Chinese social-cultural background of traditional fih'al obligation including the concept of traditional filial piety, and cultural factors and cultural values related to family caregiving is presented. Research findings, and conceptualization and operationalizationof subjective beliefs of traditional filial obligation are presented.
Concept of Traditional Filial Piety in Chinese Culture
Filial piety is the central theme of Chinese cultural values (Bush, 1970). The prime attributes of classic fih'al piety include I) dedication and loyalty to family, 2) respect, love, and taking good care of parents, 3) protection and glorification of parents, 4) obedience to parents and role defim'tions, and 5) worship of ancestors (Bush; Die, 1995, L998; Sung,
1995). In the Analects of Confucius, it says that "When parents are alive, serve them according to the rule of propriety. When parents die, bury them according to the rule of propriety, and sacrifice to them according to the rule of propriety" (Analects, 2:5, Chen,
1963, p.I3). Confucius said a filial son should behave in such a way that his parents not only get enough to eat but they also get respect (Analects, 2:6, Waley, 1938). Confucius suggests that support alone is insufficient. A respectful and warm attitude is required to 47 support and care for parents (Analects, 2:6, Waley, 1938). Confucius also said, "A. young man's duty is to behave well to his parents at home and to his elders abroad, to be cautious in giving promises and punctual in keeping them, to have kindly feelings toward everyone, but seek the intimacy of the Good" (Analects, 1:6).
Furthermore, the Book of Rites, one of the earliest books describing the concept of filial piety in Chinese culture, also emphasizes, "A filial son should bring no dishonor to his parents" (Confucius, 1982). Obedience, loyalty and glorification are expected towards parents (King & Bond, 1985). BCstorically, the respectful attitude was expected toward all elderly (Sung). Mencius, a follower of Confucianism, states "As I care for my own parents, I care for other elderly persons as well" (Lang, 1946, p.lO-LL). The ideology of
"respect for the old" is applied within the family as well as in society. Harmony results firom a respectful and warm attitude between one generation and another. Mencius believes that if the whole family is harmoruous, the society will be harmonious as well.
Therefore, an individual's cultural background plays an important role in influencing human belief systems and behaviors (Chu, 1985).
In summary, individuals' behaviors are highly determined by prior sodalization, which takes place in a cultural context such as family and society (Cook, L987).
Individuals' cultural background piays an important role in shaping human belief systems and behaviors (Chu, L985). To understand why people do what they do, it is essentfal to understand their cultural influences and cultural values. 48
Cultural Factors Influencing Filial Obligation
Culture refers to ways of living and social interactions built by a sodal group of human beings, created by learned patterns of behavior, beliefs and values, shared by individuals, and transmitted from one generation to another (Walsh, Smith, Mprales, &
Sechrest, 1999). Culture is the Icnowledge of values, traditions and practices, which are largely standardized and encoded in language and daily lite (Fry, 1996) and stored in people's mind. Culture influences people's behaviors, roles, and relationships (Phillips et al, 1996).
Historically, there are two important cultural factors influencing Chinese family values and values about filial obbgation. The two cultural factors are Confiician moral and ritual doctrines and traditional religious beliefs. Confucius was the first person to propose a moral obligation as well a socially approved virtue of "Ibiao" (filial piety in English) to teach Chinese people how to pay respect to living parents and the spirits of ancestors
(Waley, 1938). In answer to his student's question about how to show filial respect toward parents, Confucius said the first thing we should do is not spoil and injure our body, skin and hair because they are the gifts fi:om our parents (Waley, 1938). This message not only indicates that life itself is the &:st important gift a child receives from his/her parents, but also implies that life is the first and greatest debt a child owes parents
(Sung, 1995). Support and care for parents are viewed as an unquestionable responsibility and children should do it with apparent pleasure (Sung).
In addition, Confiidus believed that a human is an mteractive being, who can't exist alone (Hu, 1919). The family is an important sodal institution, as well as the 49 fundamental unit of society, and individuals are bound by the need to place family interests above their own (Analects, 1963, Chen). Family life should be based on mutual assistance and interdependence throughout life (Analects; Lin, 1982). Chinese pay special attention to teaching the value of filial piety through the family, school and media. Children also observe the interactions between parent-grandparents (Die, 1998). The notion of filial piety is deeply rooted in peoples' lives and transmitted &om one generation to another.
Chinese children have been sodalized to perceive filial responsibility as a life long obligation. Consequently, collective-interdependence and family values have always been an integral part of Chinese culture.
Another predominate Chinese cultural factor is derived from traditional religious beliefs such as Buddhism, Taoism and traditional folk religions. In a study of religions and the Chinese in Taiwan, Bush (1970) indicates that" Hsiao" is the center to all Chinese reUgions. Although thoughts and activities of "Hsiao" are emphasized differently in the various religious perspectives, none of them stray from the concept of "Hsiao" (Bush,
1970). While Buddhism proposes an afterlife, which reinforces the commitment to good deeds (Bush), Taoism emphasizes stability and harmony, and traditional folk religious activities inspire family solidarity and social cohesion (Bush; Chu, 1985). All religious beliefs strongly influence the Chinese way of thinking, acting and living (Lee &. Sun,
1995). For example, the notion of the Buddhist afterlife implies that a person's actions in this life determine what will happen to him a future existence (Bush). Thus, one must, by his/her own deeds while alive, build up good deeds (e.g., showing filial respect to parents) for his/her next life. Bush suggests that it is likely that this notion reinforces Chinese 50 people's comnutment to good deeds (e.g., showing fiUal respect and engaging in caregiving responsibility).
Moreover, Taoism and folk religions emphasize that harmony within family is the
&:st priority for a stable society. This harmonious value is embodied in the respectful and warm attitude that is shown toward parents (Chu, 1985) and expressed in the practice of co-residence in the extended family (Lee & Sun, 1995).
Cultural Values Relevant to Chinese Familv Caregiving
Holrody and Mackenzie (1995) beheve that cultural values direct the meamng of family caregiving and influence the propensity of Chinese people to engage in family caregiving. To understand the meamng of Chinese caregiving, it is helpful to identify what and how cultural values are relevant to family caregiving. According to the literature
review, five fundamental cultural values relevant to family caregiving were identified.
These five cultural values are: (a) collectivism or interdependence, (b) intergenerationai caregiving, (c) life long self-sacrifice to family and family interests, (d) the abffity to endure existential difficulty, and (e) gender role, kinship role and birth order responsibility
differentiation.
Collectivism or Interdependence
In Chinese society, an individual's life and behaviors are tied to the collective
nature of the family system (King & Bond, 1985). Chinese family life is based on mutual
assistance and interdependence throughout the life span (Lin, 1982). Traditionally the
Chinese family is an extended or stem family, which is a group of related people, living
close by, and firequently interacting and exchanging mutual aid. Components of the 51 extended family comprise three generations including a married son with his wife and their children living with his unmarried siblings and parents or multiple married, sons with their
wives and children living with their unmarried siblings and parents.
Historically, a reciprocal aid system operated within the extended kinship network.
A variety of exchanges (e.g., babysitting, financial aid, mutual help with field work during
harvest season, personal advice, and moral support) commonly occurred within the
reciprocal aid system. As a consequence, a strong sense of obligation, responsibility,
family bonds, and interdependence to one's family were regarded as virtues (King & Bond,
1985). Even today, the concept of collectivism or interdependence still dominates in the
Chinese culture. For example, in Taiwan, many married adult children are still willing
residents with their parents or live near by, so they can support each other (Wu, Li &
Cheng, 1997).
fntergenerational Careeiving
Chinese children are obligated to respect, support and care for their dependent
elderly family members throughout life, particularly their parents (Lin, 1982). Confucius
stated, "be loyal to your ruler and fiKal to your parents." Intergenerational caregiving is
the central premise of filial responsibility. Traditionally Chinese elders expect to receive
assistance and support firom then: children, espedaHy firom the oldest sons (Gates, 1987;
Lin, L982). Even today, many Chinese elderly still believe that living with a married
daughter is against the sodal norms. It makes them lose face (Lee, Li, & Chang, 1995).
For Chinese, a married daughter belongs to her husband's family and should obey the 52 customs of her husband's family because she is no longer considered a member of her original family (Lee, Li, & Chang).
Chinese elders prefer to rely on family for support rather than on friends or formal support systems (Lee, Li, & Chang, 995). Emotional and financial supports are the assistance received most often from Chinese adult children (Die, 1995). Ho (1996) believes that intergenerational caregiving is still highly prevalent in Chinese culture. This is supported by Wu, Li and Cheng (1997) who found that in Taiwan, about 2/3 of elderly parents live with their adult children, espedally married sons. In general, many Chinese adult sons still prefer to live with parents (Lee, Li, Chang, 1995).
Life Long Self-sacrifice to Family and Familv Interests
Confucius stated, "As parents advance in age, a son should rejoice in their long life" (Analects, 4:18-21, Chen, 1963). Support for parents is expected to be a life long self-sacrifice for fulfilling fiUal responsibility (Die, 1998). Chinese children are taught to sacrifice their physical, social, financial and emotional interests for their parents' wellbeing
(Chow, 1983). Individuals are expected to place family interests above theur own (Chow).
A son who does not support his parents is considered to have broken the unwritten law of tradition.
AbiUtv to Endure Existential Difficulty
Among Chinese, the concept of filial piety is based on two ftindamental ideas: (a)
Confuciam'sm, the necessary and absolute sacrifice to family, and. (b) the supernatural power of traditional religious beliefs. In Confudam'sm, family caregiving is perceived as an unquestionable responsibility as well as a way of paymg back. It is likely that according 53 to traditional religious beliefs, providing family caregiving is perceived as a way to build up good deeds and credits (e.g., showing filial respect to parents) for his/her next life.
Responsibility Differences in Gender Rote. BCinship Role and Birth Order
As in many cultures, women are the gender of choice for becoming caregivers
(Holroyd & Mackenzie, 1995; D'Andrade & Strauss, 1992). Family caring behaviors are also determined by moral obUgation of kinship (D'Andrade & Strauss, 1992). For the
Chinese, in addition to the moral obligation of gender role and kinship role, birth order is traditionally involved in family caregiving (Holroyd & Mackenzie). Historically among
Chinese, the oldest son is expected to live with and provide care for his elderly parents
(Holroyd and Mackenzie), but the caring responsibility is delegated to the oldest son's wife when he marries (Ikels, 1983; Wolf, 1968). The oldest son's wife is the first source of support for family caregiving (Mackenzie & Holroyd, 1994).
Research Findings of Familv Caregiving Experiences among Chinese
Filial obligation is highly regarded among Chinese (Ho, 1996). Recent studies have shown that beliefs about filial obligation are negatively associated with caregiver burden among Chinese caregivers (Chou, LaMontagne, &Hepworth, 1999). Phillips
(1988) suggests that the demands and trends of modem society shape family caregiving experiences. Finch (1989) indicates that limiting conditions (e.g., changes in kinship ties, demographics, and social economic factors) also affect the family caregiving contexts.
Family values might be modified or decline, according to the degree of congruence with traditional filial obligation, or change to other types of family values such as individualism^ especially when people emigrate fi:om one society to another society (e.g., fixim an eastern 54 society to a western society). An inconsistency in family caregiving experiences among
Chinese living in different geographic areas has been addressed in the literature. The variation of family caregiving experiences is described as follows.
Chinese Caregivers in China
In a study comparing the health consequences of Chinese caregivers of Alzheimer's disease (AD) patients in China and the Um'ted States, Patterson and colleagues (1998) found that family caregiving of disabled family members is perceived as a culturally and sodally mandated obligation. Rndings showed that the amount of care and help with activities of daily living (ADLs) and the negative health consequences were similar among caregivers in both areas, even though caregivers in China reported less access to emotional support than caregivers in the United States (Patterson et al.).
Chinese Caregivers in Hong Kong
In Hong Kong, family caregiving of disabled family members is perceived as a culturally and socially mandated obligation (Ngan & Cheng, 1992). In interviews with 30
Chinese family caregivers, Ngan and Cheng found that due to cultural expectations.
Chinese caregivers felt guilty when they considered sending their dependent elderly relatives to nursing homes. A prolonged obligation, however, imposed great stress on
Chinese caregivers, as more long-term care was needed. Chinese caregivers felt providing care for their elders fulfilled a socially demanded responsibility rather than being an act of love. This caused them to experience hate-love feelings. While most caregivers experienced physical, financial, emotional and/or social difficulties, many caregivers stated that guilt feelings caused them the most sufiermg (Ngan and Cheng). This finding 55 supports that beliefs of filial obligation are negatively associated with caregiver burden among Chinese caregivers as reported by Jones (1995).
Chinese Caregivers in Taiwan
In Taiwan, studies show that a strong sense of traditional fiUal obligation exists among Chinese caregivers, which has influenced family caregiving experiences (Chou,
LaMontagne, & Hepworth, 1999; Hsueh, 1998; Wang, Shyu,Yao & Wu (1998). In a study of 150 Chinese Caregivers in Taiwan, using structural equation modeling to test the relationship of filial obligation and caregiver burden, Chou, LaMontagne, & Hepworth found that filial obligation positively influenced caregiving involvement. The greater the sense of fib'al obligation the caregivers had, the more they were involved in the care. In contrast, filial obligation negatively affected caregiver burden. As the sense of filial obligation is increased, the feelings of burden are declined.
In interviews with 2 Chinese family caregivers (one was a daughter caring for a mother and the other was a daughter-in-law caring for a father-in-law), in Taiwan, Hsueh
(1998) found that the two informants perceived family caregiving as a natural part of the female family rote. They stated that providing care to dependent elderly relatives was an unquestionable responsibility for adult children. They were motivated by a sense of duty to engage in the family caregiving. They also confessed that they experienced physical difficulty while providing care, however, finally, they adjusted and felt satisfied with their caregiving role.
In a qualitative study of 71 Chinese family caregivers in Taiwan, Chinese caregivers perceived family caregiving as a journey of finding a balance for their competing needs (Shyu, Archbold, & Imie, 1998; Wang, Shyu.Yao, & Wu, 1998).
Caregivers, who found a balance point for their competing needs, finally, perceived family caregiving as acceptable responsibility and felt satisfied with it (Shyu, Archbold, & Imle;
Wang, Shyu,Yao, & Wu).
Chinese Caregivers in the United States
In 1995, Jones interviewed 10 Chinese-Americans and 10 Rlipino-Americans in the United States and found Asian Caregivers who highly valued fiUal obligation felt satisfied with family caregiving. Jones believes those Chinese caregivers were motivated by a sense of fih'al obligation to accommodate and endure family caregiving demands.
Consequently, family caregiving was perceived as fulfilling a family role and as satisfying.
This result is consistent with findings among Chinese caregivers in Taiwan (Hsueh, 1998;
Shyu, Archbold, & Imle, 1988; Wang, Shyu,Yao, & Wu, 1998).
Summary
Evidence in the literature demonstrates that family caregiving of disabled family
members is perceived as a socially mandated obligation for Chinese dl over the world.
RUal piety still strongly influences Chinese caregivers and affects the consequences of
family cargiving.
Conceptuafeatfons and Qperationalizations of Filial Obligation
According to the literature review, the importance of filial obligation influences on
family caregiving role has been mentioned since 1960 (Schorr), however, the
conceptualizations and operationalizations of filial obligation has not been well defined
until very recent. Finley, Roberts, & Banahan (1988) suggest filial obligation is aduIts*^ socialization belief that motivates them to commit to helping meet aging parents' basic needs. Filial obligation was viewed as unidimensional trait that referred to adult children's attitude in accepting responsibilities to aging parents (Finley, Roberts, & Banahan).
Cicirelli (1992) defined filial obligation as an undimensional construct that indicates an individual's global feelings of obligation. Although studies have attempted to theoretically define fih'al obligation (Cicireli^ 1992; Finley, Roberts, & Banahan, 1988), the concept clarity is till limited.
In 1988, a group of psychologists in Taiwan suggested that unidimensional construct was inappropriate for measuring a multiple dimensional construct, such as filial obligation (Yang, Yeh, & Huang, 1988). They suggested the concept of fih'al obligation should reflect four attributes rooted in Confusion's teaching: paying respect to parents, giving obedience to parents, taking care of parents, and protecting and glorifying parents.
They believe that in order to denote the values and intergenerational relauonship completely, the four dimensions of fiUal obligation concept should involve cognitive aspects and behavior because cognitive drives behavior (Yang, Yeh, & Huang). Cogm'tive aspect of filial obligation includes beliefs, affection and attitude. An individual's filial beliefs refer to an individual's understanding and acknowledgement of classic fiUal piety.
Affection involves feelings and emotional reactions toward filial beliefs. Attitude emphasizes the response tendency toward filial beliefs. Behavior refers to the actual response or performance of helping behaviors defined ni filial beliefs (Yang, Yeh, &
Huang). Therefore, a complete filial obligation concept should reflect filial beliefs, affection, attitudes and behavior (Yang, Yeh, & Huang). In their study, a 4x4 filial 58 obb'gation scale was developed to reflect four dimensions of fiUal obligation including filial beliefs, affection, attitudes and behavior with four attributes rooted in Confusion's teaching: paying respect to parents, giving obedience to parents, taking care of parents, and protecting and glorifying parents.
Subjective Feelings about Role Requirements
Whereas many studies have reported the involvement of family caregiving influencing caregiving burden (Chou, LaMontagne, & Hepworth, 1999; Deimling & Bass,
1986; Johnson & Catalano, 1983; Liu, 1991), others have shown no relationship between these two factors (George & Gwyther, 1986; Zarit et al., 1980). To explain the conflicting findings, several explanations have been documented in literature (Haley et al.,
1987; Lim et al., 1996; Pearlin, Mullan, Semple, & Skaff, 1990; and Pearlin & Turner,
1987; Poulshock & Deimling). In this section, findings related to subjective feelings about role requirements are presented. Conceptualizations and operationalizations of subjective feelings about role requirements are also addressed.
Research Findings of Subfective Feelings about Role Requirements
In 1984, to understand, the status of caregivers burden, Poulshock and Demiling
proposed a caregiving burden model with three corresponding indicators: impairment,
burden, and impact and tested it on a sample of 614 caregivers of Alzheimer's elders. In
their study, the specific impairment contexts were defined as assistance with types of
Activities of Daily living (ADL) for physical impairments (e.g., eating, dressmg, toileting,
bathing, mobility and/or incontinence), and/or care for persons with mental imparanents
(e.g., cognitive incapadty, disruptive behavior or sociability incapacity) (Poulshock & 59
Deimling). Burden referred to the corresponding feelings associated with ADL impairment and cognitive incapacity. Impact was defined as the relatively objective changes in caregivers' lives and/or their families (e.g., sodal activity restrictions or
negative impact on relationship with the elder-caregiver family (ECF).
In their study, the care receivers' physical impairments were interrelated to caregivers' subjective perceptions about the ADL task burden and caregivers' activity
restrictions. Care receivers' disruptive behaviors were associated with negative impacts on the relationship with the ECF. While ADL impairments were more highly correlated
with ADL burden than social activity restriction, care receivers' disruptive behaviors were
less strongly correlated with corresponding burden than negative impact on the ECF
relationship. The higher the caregiver's perceived ADL task burden, the more they felt
activity restrictions.
Chou, LaMontagne and Hepworth (1999) reported that the level of caregiving
involvement directed caregiver burden, although filial obligation afiected caregiving
involvement and caregiver burden. Their controversial findings support Poulshock and
Deimling's beliefs that family caregiving is a highly individualized experience. While
Poulshock and Deimling emphasized subjective responses to the caregiving context related
to types of impairment, Haley and colleagues stressed Uie caregivers' cognitive appraisal.
Pearlin and associates focused oa subjective responses to the meaning of caregiving
situations, whereas Lim and colleagues highlighted cultural background.
Another study, &i attempting to identify predictors of individual differences in
caregiver adaptation, Haley and associates (1987) examined stress and coping among 54 60 family caregivers of elderly dementia patients. They believed that the stress the caregivers experienced was the result of their highly personal response to specific caregiving contexts related to care receivers' cognitive, behavioral and self-care impairments. They found that individual subjective appraisals of care receivers' problems contributed more to negative outcomes such as depression than the level of severity of these problems, especially problems related to disruptive behavior.
Similarly, in a longitudinal study aiming to validate the process of caregiving stress, Pearlin, Mullan, Semple and Skaff (1990) conducted three interviews on 555 caregivers of Alzheimer's related dementia patients over 2 years. They found that the meaning of caregiving contexts is influenced by the caregivers' abiUty and available resources to manage difficult caregiving situations related to the care receivers' cogm'tive status and/or problematic behavior (Pearlin, Mullan, Semple, & Skaff). Caregiving stress was determined by subjective responses to the meaning of caregiving contexts.
More recently, two pilot studies on 41 Anglos and 39 Mexican Americans were conducted to understand cross-cultural family caregiver burden among Anglo and
Mexican American caregivers (Lim et al., 1996). Studies showed that the assodation between the caregiver's sodal functioning burden and the care receiver's cogm'tive incapacity was less strong for Mexican Americans caregivers than for Anglos caregivers.
To interpret the findings, Lim and colleagues suggested that the caregivers' cultural values and/or cultural norms play important roles m shaping their perception of burden. In other words, the caregiver's subjective values suchas fiBal beliefs mfluence then: perceptions about providmg assistance (Lim et al.). 6L
Overall, although researchers provide different explanations to interpret their findings, they all agree that an individual's feelings toward a particular situation are highly individualized and influenced by the individual's beh'efs and values. Therefore, it is likely that subjective feelings about providing assistance is a suffident indicator to predict caregiving appraisal including perceived demands or rewards.
Conceptualizations and Operationalizations of Subjective Feelings about Role
Requirements
The concept of subjective feelings about role requirements, first mentioned by
Poulshock and Deimling in 1984, refers to caregivers' subjective view about certain tasks in relation to specific impairments. Poulshock and Deimling believe that an individual's
negative perception about providing assistance to elderly relatives is an important
predictor for caregiver burden. Poulshock and Deimhng suggest that burden measure should test how caregivers subjectively respond to specific caregiving tasks. In their study, subjective feelings about providing assistance was operationalized as an antecedent
factor indicating negative emotional responses toward specific tasks in predicting burden
in relation to specific impairments of care receivers.
Seven items were derived about how upset caregivers felt about providing Activity
of Daily Living (ADL) assistance. These sevea items were survey questions with five
response options of no help needed, help needed but no problem at all, help needed and
tiring, help needed and difficult, help needed and upsetting. The original seven items were
translated mto Spanish and validated in Mexican caregivers in the United States (Phillips
et al., 1996). 62
Perceived Burdens
Family caregiving is a valuable social resource across cultures. However, considerable literature indicates the effects of stressors on family members caring for dependent elderly relatives are burdensome (Clybum, Stones, hadijstavropoulos, &
Tuokko, 2000; Faison, Faria, & Frank, 1999; Nagatomo et al., L999). In this secuon, findings related to perceived burdens are described.
Research Findings Related to Perceived Burdens
Caregiver burdens have been identified as a sigm'ficant factor affecting caregivers' physical, mental and social well-being (Dew et al, L998; George, 1994; Raveis, Karus, &
Pretter, 1999). In a study examining the weE-being of 133 caregivers of heart transplant recipients. Dew et al discovered that caregivers' well-being declined as perceived burden increased. Similarly, in another study of 164 daughter caregivers of elderly, Raveis,
BCarus, & Pretter found that caregivers' depression was predicted by their perceived burden. This is also supported by Clybum, Stones, Haadjistavropoulos and Tuokko, who indicate caregivers' well-being is mediated by appraisal of burden in a sample of 613 caregivers of relatives with Alzheimer's disease.
Several antecedent factors associated with appraisal of burden have been documented in literature. They are task demands (Faison, Faria, & Frank, 1999) and negative responses to task demands of the caregivmg situation (Fingerman et al., 1996), and filial obligation (Chou, LaMontagne, & Hepworth, 1999; England, 1990). In a study of 88 caregivers of elderly relatives with, chronic illness, Faison, Faria and Frank found that activity of daily living (ADL) task demands were positively related to perceived 63 burden. This result is supported by Fingerman and his colleagues, who found caregivers' negative emotional responses to task demands were positively related to caregiver burden in a 143 caregivers of frail elderly. In contrast, in a sample of 168 adult children caregivers of elderly relatives with dementia, England reported there was a negative relationship between filial obligation and perceived burden. A similar result was found in a sample of 150 Chinese caregivers in Taiwan. Chou, LaMontagne and Hepworth found caregivers' beliefs of filial obligation negatively predicted caregiver burden.
Another significant factor assodated with caregiver burden is use of copmg options. Some studies suggest that appraisal of caregiving stressor (burden) influences use of coping options (Irvin& Acton, 1996; Lazarus & Folkman, 1984; O'Brien, 1993).
In a sample of 20 spouse caregivers of family members with multiple sclerosis, O'Brien found that appraisal of burden was significantly, positively related to use of coping.
Others suggest that the impact of caregiving stress on caregivers' well-being is buffered by use of coping option (Andrews & Robinson, 1991; Pearlin, Mullan, Semple, & Skaff,
1990). In an open-ended exploratory interview with 555 caregivers of relatives with
Alzheimer's disease, coping was identified as one of the significant mediators between the appraisal of burden and well-being including physical, mental and role function. In the following section, the conceptualizations and operationalizations of perceived burdens are reviewed for further understanding the inconsistency in findmgs.
Cnncepttialiyannns and OperatinnahVatinns of Perceived Burdens
In 1963, the negative impact of caring was firet recognized as an important issue in family caregiving by Grad and Samsbury. Smce then, the issues associated with caregiver 64
burden have been intensively studied from various perspectives. The concept of caregiver
burden has evolved with time. During the early eighties, researchers tended to treat
"burden" as an objective, um'dimensional construct that implied overall and potentially
negative impacts of family caregiving (Robinson, L983; Zarit, Reever, & Bach-Peterson,
1980). Later, "burden" was defined as a multidimensional construct to reflect the
multifaceted impacts on caregivers' lives (George and Gwyther, 1986; Kosberg & Cairl,
L986; Montgomery & Gonyea, & Hooyman, 1985; Lawton et al., 1989; Poulshock &
Deimling, 1984; Vitah'ano, Yong, & Russo, 1991). There are inconsistencies in
conceptualizations and measures of caregiver burden. To further understand
inconsistencies in burden measures, an overall review of relevant literature is presented as
follows.
Burden as An Objective Unidimensional Construct
Burden, as an objective unidimensional construct, was described as the most
frequently mentioned discomfort, by caregivers, in such areas as physical and
psychological wellbeing, finances, social life and their relationships with care receivers
(Zarit, Reever, L980). They believed that feelings of burden would be related to the
extent of behavior impairment manifested by the person with senile dementia" and "the
relationship of the caregiver to the impaired person affected burden" (p. 651-652). In
their study, burden was operationalized as caregivers' feelings about giving care and its
effects on their lives.
Robinson (1983) suggested burden is the subjective perception of strains of the
caregivmg situation. Burden referred to results of common stressors assodated with 65 family caregiving in a stud of informal caregivers of elderly patients following acute hospitalization for hip surgery or heart disease. However, the burden concept failed to clearly define and specifically measure burden by using total scores to indicate overall levels of burden. However, the issues related to confounding between subjective responses (e.g., experiendng emotional adjustments) with objective impacts (e.g., changing personal plans) were not clearly described.
Burden as Mutildimensional Construct
More recently, a number of researchers suggested that caregiver burden is not well defined unless it is conceptualized as a multidimensional construct that reflects the multifaceted impacts on caregivers' lives (George & Gvvyther, 1986; Montgmery, Gonyea,
& Hooyman, 1985; Poulshock & Deimling, 1984; Vitaliano, Yong, & Russo, 1991). The terms burden (Montgomery & Gonyea, & Hooyman; Novak & Guest, 1989; and
Poulshock & Deimling), impacts on wellbeing (George & Gwyther), cost (Kosberg &
Cairl, 1986), and stressors (Lawton et al., 1989) were used to describe the negative consequences of family caregiving. While Poulshock and Deimling emphasized subjective emotional and affective responses, Montgomery Gonyea and Hooyman stressed subjective feelings and objective impacts on the caregivers' lives and relationships. George and
Gwyther highlighted the negative impacts on different aspects of the caregivers* wellbeing.
Actual and potential negative impacts of caregiving were conceptualized as burden in the works of Kosberg and Cairl, and Lawton and colleagues.
Poulshock and Deimling (1984) were the first researchers to clearly and specifically conceptualize burden as a result of caregivers' highly personal and 66 mdividualized responses to specific caregiving contexts (Vitaliano, Yong, & Russo, 1991).
They assumed that "burden is the result of highly personal and individualized response to specific caregiving contexts" (Poulshock & Deimling, p. 231). They suggested that "the linkage between impairment and impact is centered on the subjective view of caregivers that certain tasks performed and derived firom the impairments were burdensome" (p.231).
They indicated that the concept of burden should refer to the caregivers' subjective perceptions and interpretations in relation to the elder care receivers' specific impairments.
That is, "the degree of burden is associated with subjective perception about caregiving contexts and is related to relatively objective changes in caregiver's lives and families"
(p.235).
To test the above hypotheses, Poulshock and Deimling proposed a burden model, which depicts the process of burden in two consecutive categories: elders' impairment, and impact/burden. Elders* functional/behavioral impairments are defined as three independent variables. Assistance with types of Activities of Daily Living (ADL) for physical impairments (e.g., eating, dressing, toileting, bathing, mobility and/or incontinence), and/^orcare for persons with mental impamnents (e.g., cognitive incapadty, disruptive behavior or sociability incapacity). Burden is "a mediating force between the elders' impairments and impact on caregivers" (p.230). Burden refers to the emotional responses of the caregivers toward providing assistance with ADL for physical impairment and disruptive behavior for cognitive mcapacity. Impact is treated as a dependent variable
that refers to the relatively objective changes in caregivers' lives and/or their families (e.g., sodal activity restrictions, job conflict, finances, or negative impact on relationship with 67 elder-caregiver family (ECF). Caregiving burden is conceptualized as a multidimensional construct, which includes subjective emotional responses to and the objective effects of caregiving, although the terms "subjective" and "objective" burden were not used.
Whereas subjective emotional responses refer to feeling embarrassed, overloaded, trapped, resentful, objective effects reflect disruptions in finances, role, family life and the relationship of caregiver to elderly care receiver and caregiver activities, Poulshock and
Demiling disclosed that burden was assodated with severity of ADL impairments and behavioral problems. While the ADL impairment was highly associated with ADL burden, disruptive behavior was strongly correlated with negative impact on relationship with elder-caregiver family (ECF) relationship.
Poulshock and Demiling's study made several excellent contributions to the conceptualization of burden. E^t, their study clearly and specifically conceptualized burden as a subjective phenomenon by proposing a conceptual model to examine the corresponding relationship between functional/behavioral impairments and burden.
Second, they obtained evidence of construct validity by testing hypotheses in actual caregiving problems and care recipient impairments. Third, they enhanced the construct validity for measuring burden by recognizing the potential relationships between caregiver burden and general psycholgical criteria (e.g., depression). Fourth, they differentiated subjective burden measures firom impact measures. Fifth, they established a reliable and usefiil tool for impact measurement by obtaming a good mtemal consistency coefficient of alpha .88 for the subscale of negative changes in family relationship, and alpha .80 for the subscale of restrictions ia activity (Derailing & Bass, 1986). 68
On the other hand, some drawbacks threatened conceptual clarity. Rrst, the distinction between sources of objective burden and subjective burden were not clearly defined. Second, psychological distress such as depression was defined as an antecedent to burden. However, no substitute psychological outcome measures such as mental health, distress or role satisfaction were considered to test feelings on the negative impacts of caregiving. Third, the usefulness of the impact scale is limited for studies with different types of care recipients (e.g., individual with cancer or stroke) because it was designed for caregivers of dementia elders (Haley et al., 1987).
Montgomery, Goynea and Hooyman (1985) were also interested in understanding the relationships between specific caregiving responsibilities and burden. Burden was conceptualized into two categories: objective and subjective burden. The term burden covered everything firom disability of the care receiver and the feelings of the caregiver to the effects that caregiving has on the famEy firom either a subjective or objective perspective. They bplieve that factors contributing to subjective burden are different from those contributing to objective burden. Subjective burden was defined as "the caregiver's attitudes toward or emotional reactions to the caregiving experience" (p. 21). Objective burden referred to "the extent of disruptions or change in various aspects of the caregiver's life and household" (p. 21). They suggest that a given family caregiver might not only experience subjective or objective burden but simultaneously experience a high level of objective burden and a low level of subjective burden or vice versa. It is important to make a (fctinction between subjective burden and objective burden measures. They were the first to distinguish subjective burden firom objective burden. 69
George and Gwyther (1986) conceptualized burden as the consequences of fanuly caregiving on caregivers' wellbeing ia areas such as physical health, naental health, financial resources and social partidpation. They suggested that "caregiver burden and caregiver wellbeing are the opposite sides of the same coin" (p. 253). Burden is synonymous with declines in wellbeing. Caregiver burden refers to "the physical, psychological or emotional, social, and financial problems that are experienced by family members caring for impaired older adults" (George & Gwyther 1986, p. 253). They
beUeved that caregiving contexts and available resources shape the caregivers' wellbeing.
The caregiving contexts referred to personal and situational characteristics including characteristics of patient illness, type of relationship between caregiver and care
precipitant, care recipient living arrangements. Care recipient living arrangements were
identified as living with caregiver, living in nursing home, and living with someone else.
The patient illness characteristics were coded into two categories, duration of the illness and severity of the symptoms. George and Gwyther suggested that the longer time the
caregivers cared for their dependent family members with severe symptoms the greater the
declines in their wellbeing. George and Gwyther made a distinction between the
dimensions of burden by using different scales to reflect various areas of caregiver's
wellbeing and function.
To measure different dimensions of caregivers' wellbeing, George and Gwyther
applied four different instruments: the Short Psychiatric Evaluation Schedule (Pfeiffer,
1979), the Affect Balance Scale (Brabdum, 1969), Subjective Assessments of Economic
Wellbeing, and Sodal wellbeing. Each instrument based on reliable evidence reflects 70 different dimension-specific patterns of caregiving impact. The internal consistency
reliabilities for the Short Psychiatric Evaluation Schedule and the Affect Balance Scale
were alpha .85 (Pfeiffer) and .89 (Brabdum), respectively. For the Subjective
Assessments of Economic Wellbeing and Sodal Wellbeing, the internal consistencies were
.85 and .79, respectively (George & Gwyther, 1986). The findings showed that burden is a subjective rather than an objective phenomenon. There was no difference in caregiver
wellbeing among the following variables: types of relationship between caregivers and care
receivers and the characteristics of patient illness. In contrast, caregivers' wellbeing was
strongly associated with availability of resources. Furthermore, caregiver wellbeing was
directly associated with the characteristics of the caregiving situations and the resources
available rather than the condition of the patient.
George and Gwyther made a distinction between the dimensions of burden by
using different scales to reflect various areas of caregiver's wellbeing and function.
Although this study was not designed to provide full psychometric development of indices,
George and Gwyther's study provided an opportunity to derive some measure of reliability
of those four scales.
Kosberg andCairl (1986) also suggest burden should reflect adverse subjective
and objective consequences of caregiving. They believe that "caregivers not only are
over-burdened by the needs of the care receivers but also they nuiy be mcompetent to
provide the needed care" (p. 274). Therefore, burden was conceptualised as multifaceted
costs of caregiving mcluding caregivers' personal-sodal restrictions, physical-emotional
health, economic costs, perceived care recipient as troublemaker, and value tnvesiment ni 71 caregiving. In their study, costs referred to actual and potential consequences or impacts of caregiving (Kosberg & Cairl). By reviewing literature in the areas of child abuse, spouse abuse or family conflict, they suggested cost of caregiving should theoretically represent psychopathological, sociological, sodal psychological, life crises, sodal structural, intergenerational, and human ecological explanations of abusive behavior.
They also developed a cost of care index targeting caregivers of elderly relatives for screening early sign of abusive caregiving.
Lawton et al. (1989) defined burden as an external demand or potential threat that
has been appraised as a stressor based on Lazarus and Folkman's (1984) stress-coping
theory. They believe that a potential stressful situation arouses a process of appraisal and cogm'tive coping. They defined the cognitive coping process as an integrated part of the
reappraisal process. Burden is conceptualized as a result of appraisal and reappraisal.
They suggest that both positive and negative aspects of appraisal are possible responses to
a potential stressor, in this case caregiving.
Summary
Caregiving affects many aspects of caregivers' lives including health status,
disturbed feelings, finandal strains, social activity, and family relationships. The caregiver
burden concept has been defined as features or characteristics of a relationship between
the caregiver and the care receiver or other family members rather than personality traits
of the caregiver (Vitaliano, Young, Russo, 1991). Caregiver burden has a relational
characteristic and it has usually been measured withm the context of other variables such
as caregiving involvement, use of coping strategies and depression (Vitaliano, Young, &. Russo). Those context variables were either identified as predictors, mediators or outcomes of, or subsequent theories about caregiver burden.
Overall, caregiver burden refers to negative consequences of family caregiving, which include experiencing problems or difficulties in such areas as physical, psychological, social or financial wellbeing, or family relationships, or all of them at once, at some point during providing care (George & Gwyther, 1986; & Poulshoch & Demiling,
1984). Researchers suggest that a multidimensional instrument provides aa advantage for making a distinction among dimensions of burden (George & Gwyther, 1986; Kosberg &
Cairl, 1986;Lawtonetal., 1989).
Perceived Rewards
In contrast to many studies focusing on negative perceptions and consequences of caregiving, a number of researchers discovered that family caregiving has been perceived as a positive experience among many caregivers (Brody, 1989; Cohler, Groves, Borden &
Lazarus, 1989; Given, King, Collins & Given, 1988; Jones, 1995, 1996). Among those studies, positive experiences and consequences include positive feedback from the care receiver (Brody, 1989), enhanced self-regard from caregiving (Cohler, Groves, Borden, &
Lazrus, 1989), positive reactions (Given, King, Collins, & Given, 1988), positive feelings of obligation and satisfaction (Jones, L995; 1996).
Research Findings Related to Perceived Rewards
In addition to the above findings, it is believed that there is a relationship between caring and cargiving. In 1983, in the book A Labour of Love. Women. Graham attempted to describe the relationship of carmg and caregiving. Graham regards caring and 73 caregving as two sides of the same coin with two interdependent dimensions: love and labor in an interdependent relationship. Caring is love, which, refers to the emotions and feelings associated with care. Caregiving is labor, which consists of tasks and activities associated with care. Caring is the basis for caregiving and caregiving is manifested in caring. An individual's caring character may direct the levels of caregiving involvement and may also shape feelings about caregiving. This notion implies that the caregiving experience consists of love and responsibih'ty.
Klein (1989) believes that caregiving experience is related to moral development and involves feelings of pride and suffering. An individual's moral development determines the person's moral beliefs and judgments about caring and shapes the person's subjective feelings about caregiving (Klein). EQein suggests when burden is perceived as an obligation requiring a cost and as merit for suffering, positive caregiving experiences are part of burden. For example, when a caregiver perceives a balance between moral beUefs and judgments about caring, family caregiving is perceived as bearable.
Furthermore, when family caregiving is perceived as sacrifice for an ultimate goal, (e.g.,
perceiving caregiving as a way of paying respect or paying back, or faith in a loving God),
the caregiver may experience satisfaction with caregving or feelings of pride (Klein).
There is a tendency that subjective beliefs about caring (beliefs and judgment about caregiving) direct levels of caregiving involvement (providing family caregiving activities)
and subjective experiences (subjective perception about caregiving), especially positive
experiences or feelings. These theoretical notions provide explanations for why and how
some caregivers experience less feelmgs of burden than others, or even have positive 74 experiences, while others do not. These theoretical notions have also been supported in the literature.
In 1988, Picot studied 81 African American caregivers and discovered that when caregivers felt love and responsibility for the care receivers, they had positive feelings of self-worth. Picot found that caregivers' positive experiences were significantly assodated with caregivers' sense of duty derived from religious beliefs (Picot, 1988). Murrary and colleagues (1999) revealed a similar finding on 20 spouse caregivers of Alzheimer's patients in London. They discovered that perceiving family caregiving as a way to fiilfiH the sense of duty directs positive caregving experiences. This finding has also been supported in studies with Asian Caregivers (Jones, 1995; Lee, Kim & You, 1997). Jones interviewed 10 Chinese-American and 10 Filipino American and discovered that Asian caregivers who had a strong of sense of duty and paying respect to elders experienced feelings of satisfaction toward caregiving. In another study of 54 Korea caregivers living in rural and urban areas, Lee, BCim and You (1997) disclosed that rural caregivers had greater positive feeh'ngs of self-growth and protectiveness of their parents than urban caregivers. They also discovered a greater adherence to traditional cultural values among rural caregivers with a greater, positive, caregiving experience. Since paying respect to elders has long been perceived as part of the traditional custom and moral obligation in
Asian culture, it is likely that moral obligation plays a very important role in influencing positive caregiving experiences. 75
Cnncp.ptiiah'yations and Operatfonalizations of Perceived Rewards
Although positive experiences of caregiving have been widely described, in the literature, very little research, addresses the theoretical definition of perceived rewards.
The term "rewards" was first used by Picot (1991) to acknowledge positive experiences.
"Rewards" is operauonally defined as a unidimensional construct that indicates an
individual's positive appraisal of caregiving including positive experiences and positive changes in life. The positive experience refers to pleasant feelings and satisfaction in
response to family caregving role. Positive changes in life include positive consequences such as financial condition or relationships among family members.
Use of Coping
Coping has been widely studied in the behavioral sciences such as psychology, sodology and nursing. Coping has shown to be an important factor influencing family
caregivers burden (Bryan, 1990; Chan, Lee, & Lieh-Mak, 2000; Chou, LaMontagne, &
Hepworth, 1999; Hull, 1992; Irvin & Acton, 1996; O'Brien, 1993) and health
(Neundorfer, 1991; Pruchno, & Kleban, 1993; Quayhagen, & Quayhagen, 1988). In this
section, findings related to coping options are described.
Research Enndings Related to Use of Coping Options
According to the literature review, there is no clear consensus regarding
consequences of use of coping. In a study of 150 Chmese caregivers in Taiwan, Chou,
LaMontagne, & Hepworth (1999), coping has shown to be a beneficial mediator for
caregivers burden. This finding is supported by a longitudinal, qualitative study of 14
hospice caregivers of dymg relatives at home (Hull, 2000). In contrast, O'Brien (1993) 76 indicated that use of coping was significantly positively associated with caregiving burden in a study of 20 spouse caregivers of multiple sclerosis. In another study of 60 spouse caregivers of dementia, Neundorfer found that caregivers' health was influenced by use of coping.
Various findings in literature nught result firom different coping patterns were developed and used during the process of caregiving (Haley et al., 1987). Haley and colleagues stat that durmg the process of family caregiving, different coping patterns were developed to respond to different stressor or difficulties. Although the ftinction of coping responses is to regulate emotions and deal with problem through action, coping responses also can adversely affect health when coping options assodated to maladaptive or impeding adaptive health/illness-related behaviors are used (Lazarus & Folkman, 1984).
This has been empirical supported by studies (Almberg, Grafstro, & Winblad 1997; Chou,
LaMontagne, & Hepworth, 1999; Li, Seltzer, & Greenberg, 1999; Morese & Fife, 1998;
Shiau et al., 1999). Empmcal evidences indicate that the use of coping strategies vary with different slressfiil situations (Folkman & Lazarus, 1980; Vitaliano, DeWolfe, Maiuro,
Russo & BCaton, 1990a). For example, although both "problem-oriented" and "emotion- oriented" coping have shown mediate relationships with health outcomes (Folkman &
Lazarus, 1988a), problem-oriented coping was used more often and. more useful with changeable situations. On the other hand, emotion-oriented coping was used more firequently in and was more beneficial m unchangeable situations (Vitaliano, DeWolfe,
Maiuro, Russo & Katon). 77
Although there is inconsistency in findings on the contributions of coping options to health outcomes, coping efforts significantly influence the impact of caregiving on caregivers' health (Haley et al., 1987; Neundorfer, 1991; & Quayhagen & Quqyhagen,
1988). For example, Haley and associates found behavioral strategies were the most effective contributors to heath outcome among caregivers. Neundorfer discovered that confrontive coping was the best predictor of anxiety and escape-avoidance was the best predictor of overall psychological distress and interpersonal conflict. Quayhagen and
Quayhagen report this effective contribution present among spouse caregivers, but not in daughters.
Conceptualizations and Operationalizations of Coping Options
According to the literature review, coping was recognized as a multiple dimensional construct by researchers. However, while some researchers consider coping to be a stable inherent disposition or personality trait (Caver et al., 1989; Haan, 1977), others suggest that it is a changing process contributed with adapting to reality (Amirkhan,
1994; Folkman and Lazarus, 1984; Moos, 1977). Caver and colleagues defined coping as a personality trait, which an individual's behaviors are attitudinally and cognitively consistent across situations. Hann suggest coping is an individual's stable preferences that emphasize how a person usually handles a variety of stressful situations.
As opposed to it being a trait» coping is defined as "constantly changing cognitive and behavioral efforts to manage specific external and/or intemal demands that are appraised as taxing or exceeding the resources of the person (Lazarus & Folkman, 1984, p.l4). Coping is characterKed as an ongoing process and that a person might use one 78 strategy in the faeginning of a particularly stressful situation but later might change to another strategy (FoQcman & Lazarus, 1980). Coping concerns with how a person actually handles a specific, stressfiil situation, was suggested to be appropriate for understanding the concept of coping (Folkman, 1988). An individual's coping styles are situational responses to specific stressor (Folkman & Lazarus). They suggest that a person might use more than one strategy at the same time (Folkman & Lazarus, 1984).
Folkman and Lazarus (1980) also challenged the coping trait approach as inappropriate for measuring a person' ability to cope in a specific context because the coping trait approach focuses on the stability of a person's characteristics rather than the stability of how a person copes. They believed that coping process approach is appropriate for measuring the complexity of the multidimensional quality of actual coping for people with a long-term stressful situation such as caregiving (Folkman, 1988). They also suggest the function of coping is to manage demands (Folkman, 1984), which include regulating emotions and dealing with the problem through action (Folkman & Lazarus,
1980; 1984). To acknowledge the complexity and variability of the actual coping process, three essential criteria were suggested for opterationalized the coping concept. These are
(1) coping must be examined within the context of a specific stressful encounter, (2) what the person actually does must be described, and (3) there must be multiple assessments to examme the variability of coping with the respect to specific stressful encounters over time
(Folkman & Lazarus).
Literature indicates that attributes defining the coping concept range fi:om two to fourteen dimensions. Various attributes with inconsistent and overlappmg labels were 79 identified for both coping trait and status tools. In this section, two-dimensional versus multidimensional approach is used to address the essential attributes of coping
In 1980, Lazarus and. Folkman developed a coping checklist and two types of coping strategies were derived: (1) problem-focused strategies and (2) emotion-focused strategies. While problem-focused strategies emphasize doing something better to change problems causing the distress and emotion-focused strategies stress using a cognitive process to react to a stressful situation or regular emotional response (Folkman, 1988).
Since then, the coping scale has evolved. Although the number of factors stays constant, other interchangeable terms (Haley et al. 1987; Segall & Wykle, L988; & Wood, 1987) new terms with respect to different emphases (Patterson et al., 1995; &Tobin et al., 1989) were used. The term problem-focused has been changed to other alternative terms such as behavioral-oriented or action-oriented (Wood, 1987). Problem-focused strategies include problem-solving/help-seeking (Haley et al. 1987) and instrumental help (Zarit et al.,
1986).
Similarly, the term emotion-focused has been changed to terms such as cognitive- oriented or affection-oriented. Emotional-oriented strategies include reforming logical analysis (Haley et al.), positive self-taDc (Levine, Dasloor, & Gendron, 1983), accommodation (Segall & Wykle, 1988) fantasy and blame (Quayhagen &. Quayhagen,
1988), passive forbearance (Strong, 1984), prayer to God (Segall & Wykle), or no response (Irvine, Dastoor &. Gendron). Some researchers attempted to emphasize the importance of one set of coping over another. New terms based on the ego-psychology concept to classify coping strategies were created. For example, the two types of copmg strategies were changed into "engagement" and disengagement" (Tobin et al., 1989),
"repression" and "sensitizau'on" (Cohen, 199 L), and "avoided" and approach" (Patterson etal, L995).
On the other hand, more recent, in order to reflect the complexity of coping options, several researchers identify multiple attributes by applying more advanced statistical techmques such as confirmatory factor analysis (Amirkhan, 1990; Carver et al.,
1989; Jalowiec, 1979; 1988; Folkman, 1988; Folkman &. Lazarus, 1988; Quayhagen &
Quayhagen, 1988). A number of measures with overlapping and interchangeable labels were used developed (Carver et al; Jalowiec; Amirkhan, 1990; Folkman & Lazarus, 1984;
Moos, 1990), however, the conceptualization of coping is till vague.
Folkman, Lazarus and Dunkel-Shetter (1986) revised the Ways of Coping checklist to develop the Ways of Coping questionnaire (WOC) (Folkman & Lazarus,
1980). The WOC treats coping as a process of dynamic situation-specific responses with eight attributes. The eight factors are confirontive coping, distancing, self-controlling,
seeking social support, accepting responsibility, escape-avoidance, and positive appraisal
(Folkman, Lazarus, Dunkel-Schetter). Confirontive coping suggests aggressive efforts to
alter the situation, which refers to a degree of hostility and risk-taking. Distancing coping
is defined as efforts to detach oneself of create a positive outlook. Self-control coping
refers to efforts to regular one's own feeling and actions. Seeking social support coping
suggests efforts to seek informational and emotional support. Accepting responsibility
refers to acknowledging one's own role in the problem, which suggest concomitant theme
of trymg to put things right. Escape/avoidance coping describes wishful thinkmg and 81 behavioral efforts to escape or avoid. Planfiil problem-solving coping is operated as eliberate problem-focused efforts to alter the situation. Positive reappraisal refers to efforts to CTeate positive meaning by focusing on personal growth. Although the WOC attempted to theoretical define each factor, problem of unstable factor structure has been noted by the developer (Folkman & Lazarus, 1988) and criticized by Ayers (2001).
The WOC has been tested in various samples including healthy individuals
(Folkman and Lazarus, 1985; Smyth &Yarandi, 1996), climcal populations (Cohen, 1991;
Mishel & Sorenson, 1993; Wineman, Durand, & Steiner, 1994), and caregivers of different types of dependent relatives (Hooker, Frazier, & Monahan, 1994; O'Brien, 1993;
Phillips & Thomas, 1996; & Rosengard & Folkman, 1997). It has been criticized also demonstrated cross-cultural validity in Chinese Caregivers in Taiwan (Chou, LaMontagne,
& Hepworth, 1999), and AMcan American Caregivers in the Um'ted States (Smyth and
Yarandi, 1996). Although the WOC is a well-known tool, however, its factor construct was unstable as various numbers of factors in different samples were reported.
Moos (1990) developed a 48-item of Coping Responses Inventory (CRI) to organize coping efforts into four sets of responses: approach, avoidance, cognitive and behavior. The CRI assumes effort as a situation-specific response to stressfijl event. The
CRI has been suggested as good reliable useful tool ui identifying and describing types of coping styles by patients (Ayers, 1999; Milne, 1992). Ayer believes the CRI has strong face validity and clear, comprehensive structure. In 1990, Moos revised the CRI into a 2
X 4 of CRI-adult scale. The CRI-adult defined approach copmg as logical analysis, positive reappraisal, seeking guidance, and problem solving. The avoidance coping 82 options referred to cognitive avoidance, acceptance/resignation, alternative rewards, and emotional discharge. However, no clear theoretical definition for each subject concept was given. In addition, lack of stable factor structure was noted in literature; a variation of three to ten different factors was reported in different studies with different populations
(Moos, 1988; Moos et al., 1990, 1995; Moos, et al., 1992).
Summary
Use of coping options has been recognized as an important indicator in understanding the processes and consequences of family caregiving. Coping options among caregivers vary with different stressful situations across cultures. The contributions of coping options to health outcome among caregivers were also various with different stressful situation and across cultures. In addition, the effects of coping on caregivers' health outcome differ with different coping efforts.
Coping is defined as a multidimensional construct. Copmg options have been conceptualized as a personality trait and status (process). The coping trait was also called dispositional coping styles or ego psychological oriented approach. The coping status was also named situational coping styles or specific situational oriented approach. While the coping trait focuses on ability in mastering everyday stressfiil situations, coping status focuses on the efforts used to manage stressfiil situation.
A number of measures of coping with, widely variety of attributes ranging firom two to fifteen were developed. While the number of attributes stays consistent, overlapping and interchangeable labels were used for those measures. 83
Caregivers' Health Status
Recent caregiving studies disclosed that caregiver burden was a significant predictor for caregiver health. The majority of studies suggested that there is an inverse relationship between health status and caregiver burden (George &. Gwyther, 1985;
Griggs, 1995; & Martinson, Gaspers, & Muwaswes, 1998). An individual's health status has long been recognized as an important index for health-related outcome assessment. It is suggested that health-related outcome assessment is an important indicator for health saeening, health promotion or disease prevention (Ware & Sherboume, 1992). To understand the relationship of caregiving experiences, either negative or positive, and the caregivers' health, the measures of health status and findings relevant caregiver health are discussed in the following sections.
Research Findings Relevant to Health Status of Caregivers
In a study of 122 family caregivers of hepatoma patients in Taiwan, Lin and colleagues (L996) discovered that caregiving stress was significantly correlated with the caregiver's physical, psychological and social health. Similar result was supported in study of 510 caregivers of relatives with dementia (George Sc Gwyther, 1986). George and Gwyther compared caregivers to non-caregivers and found that caregiver mental health and social well-being were primarily affected by caregiver burden. In another study, Gaynor (1990) found that there is a positive relationship between caregiver burden and mental health among older spouse caregivers. However, in a longitudinal study. Bull
(1990) mterviewed caregivers of discharged patient hospitalized for an acute illness episode discovered that the level of burden had no significant directly effect on caregiver 84
physical health status. In addition, when Gaynor compared level of burden and health status between younger and. older spouse caregivers, a negative relationship between level of burden and mental health was discovered. Gaynor found that a higher level of burden
and better health status experienced, by younger spouse caregivers. This finding revealed
that the relationship of level of burden and. health status was also confounded by age of the
caregiver.
While many studies focused on relationship between the negative caregiving
experience and health status, Quayhanger and Quayhangen (1988) found a positive
relationship between positive caregiving experiences and health status among female
spouse caregivers. Since many nunority caregivers perceived caregiving as an imperative
obhgation for committing the law of tradition, in this case, it is difficult for them to
confess that caregiving is burdensome. There is evidence in the literature that satisfaction
was the most common consequence for minority caregivers such as Africa-American,
Chinese and Korean. It is likely an associated relationship between caregiver burden or
positive caregiving experience, and health status among minority caregivers such as
Chinese. Currently, in caregiving literature, caregivers' health has been recognized a
major concern for nursing and is a critical indicator for measuring caregiving outcome.
However, almost no literature documented the relationships between caregiving
experiences, either negative or positive, and health status among Chinese caregivng &i the
Um'ted States. 85
Conceptualizations and Qperationaliziitinns of Health Status
The concept of health status has been well defined in 1958 by WHO in which it refers to "a state of complete physical, mental and sodal well-being and not merely the absence of disease and illness". To distinct firom medical health, the term of "perceived health" was suggested to refer health as an individual's subjective perceptions of his/her capadty (Suchman, Streib and Phillips, 1958). Later, it was exph'cated and defined as
"fiinctional adequacy of an individual to fulfill the role which a healthy member of his/her age and sex. is expected to fulfill" (Sanders, 1964). Health reflects people's ability to perform sodally allocated tasks of daily living (Sullivan, 1966). Health refers to a level of process, which raises or declines over time (Sullivan). Therefore, an individual's health can be indicated by individual's perception of changes in usual feelings and fiinction
(Sanders). Although health has been wett defined, no explanation about how to measure the subjective perceptive health was documented.
In 1966, Rosow and Breslau suggest that health is a multifaceted concept that reflects an individual's perceptions about his/her own overall functional status and subjective well-being. In their study, the concept of health was operationalized as "the degree to which individuals claims of the specific capadties involved". In order to capture the actual meaning of health for an individual, they suggest that the assessment of health should be evaluated firom a client perspective. A. 25-diversified Guttman health scale was developed to question about the respondent's perceptions of his/her health lunits, and activities or social function constrains to capture the respondent's overall functional health. 86
Since then, several health-related outcome assessment tools from the client's perspective have bee developed for general populations (Chambers et al, 1976; Ware,
1976; Ware and Kamos, 1976; Ware et al., 1980; Ware et al., 1993). Chambers and colleagues developed a 59-item of McMaster health index questionnaire (MHIQ) based on literature review of health status measurements and consultation of expert panels to predict care need for general populations. The MEHQ emphasizes on evaluation of respondents' ability in relation to physical, emotional and social functioning.
In 1976, Ware and associates suggest assessment of health should respect to time as health described a process and changes with time. They developed a 36-item of health rating scale- the health perceptions questionnaire (HPQ) to reflect eight perceptual dimensions in relation to general health and sick role propensity for general populations.
The eight dimensions are (I) past health, (2) present health, (3) future health, (4) health- related worries, (5) health-related concerns, (6) susceptibility to illness, (7) resistance to illness, and (8) attitude toward going to the doctor, to detect a global picture of overall health.
Later, a more comprehensive and with better psychometric properties scale-the 36- item Short-Form health survey (SF-36) for health insurance and medical outcome studies was developed (Ware et al., 1993). The SF-36 is the most popular and widely used tool for screening health in clinical or commura'ty patients (Ware & Sherboume, 1992). It is a combmation of Lflcert-type and dichotomous survey questionnane. The 36-item Short-
Form Health Survey (SF-36) is a generic measure of health status on eight areas of wellbeing (Ware & Sherboume, 1992). These eight areas of wetlbeing are (I) general 87 health perceptions, (2) general mental health, (3) limitations in physical activities, (4) limitations in social activities, (5) linutations in usual role activities, (6) bodily pain, (7) general mental health, and. (8) vitality (energy and fatigue) (Ware & Sherboume, 1992).
The SF-36 is a well-known and well-developed, tool that has been widely used and vah'dated for a large heterogeneous population across cultures including Australia, China,
Deiunark, Germany, Sweden, the Ura'ted States, and Taiwan.
Summarv
Caregivers' health is a major concern for nursing. It has been recognized as an important indicator in understanding consequences of family caregiving. Factors influencing caregivers' health include subjective cultural beliefs, perceived burdens, perceived rewards, and use of coping. However, there is inconsistency in findings.
Researchers suggest perceived health as a multipledimensional concept reflecting individuals' subjective perceptions of their ability to perform sodally allocated tasks of daily living. Several assessment instruments firom the client's perspective have bees developed for measuring overall health for general populations. Among them, the SF-36
is a widely used short form measure of subjective health status that was psychometrically sound for general populations across cultures. 88
CHAPTERS
This chapter includes research design^ description of sampling procedures, training
of recruiting representation and data collection protocol. Instrumentation is presented.
The data analysis plan is addressed.
Research Design
A descriptive correlational design was used to explore the effects of subjective
beliefs of filial obligation on Chinese caregivers' health. A descriptive design is useful to
describe the essential components of concepts and the relationships being studied
(Habennan & Lewis, 1990).
A causal modeling approach was used to identify and test hypothesized causal
relationships within a conceptual firamework. The advantage of causal modeling is to allow
researchers to test hypotheses and make causal inferences, without applying experimental
manipulation and random assignment (Joreskog, 1973; Wiley, 1973). It is also useful in
testing theory in a situation where the theory is underdeveloped or when an experimental
design cannot be used due to ethical considerauons (Bentler, 1980). Therefore, causal
modeling can maximize the use of empirical data in a nonexperimental design.
A modified cross-cultural approach was also applied to test the phenomenon —
caregiving experiences in one specific ethnic group (Chinese caregivers) living in the
Um'ted States. Although this study was not designed to compare or test cultural distance
across cultures, theoretical perspectives and instruments were originally developed in
English. 89
To further describe the research design, in the following sections, steps and the philosophical underpinnings for causal modeling are addressed. Statistical techmques for causal model testing are described. Furthermore, relevant knowledge about the cross- cultural method including essential components of a cross-cultural design and procedures for cultural equivalences are presented.
Causal Model Testing
Causal modeling is a method founded on the philosophical notions of causality. It is designed to test and generate theory based on scientific processes and empirical data
(James, Mulaik, & Brett, 1982). Causal inferences can be made from the empirical data based on the underlying model when the operationalized hypothesized causal relationships among variables in a theoretical model are supported by empirical data through a modeling process.
Two interconnected steps are involved &i causal modeling. The first step is theory building. The second step is model testing by applying statistical procedures such as structural equation modeling using computer programs, for example, equations (HQS)
(Bentler, 1995), linear structural relationships (USREL) (Joreskon & Sorbom, 1993) or
Analysis of Moment structural equations (AMOS) (Arbuckle, 1997). In the first step,
theory building, a causal model is postulated by identifying a set of variables and defining causal relationships among those variables based on the requ^ments of causality
(Blalock, 1985). In this step, hypothesced causal explanations for a. phenomenon of
interest were developed based on previous research or theories. 90
Three philosophical underpinnings for causality are required for causal modeling:
(I) a temporal ordering, (2) measurable relationships and (3) no spurious relationships among variables (Kenny, 1979). First, a temporal ordering between the variable of X and
Y implies that the cause (X) must happen before effect (Y). For example, if high blood pressure is a cause of stroke, the high blood pressure must occur before the stroke.
Second, there must be an observed and measurable relationship between the variables X and Y. Third, there must be no spurious relationship, which means that the relationship between two variables must not be the result of a third variable that is related to both. In this study, all three conditions were met by specifying the hypothesized causal relationships between the major variables within the theoretical framework and by assuming all major variables were included in the theoretical framework based on literature review.
During recent years, statistical techniques have been developed for causal modeling. Two sophisticated statistical techniques, path analysis and structural equation modeling (SEM), are widely used for testing causal models (Bender, 1998; Blalock, 1961,
1963,1964,1965; Byrne, 1994:Joreskog, 1973; Kenny, 1979; Wiley, L973). Although both techniques rely on linear multiple regression to examine patterns of causation m a set of variables withm a causal model, SEM has several advantages over path, analysis based on the different estimation methods used. Path analysis uses least-squares (LS) estimation, which requires many restrictive assumptions for multiple regression
(PedDiazur, 1982). In addition, path analysis assumes that the variables in the model are 91 measured without error, the residuals are not correlated, and the causal flow is not a reciprocal direction because no recursive model is allowed (Pedhazur).
In contrast, SEM uses other estimation methods such as maximum likelihood (ML) and General Linear Squares (GLS), which are more powerful analysis techniques and avoid the problems of path analysis (Bentler, 1998; Bollen, 1989; Fox, 1984; Joreskog,
1973; Joreskog & Sorbom, 1989). In the equations for SEM, the matrix of coefficients of the effects of all dependent variables is nonsingular. Latent exogenous variables are uncorrelated with the residuals or errors.
As a result, unlike path analysis, SEM can accommodate measurement errors, correlated errors and correlated residuals (Bollen). In SEM, error terms in the model are not correlated with any of the latent variables (Fox, 1984). Measurement error is estimated and removed fi:om the relationships between theoretical constructs to obt;iin a more precise test of the theory (Bollen, 1989). SEM can also be used to test nonrecursive models, which refers to models with two-way paths (Bollen).
Furthermore, SEM allows researchers to test the measurement of theoretical constructs and a theoretical structural model simultaneously, which is impossible using path analysis (Bollen, 1989). In SEM, construct validity can be examined, by testing the measurement of theoretical constructs. SEM allows "evaluation of an entire model, which brings a more macro-level perspective to the analysis" (Rex, 1998). In other words, it allows the researcher to understand patterns of correlations among a set of variables within a whole model rather than, individual effects. The crucial component in SEM is to dedde that "should it be accepted, rejected, or revised" by using fit indices, which will be 92 described later. If it is rejected, individual effects wiH not be interpreted meaningfutty. A revised model is proposed. In this study, SEM was chosen to test the hypothesized causal relationships proposed in the conceptual model.
In SEM, based on the Bentler-Week's system (1980), all variables are hypothesized as either measured (observed) variables or unmeasured (latent) variables (Byrne, 1995). A latent variable is regarded as a factor measured by at least two or more observed variables.
Each observed and latent variable has an associated error (residual) (Byrne) based on classical measurement theory, which the total observed scores for each variable is equal the true score plus measurement error (Walsh & Betz, 1985). SEM is a statistical technique for testing how well the data fit the hypothesized model. In another word, SEM tests how much discrepancy there is between the observed data and hypothesized model
(Byren, 1995). Simply, the relationship between the data and the model can be stated as data equal model plus residual; Data =Model + Residual (Byrne).
In SEM, a full latent model, either standard or non-standardized contains two types of models, the measurement model and the structural model. The correlation coefficients between variables within the measurement and structural model are referred to as path coefficients or parameters (Byrne). The parameters in SEM include the regression coefficients, variance and covariance. The regression coefficients are relationships between independent variables and dependent variables. The variances are residual errors assodated with dependent variables, either unobserved latent or observed variables.
Covariances are correlation coefficients between independent variables. 93
In the structural model, the direction and strength of the paths between two variables indicates the direction and degree of causal relationship between these two variables (Byrne, 1995). The structural model tests how well the hypothesized paths fit with the data. The model presented in Chapter I (Rgure I, page 34), is an example of structural model guiding this investigation. In the measurement model, the strengths of the paths between the latent variables and the observed variables indicate the degree of convergence between the latent and observed variables. In other words, the measurement models test the construct validity of the measurement scales used for the latent variables.
The measurement models for this study will be described in the section of hypothesized causal statements in the conceptual model and presented in figure 2 (page 106).
For the model diagram, to differentiate the regression coeffidents in the measurement model firom those in the structure model, different Greek letters are suggested (Byrne, 1995, p.48). The lambda coefficients (X) refer to the coefficients between indicators and their latent factors in measurement model. The beta coeffidents
(P) refer to coefficients between endogenous variables and gamma (y) refers to the coefficients between exogenous and one endogenous variable in structure model. Those symbols will be used for the figures in the study.
Modified Cross-ultural Approach
Overall, cross-cultural studies have either operational or comparative purposes
(Werner & Compbell, 1980). An operational cross-cultural study is designed to test the cultural differences between groups and the way ui which acculturation influences perceptions and behaviors (Werner &; CompbeD). When two or more phenomena in 94 different cultural groups are compared, the study is considered to be a comparative study
(Werner & Campbell). This study was designed to explore the effect of subjective beliefs of filial obligation on caregivers' health among Chinese caregivers in the United States by applying theory and instruments developed in English. For the purpose of this study, the operational design was used to develop a modified cross-cultural approach. Even though, the current study was not designed to test cultural differences between two groups or to study acculturation, it did explore the applicability of theoretical assumptions made for one group to another group. To further understand the nature of cross-cultural research, the essential components of and strategies for cultural equivalence are reviewed. Processes of cultural equivalence for cross-cultural designs are also presented.
The important goal of cultural equivalence is to enable the researcher to describe the universaUty or specificity of phenomena of interest among cultures in a common underlying process, it was also called an identical categorical system (Berry, 1980). Three essential components were suggested for cultural equivalence: functional, conceptual
(language), and metric equivalence (Berry & Dasen, 1980). Functional equivalence concerns the measurement validity in relation to the equivalence of behaviors or experiences naturally occurring in two or more ethmc groups being studied (Berry). For fiinctional equivalence, it is not necessary for the meaning of behaviors and experience to be identical (Berry, 1992). For instance, family caregiving is a universal phenomenon but it is not identical across cultures.
Conceptual (language) equivalence stresses the measurement validity assodated with the conceptual and operational de&iitions in the original and target language (Phillips 95 et al, 1994). It should be addressed when measures developed for one population are used in another population (Phillips et al). A series of procedures for conceptual equivalence (translation and validation of translations) were identified in literatures and will be discussed in the following sections.
Metric equivalence refers to emic and etic concepts (Berry, 1992). The emic concept is derived firom the culture under study. It implies culturally specific behaviors or experiences that only exist in a particular culture, for example, worship of the ancestors among Chinese people. The emic concept might be drawn firom the transcript of a qualitative study (Jones, 1992). In contrast, the etic concept is culture-firee or universal concept that can be measured through vah'd instrumention because it operates in more than one society, for example, caregiver burden. The etic concept can be derived firom the research standard, as an instrument uniquely described by cultural standards (Jones, 1987).
Metric equivalence focuses on the psychometric properties of cross-ethnic measurement vah'dity. It should be considered when new mstruments are developed because no current instrument is available.
The purpose of this study was to test a theoretical model by adapting available instruments in English (one cultural group) rather than developing new instruments for
Chinese caregivers (another cultural group) or comparing two cultures. A modified cross- cultural design, which focused on assuring measures were culturaEy and linguistically appropriate for Chinese caregivers and interpreting what they were intended to measure, was applied. Three distinct but mter-related strategies were used: (1) assurance of 96 functional equivalence for selected instruments, (2) translation and validation of translation, and (3) ethic identity in recruitment.
Assurance of functional equivalence for the selected instruments is the &st essential step for a cross-cultural study (Phillips et al., 1994). This step helped to (1) understand the meaning of study concepts, (2) examine the relevance of the concept for various ethnic groups such as Chinese around the world, (3) identify appropriate instruments for the study and (4) assess the items on the selected instruments for applicabffity to the target culture. Two strategies, (1) a review of h'terature on the selected problem and (2) expert experiences related to the selected problem (Phillips et al.) were used in this study. Literature on major variables including filial obligation, perceived burden, perceived reward, use of coping and health status relevant to family caregiving for various ethmc groups around the world were extensively reviewed. The review of literature revealed that family caregiving is functionally equivalent among cultures although the meamngs and experiences are not identical. Eleven measures were selected based on the evaluation of applicability and appropriateness for Chinese caregivers. Of them, two were Chinese origin and nine were English origin. All measures were entirely applied except for the two Chinese indicators, which were partially adopted due to the limitation of their applicability for Chinese caregivers. The second strategy was used for the two partially adopted measures. This strategy helped to establish content validity by inviting four Chinese experts selecting relevant and appropriate items from the Chinese origin instrument. In the section on mstrumentation, the detailed procedures for selecting 97 and vaUdating items for the two partially adopted indicators are described. The reh'ability, validity and scoring for each selected measure is also presented.
Translation and validation of translation are essential and required procedures for any cross-cultural research (Brislin, 1970). In 1994, Phillips and colleagues suggested that no matter what the purpose of a cross-cultural study, comparative or operational, conceptual equivalence, which concerns conceptualization and instrumentation is required.
Translation refers to transfer of thoughts or ideas from one language to another (target) language (Brislin, 1970). In terms of translation, Irvine and Carroll (1980) identify two types of translation, asymmetrical and symmetrical, based on the goal of study. They believe that the goal of a study guides translation procedures and interpretation of research findings. Therefore, to validly interpret study findings, translation procedures should correspond with the goal of study. An. operational cross-cultural study requires asymmetrical translation, which means the translation must be loyal to the original language. This often results in awkward wording in the target language (Irvine & Caroll,
1980). A new vocabulary is introduced when concepts are unavailable in the target language (Werner & Campbell, 1980). For a comparative cross-cultural study, symmetrical translation is needed. Symmetrical translation relates assessing conceptual equivalence across cultures (Brislm, 1970). It focuses on equal famiKarity and coDoquialness in both the source and target languages and identification of common items in both cultures (Chapman & Carter, 1979). For a symmetrical translation, some items on the original instrument may require change to achieve conceptual equivalence across cultures (Jones, 1992). To assure a good quality translation, Brislin (1970) suggests ten principles. These ten principles are (1) use of short, simple sentences less than 16 words, (2) use active instead of passive words, (3) employ specific instead of general terms (e.g., using cow or chicken instead of hvestock), (4) repeat nouns instead of pronouns, (5) add sentences that provide context for key words, (6) avoid use of vague word to mean something (e.g. probably), (7) avoid use of subjective mood words such as could and would, (8) avoid use of verbs or proposition to tell when or where (e.g., upper or beyond), (9) avoid use of colloquialisms or metaphors because such phrases are not likely to have equivalents in a target language, and (10) avoid use of two verbs that suggest two different actions. All principles were applied for the Chinese language translation, except for the first principle because it is not appropriate for Chinese translation.
Validation of translation (instruments) was done after translation. Four usefiil procedures for validating translations suggested by Brislin (1970) were used. These four procedures are which are (1) back-translation, (2) the bilingual technique, (3) the committee approach, and (4) the pre-test procedure. Back-translation is the transfer of thoughts or ideas from the target language back to the original language to generate a second original language version (Brislin). The bilingual techmque involves asking a bilingual person to answer the test, or having a group of bilinguals take different halves of the test, and compare the answers (Brislin). The committee approach (also called expert opinions) mvolves having a group of bilinguals involved in the process of validation of translation (Brisin). The pre-test procedure involves a field study in which the newly translated target version instrument is given to a group of people to examine the 99
readability of the new instrument, and to assure that the people who are expected to
respond the test can comprehend the materials (Phillips et al., 1994). AH procedures were applied and the specific steps are described in the section on translation and validation of
translation.
Ethnic identity is important for recruitment. The ethmc identity refers to self-
identification of ethnic background or belonging. Since cultural diversity is a social
phenomenon, how to differentiate culturally diverse populations should be considered
when conducting a cross-cultural research (Phillips et al., 1996). Ethmc identity should be
the first step to recnut the target population for study involving a cross-cultural
methodology because many ethnic groups have many subgroups, (Phillips et al., 1994).
For example, Americans consist of European Americans, Afincan Americans, Hispanics
Americans, Asian Americans, Chinese Americans and so on. For this study, conscious
awareness of a participant's ethmc identity was used to recruit the target population
(Aboud, 1988; Phillips et al., 1996).
Hvpothesized Causal Statements in the Conceptual Model
The purpose of this study was to identify causal relationships among factors that
contribute to caregivers' health, status. Specifically the study attempted to examine the
impact of subjective beliefs of filial obligation on caregivers' health through the mediated
effects of perceived burden, perceived reward and use of coping options in Chinese
caregivers. Eleven variables were used to propose a non-standard ftiH model.
Figure 2 is the complete original model with four measurement models and two
manifest variables. The capital letter "F" mdicates the latent variables. The capital letter 100
"V" reflects manifest variables. The capital letter "E" means the residual variation in manifest variables. The four measurement models within the original model are shown in figure 3. The first measurement model was formed by the latent of subjective beliefs of fih'al obligation and indexed by two manifest variables, fib'al values and filial affection. The second measurement model wa.s CTeated for the latent of perceived burdens and measured by three manifest variables: personal-sodal cost, physical-emotional cost and economic cost. The third measurement model was for the latent of use of coping and measured by behavioral-oriented coping and cogm'tive-oriented coping. The fourth measurement model is the latent of health status. It was hypothesized an outcome variable and measured by two manifest variables, physical health and mental health.
Within the original model, there was one purely exogenous variable (subjective beliefs of traditional fib'al obligation) and one true endogenous variable (health status).
Four mediators fiinctioned as exogenous and endogenous variables. The four mediators were subjective feelings about role requirements, perceived burdens, perceived rewards and use of coping. Figure 2. The Original Model + Y
-P
El •
E2—p. V2
ElO Ell
E3 E7 Note. Fl; Subjective Beliefs of Traditional Filial Obligation VI: Filial Values V2: Filial Affection F2: Perceived Burdens V4: Personal-social Cost V5: Physical-emotional Cost V6: Economic Cost F3: Use of Coping V8: Behavioral-oriented Coping V9: Cognitive-oriented Coping F4: Health Status VIO: Physical Health V11 Mental Health V3: Subjective Feelings about Role Requirements V7: Perceived Rewards Personal- Physical- Economic Physical Mental emotional social Cost Health Health Cost
Subjective Beliefs of Perceived Use of Traditional Filial Burdens Coping Obligation Health Status
Values Affection Behavioral- Cognitive- oriented oriented Coping Coping
Figure 3 Measurement Models 103
Five direct and four indirect causal hypotheses (statements) were specified to predict the effects of subjective beliefs of filial obligation on caregivers' health status.
Each relationship specified in the structural model represented a specific causal statement that was tested. In the following hypotheses, the endogenous variable (health status) predicted by exogenous variable(s), including direct and indirect effects are described.
Direct Effects
H1. Caregivers' health is directly positively affected by subjective beliefs o f traditional filial obligation, perceived rewards and use of coping, and directly negatively impacted by perceived burdens.
H2. Use of coping is directly positively influenced by perceived burdens.
ED. Perceived burdens are directly affected by subjective beliefs of traditional fiUal obligation.
H4. Perceived rewards are positively influenced by subjective beliefs of traditional filial obUgation and negatively affected by subjective feehngs about role requirements.
H5. Subjective feelings about role requirements is negatively influenced by subjective behefs of traditional filial obligation.
Indirect Effects
HI. Caregivers' health is indirectly affected by subjective behefs of traditional filial obligation through the mediated effect of perceived burdens.
E£2. Caregivers' health is indirectly affected by subjective beliefs of traditional
filial obligation through the mediated effects of perceived burdens and use of coping. 104
H3. Caregivers' health is indirectly affected by subjective beUefs of traditional filial obligation through the mediated effect of perceived rewards.
H4. Caregivers' health is indirectly affected by subjective behefs of traditional fih'al obligation through the mediated effects of perceived rewards and subjective feelings about role requirements.
Sampling Procedures
In this section, sampling procedures including ethical considerations, subject selection criteria, data collection sites and recruiting are described.
Ethical Considerations
Prior to data collection, approval from the human Subject Committee of the
University of Arizona was obtained (Appendix A). A written disclaimer with a statement about withdrawing at any time was used (Appendix B). In the disclaimer, the purpose of the study was briefly described. No names were used to assure confidentiality.
Subject Slection Criteria
The criterion for sample selection was Chinese primary caregivers who (1) had carried primary responsibility for providing assistance with at least three activities of daily living to a family member over the age of 55, (2) had provided informal support to their
family member for at least 3 months while the elder lived in the United States, and (3)
were able to speak, read, and write adequate Chmese or English.
Data Collection Sites
Sevea cities from five states including Arizona, California, Georgia, Maryland, and
Massachusetts were used for data collection sites. The involved dties mcluded Atlanta, 105
Boston, Los Angeles, San Francisco, Washington District ofColorabia, Phoenix and
Tucson. The seven cities were selected based on the density of Chinese population and accessibiUty of recruitment. Table 1 lists the number of Chinese people among Asian population in the selected cities.
Recruiting
The invesugator and six masters prepared recruiting representatives conducted the recruitment for this study. The criterion for selection the recruiting representatives was
(1) their education background qualified them to be recruiting representatives, (2) they lived in cities with large Chinese populations and (3) they held leadership positions in the
Chinese commum'ties. A $15 to $40 dollar gift was offered to the recnuting representatives in appreciation of their help. In addition, for those who partidpated, a thank-you card with sincere expressing appredation was sent to them.
Prior to recruitment, an initial estimation of possible number of subjects was obtained from each of the recruiting representative. A total of 190 pssible subjects was estimated. Representatives were sent recruitment packages, based on their estimates.
Two to three extra packages were also mailed to each selected site. Rigorous training for data coDection was given to the seven recruiting representatives. The training procedures are described in the following section. 106
Table 1
Chinese among Asian Population in the Study Sites
Chinese Asian
Atlanta 27446 173170 Boston 84392 238124 Phoenix 5761 26449 Tucson 4390 17213 Los Angeles 63075 369254 San Francisco 152630 239565 Washington District Colombia 3734 15189 107
Training oFRecruiting Representatives
To avoid sampling bias as much as possible, prior to data collection, a standard data collection plan was described to each recruiting representative by phone. Rve procedures were involved in the standard data collection plan. First, each study package was double checked with the investigator to assure it contained a study disclaimer, an envelope with return address, agift of 10 books of 33 cents of U.S. postage stamps, and a booklet of questionnaires. Each questionnaire booklet had eight pages including one participant proRle form and eleven instruments.
Second, the purpose of the study and sampling criterion were briefly described.
This step helped the recruiting representauves get a sense of the linkage between the purpose of study and the sampling criterion. For this step, verbal comments indicative of understanding the linkage between the purpose of study and the criterion of sampling were obtained.
Third, the principal investigator explained each questionnaire, item by item, to clarify any possible question, which might be raised by the recruiting representative. This step helped the recruiting representatives get a sense of how to answer possible questions the participants might have. All representatives indicated they would answer any possible question based on their best knowledge about this study. If they could not answer, they would refer the questions back to the prmcipal investigator. These questions were answered directly by the principal investigator or indirectly through the recruiting representative who referred the question on subjects' behalf. 108
Fourth, recruiting representatives and the investigator (whoever mailed study packages to subjects) called subjects one-week after study packages were distributed.
This step helped to assure the complete study package was noailed to the intended subject.
To enhance the response rate as much as possible, two reminder calls were given to subjects who did not return study booklets within the expected time. The time schedule for making reminder phone calls was as follows. The first reminder phone call was made two weeks after the subject received the study packages. If the study booklet still was not after two weeks, the second reminder phone call was made to remind the subject to return it.
Data Collection Protocol
Several steps were involved in the data collection protocol. First, potential subjects were identified either by the investigator or the recruiting representatives.
Second, once potential subjects were identified, a brief description of the study was given
by the investigator or recruiting representatives (whoever identified them), either through
face-to-face contact or phone contact depending on their preference. Information about
the purpose of the study, the gift of appredation, and the protection of confidentiality was
described to the potential subjects. They were told 25 to 35 minutes were needed to
complete the questionnaires.
Third, if the potential subjects agreed to participate, a study package was
distributed to them. The participants' names, addresses, phone numbers were obtained, in
case fiirther contact was necessary. One week after the study package was distributed, a 109
phone call was made fay the investigator or the recruiting representatives (whoever distributed the study packages) to assure the complete study package was received.
Fifth, if the study booklets was not returned, two reminder calls two weeks apart
were made, by the investigator or the recruiting representatives to remind the partidpants
to return the study booklets.
Instrumentations
The participant profile was used to obtain baseline information about caregivers
and care recipients. Variables used to describe the caregivers included: age, gender,
socioeconomic status (SES), length of time in the Um'ted States, level of involvement with
and relationship to the care receiver. Caregiver's education, employment status,
occupation, annual household income were used to reflect their socio-economic status
(SES). Information related to care recipients included age, gender, type of health
problems, cogm'tive status, mobility and types of assistance needed.
Eleven measures were used. Of these, there were two subscaies from the Yang's
Mai Obligation Scale (YFOS), a 7-item of subscale from Poulshock and Deimling's
burden inventory (BI); the 24-item Picot's caregivers reward scale (PCRS), three subscaies
from the Costs of Care Index (CCI), two subscaies from the Ways of Coping
Questionnaire (WOQ and two subscaies from the Short Form 36 Health Survey (SF-36).
The two subscaies from the YFOS were the revised Yang's Filial Values Subscale (YFVS-
R) and the revised Yang's Filial Affection Subscale (YFAS-R). The three subscaies from
the CCI were the Personal-Social Cost Subscale (PSCS), the Physical-Emotional Cost
Subscale (PECS), and the Economic Cost subscale (ECS). The two subscaies from the 110 woe were the Behavioral-Oriented subscale (BOS) and the Cognitive-Oriented subscale
(COS). The two subscales from SF-36 include were the Physical Component Summary
Subscale (PCSS) and the Mental Component Summary Subscale (MCSS).
The YFVS-R and YFAS-Rwere originally developed in Chinese and had only been tested in healthy Chinese adults. Since this study was the first time they were used with
Chinese caregivers, the processes for initialing and estabb'shing content and face validity for the items selected is addressed in the following section. The remainding nine measures were originally developed in English. The appropriateness of the instrument for measuring the theoretical concept, the scaling and reUabiKty and validity are presented. To assure the concept equivalence between English and Chinese, they were translated into Chinese, after permission for translation was obtained. The processes for conceptual equivalency
(translation and validation of translation) for all EngUsh instruments will be described later.
The Revised Yang's Filial Values Subscale (YFVS-R)
The YFVS-R was used to measure the degree to which the subjects agree with traditional 6Ual values. The YFVS-R is an 8-item subscale adopted from Yang's fiUal obh'gation scale (YFOS) (Yang, Yeh, & Huang, 1988).
The original YFOS contains four subscales including beliefs, affection, attitude and behavior subscales. For the beliefs, attitude and behavior subscale, a pool of 86 items derived from a literature review and tested for face validity with a large sample of over
10,000 healthy people, consisting of non-student adults and students from middle school to college. Three different verb phrases were used with the same 86 items to generate three subscales: filial belief, filial attitude, and filial behavior subscales (Yang, Yeh, & Ill
Huang). For example, the item of "Do vou agree that voa should sacrifice yourself to take care of your parents?" is for filial belief subscale. When the same item was change to 'T)o vou agree that vou will sacrifice yourself to take care of your parents?", it belonged to the filial attitude subscale. Similarly, when the same item was changed to 'T)o vou agree that vou really sacrifice yourself to take care of your parents?", it became an item in the filial behavior subscale.
A 52-item instrument with four factors was derived from the original 86 items.
The four factors are (1) paying respect parents, (2) giving obedience to parents, (3) taking care of parents, and (4) protecting and glori^ng parents. Of the 52 items, there were 20 items for paying respect to parents, 12 items for obedience to parents, 10 items for taking care of parents, and 10 items for protecting and glorifying parents. The 52 items used 4- point Likert type scales with response ranging from "not at all" = 0 to "very much" = 3.
The possible total scores for each subscale ranged fromO to 156.
For the Filial Affection subscale, there were 19 items on caregivers' positive and negative affection toward their parents. Of the 19 items, 12 are positive affection and 7 are negative affection. The Filial Affection subscale was a 4-point Likert type scale with response ranging from "not at all" = 0 to "very much" = 3. The possible total scores range fromO to 57.
The Yang's Enlial Obligation Scale (YFOS) was used because it was originally developed for Chinese. It appeared simple and easy to admimster. Items reflected four common attributes of the concept of filial obligation embodied in Chinese culture. The
YFOS had been tested on a large heterogeneous population including students ranging 112 from, middle school to coDege and non-students. The internal consistency reliabiKty for different subscales varied from alpha = .46 to .96 on different populations. For the population of non-student adults, the internal consistency coefficients for different subscales were alpha = .66 to .93.
For this study, in order to ensure that items were relevant to family caregiving and applicable for this study, eight items were adopted from the revised YFOS to create the revised Yang's Filial Values Subscale (YFVS-R) to determine the degree in which the subject ascribed traditional filial values. Eight items with the verb phrase "I feel that I should" were used. A. 4-point LQcert scale with response ranging from "not at all" =0, "a little" = I, "moderate" = 2 to "very much" = 3. The possible total scores will be from 0 to
24.
Items were selected based on an extensive literature review on the attributes of the filial concept in Chinese culture. Items were also evaluated by an expert panel to establish content validity. The expert panel consisted of four bilinguals; one who had earned a
Ph.D. in the field of nursing, one who had earned a Ph JD. in the library science, one who had earned a post doctoral degree in the field of chemistry and one who was a medical doctor. They were all foreign-bom Chinese immigrants. Face validity was established by conducting a pilot study on six Chinese adults. All agreed no change in the items was needed. The same criteria were used for selecting items for the revised Yang's Filial
Affection subscale (YFAS-R), which is described in the followmg section. 113
The Revised Yang's Fih'al Affection Subscale CYFAS-R)
The YFAS-R is another 8-item subscale adopted from the YFOS to measure the degree in which the subjects indicated their affection toward their parents. To assure that items were relevant to family caregiving and applicable for this study, the same criteria for
YFVS-R were used to develop the YFAS-R. Only items relevant to family caregving were selected. The YFAS-R consists of 4 positive affection and 4 negative affection items. It a4-point Likert scale with response ranging from "not at all" =0, "a little" = 1,
"moderate" = 2 to "very much" = 3. The possible total scores will be from 0 to 24.
Poulshock and Deimling Burden Inventorv (BP
Seven items from the Poulshock and Deimling's BI (1984) were used to index the subjects' subjective feelings about performing activities daily living (ADL) tasks required
by family caregiving role. The seven items were survey questions that addressed negative emotional responses toward specific tasks performance to capture whether caregivers felt
upset about providing ADL tasks. There were five response options including "no help
needed", "help needed but no problem at all", "help needed and tiresome", "help needed
and difficult", and "help needed and upsetting. The seven items were simple and easy to
admimstrate. .
In terms of scoring, for answers of either "no help needed" or "no problem at all",
a value of 0 was assigned. For answers of "tiresome", "difficult" or "upsettmg", a value
of 1 was assigned. For the combination answers of "no problem at all" and "tiresome", a
value of 1 was assigned. Similarly, for answers of "no problem at all" and "difficult", a IL4 value of I was assigned. For answers of "no problem at all" and "upsetting", a value of L was also assigned.
Other combination of responses, for example, answers of "tiresome" and.
"difficult", "tiresome" and "upsetting", or "difficult" and "upsetting", a value of 2 was assigned for each. For the combination answers of "tiresome", "difficult" and "upsetting", a value of 3 was assigned. Therefore, the possible total score for each item ranged from 0 to 3. A higher score indicated a greater negative feeling about providing assistance. The possible total score for this scale ranged from 0 to 21.
Personal-Social Cost Subscale (PSCS)
The PSCS was used to measure the degree to which the subjects perceived personal and social impacts on their life in response to the family caregiving role. The
PSCS is one of the subscales from the Cost of Care Index (CCI) (Kosberg & Cairl, 1986).
The CCI contains 20 items to identify five specific problems related to family caregiving. The five areas of problems were (1) personal-social cost, (2) physical- emotional cost, (3) economic cost, (4) perceived caregiving as a provoking cost, and (5) perceived caregiving as a value investment cost. The CQ was originally designed, to provide a fiiH psychometric development of indices for measuring burden (Kosberg &
Cairl, 1986). Subjectivestatementsofcaregivers wereusedtoconstructthisscale. 'Tfeel that as a result of caring for my elderly relative I do not have enough, time for myself' is as sample item from the scale. The CQ is a 4-point Likert scale with responses ranging from
"strongly disagree" = I, "disagree" =2, "agree" =3, to "strongly agree" =4. The possible total scores range from 20 to 80. The internal consistency reliability of the CQ was alpha LL5
= .91 for the overall scale (Kosberg & Cairl). Construct validity was established in relationship to caregiver and care receiver characteristics.
The CCI appeared to be a clear, specific, useful and reliable tool (Kosberg & Cairl,
1986). It had also been validated in other cultures such as African American, Chinese and
Japanese with good reliability results (Chou, 1998; Mackenzie, Lee, Dubley-Brown, &
Chin, 1998; Mizoguchi, lijma, Niino, &Orimo, 1995; Picot, 1995). In 1995, the CCI was validated on a sample of 42 Japanese caregivers of elders, 27 with dementia and 15 with other neurological disorders, and obtained a good internal consistency coefficient alpha
S3 was obtained (Mizoguchi, lijma, Niino & Orimo). In the same year, Picot applied it to
81 Affican American caregivers; however, no informauon about the reliability and validity was provided. Three years later, in 1998, Mackenzie, Lee, Dubley-Brown and Chin admimstered it to evaluate a pilot case management project in Hong Kong. Similarly, no information about the rebability and validity was provided. In the same year, the economic subscale was validated on 150 Chinese caregivers of demented elders in Taiwan and a good internal consistency coefficient of alpha .93 was obtained (Chou).
For this study, the Personal-Social Cost Subscale (PSCS) and the other subscales, the Physical-Emotional Cost Subscale (PECS) and Economic Cost Subscale (ECS) were used. The PSCS consists of 4 items and used a 4-pomt rating scale with responses ranging from "strongly disagree" = 1, "disagree" = 2, "agree" = 3, to "strongly agree" = 4.
The possible total scores range from 4 to 16. 116
Phvsical-Emotional Cost Subscale (PECS)
The PECS was used to measure the degree to which the subjects perceive physical and emotional impact on their life in relation to family caregiving role. It is another subscale from the Cost of Care Index (CCI) (Kosberg & Cairl, 1986). The PSCS contains
4 items with responses ranging "strongly disagree" = 1, "disagree" = 2, "agree" = 3, to
"strongly agree" = 4. The possible total scores range from4 to L6.
Economical Cost Subscale (ECS)
The ECS was used to measure the concept of personal and social burden. It is another subscale from the Cost of Care Index (CCI) (Kosberg & Cairl, 1986). The PSCS has 4 items with responses of "strongly disagree" = I, "disagree" = 2, "agree" = 3, to
"strongly agree" =4. The possible total scores range from 4 to 16.
Picot's Caregiver Reward Scale fPCRSI
The PCRS was used to index the degree to which the subjects perceived positive aspects of caregiving including pleasant feelings, positive consequences and satisfaction in response to family caregving role. The original PCRS was a unidimensional scale with 24 items derived from literature review (Picot, 1991). The PCRS was a 5-point rating scale with responses ranging from "not at all" =0, "a little bit" = I, "somewhat" =2, "a quite bit"
= 3, to "a great deal" =4. The total score of the PCRS was calculated to reflect the overall rewards. The possible scores were 0 to 96.
The original 24 items of PCRS were tested usmg a nonrandom sample of 83 back
female caregivers and a random mixed sample of256 black: and white female and male L17 caregivers (Picot, 1990; L991). A revised l6-itera scale was identified with internal consistency coefficients of alpha = .83 to .88.
For this study, the original 24 items of PCRS were used as suggested by the author since it was being tested for the first time with Chinese. In addition, it was a simple, comprehensive and reliable tool (Picot).
Behavioral-Oriented Subscale (BOS')
The BOS was used to measure the extent to which the subjects used behavioral
reactions for managing stressful situations resulted firom family caregiving role. The BOS
was derived from the revised Ways of Coping Scale (WOC-R) (Folkman 8c Lazarus,
1988).
The im'tial WOC-R is a 66-item scale with eight factors reflecting the complexity of
coping options. In im'tial testing, the authors obtained an alpha = .70 for the overall scale
(Folkman, Lazarus, & Dunkel-Schetter, L986). The eight factors were confrontive
coping, distancing, self-controlling, seeking social support, accepting responsibility,
escape-avoidance, and positive appraisal (Folkman, Lazarus, & Dunkel-Schetter).
Recently, the eight factors of WOC-R have been grouped into two categories: behavioral-
oriented and cognitive-oriented coping strategies. The behavioral-oriented subscale
consists of confrontive coping, seeking social support, accepting responsibility and planful
problem-solve coping. The cognitive-oriented subscale encompasses distancing, self-
controlling, escape-avoidance, and positive appraisal. These two subscales were validated
on a sample of 150 Chinese caregivers in Taiwan. Internal consistency of alpha .86 for 118 problem-oriented subscale and .79 for emotion-oriented subscale were obtained (Chou,
1998; Chou, LaMontagne, & Hepworth, 1999).
The WOC-R is a well-known tool and has been widely used by numerous researchers for a variety of samples. It has been tested with healthy individuals (Folkman
& Lazarus, 1985; Smyth &Yarandi, 1996), clinical populations (Cohen, 1991; Mishel &
Sorenson, 1993; Wineman, Durand, & Steiner, 1994), and caregivers with different types of dependent relatives (Hooker, Frazier, & Monahan, 1994; O'Brien, 1993; Phillips &
Thomas, 1996; & Rosengard & Folkman, 1997). It has also demonstrated cross-cultural validity in Chinese Caregivers in Taiwaa (Chou, LaMontagne, & Hepworth, 1999), and
African American Caregivers in the United States (Smyth & Yarandi, 1996).
In this study, the BOS consisted of 39 items. It was 4-point Likert type scale with response from "not used" = 0, "used somewhat" =1, "used quite a bit' =2, to "used a great deal" = 3. The possible total scores range from 0 to 87. The BOS has been tested in a sample of 150 Chinese caregivers in Taiwan. A good internal consistency of alpha = .86 was obtained (Chou, 1998; Chou, LaMontagne, & Hepworth, 1999).
Cogm'tive-Oriented Subscale fCOS)
The COS was used to index the extent to which the subjects used cognitive reactions for managing stressful situations resulted from family caregiving role. It contained 39 items derived from the revised 66-item Ways of coping scale (WOC)
(Folkman & Lazarus, 1988). The COS is also used a 4-point Likert type scale with
response from "not used" =0, "used somewhat" = 1, "used quite a fait" =2, to "used a great
deal" =3. The possible total scores range fromO to 111. The BOS had an acceptable 119 internal consistence coefficient of alpha=J9 with same 150 Chinese caregivers in Taiwan
(Chou, LaMontagne, & Hepworth, 1999).
Physical Component Summary Subscale fPCSSI
The PCSS was used to determine the degree to which the subjects perceived their own state of complete physical health. It is one of two subscales from the 36-item Short-
Form Health Survey (SF-36) (Ware et al., 1993).
The SF-36 is a combination of Likert-type and dichotomous items. The 36-item
Short-Form Health Survey (SF-36) is a generic measure of health status on eight dimensions of well-being (Ware & Sherboume, 1992). These eight areas of well-being are
(1) general health perceptions, (2) limitations in physical function, (3) limitations in role-
physical function, (4) bodily pain, (5) general mental health, (6) limitations in social
activities, (7) limitauons in role-emotional function, and (8) vitality (energy and fatigue)
(Ware & Sherboume, 1992). The reUability of the eight scales has been examined by
using internal consistency and test-retest. The internal consistence reliability for each
subscale ranged from alphas over of .70 to .80 (Mchomey et al., 1994; Ware et al., 1993).
Some subscales, for example the physical and mental subscales, even have reliabiUty
coefficients alpha over .90 (Ware et al., 1993). The content validity of the SF-36 was
compared to other widely used generic health surveys (Ware et al., 1995). The method of
known-groups validation was used to establish construct vah'dity in many cross-cultural
and longitudinal studies involvmg physical and mental health (McHomey et al., 1993).
The SF-36 is a well-known and well-developed tool that has been widely used and 120 validated for a large heterogeneous population across cultures incradlug Australia, China,
Denmark, Germany, Sweden, the United States, and Taiwan.
The PCSS contains 21 items with four subscales including (1) general health perceptions (GH), (2) limitations in physical function (PF), (3) limitations in roIe-physical function (RP), and (4) bodily pain (BP). The PCSS consisted of three Likert-type subscales and one dichotomous subscale. The GH is a 5-item of Likert-type scale with responses ranging from "poor" =1 to "excellent" =5 or "defim'tely false" =l to "definitely true" =5. A higher score indicates better health. The PF is a lO-item, Likert-type scale with responses ranging from "yes, limited a lot" = 1 to "no, not limited at all"
= 3. The RP is a 4-item, dichotomous scale with responses ranging from "yes" = 1 to "no"
=2. The BP is a 2-item, Likert-type scale with responses ranging from "very sever" =1 to
"none" =6. Higher scores indicate better physical health. The possible total scores for the
PCSS are 21 to 75.
Mental Component Summary Subscale (MCSS)
The MCSS was used to measure the degree to which the subjects perceived their own state of complete mental and sodal wett-being. The MCSS is the other subscale from the 36-item Short-Form Health Survey (SF-36) (Ware et al., 1993). The MCSS contains
14 items with four subscales including (1) general mental health (MEi), (2) limitations in social function (SF), (3) limitations in role-emotional function (RE), and (4) vitality
(energy and fatigue) (VT). The MCSS contained, three Likert-type subscales, and one dichotomous subscale. The MH is a 5-item, Likert-type scale with responses ranging from
"all the time" =1 to "none of the tune" =6. The SF is a 2-item, Likert-type scale with 121 responses ranging from "extremely" =1 to "not at all" =5. The RE is 2-item, dichotomous subscale with responses ranging from "yes" =1 to "no" =2. The VT is 4-item, Likert-type scale with responses ranging from "none of the time" =l to "all of the time" =6. Higher scores indicate a better mental health. The total possible scores for the PCSS is 14 to 70.
The PCSS has beea tested in various populations in different studies and obtained good reliability with alpha over .90 (Ware et al., 1993).
Summary
Reh'ability of eleven measures was reviewed. Numbers of item and rang of scoring for each measure were address. For the partially apph'ed measures, criteria and justification for selecting items were described.
Processes for cultural equivalence for English version instruments
Four steps were taken to establish cultural equivalence for all English instruments.
These four steps are (1) translation and back-translation, (2) validation of translation, (3)
bilingual technique for evaluating the adequacy of translation and (4) empirical test for
estabhshing face validity.
First. Translation and Back-Translation
Before translating, the meanings of each item in the English versions were double-
checked with two native English speakers to ensure the original intended meanmgs for the
translated versions were clearly understood. Both native English speakers had a graduate
education, one in nursing and the other in sociology.
Instruments were iniu'ally translated mto Chinese by the investigator. The principal
investigator is a bilingual, who was bom in Taiwan and had lived the United States for 122 nearly 7 years. Phillips and colleagues (1996) suggest the use of a bilingual person is essential for developing a reliable instrument. Brislin's (1970) ten translation principles were used to assure the quah'ty of translation. Irvine and Caroll's (1980) asymmetric translation procedure was applied in this study.
After the first Chinese versions were generated, two Chinese bilinguals back- translated them into English to generate the second English versions. One of these translators had earned a master's degree in computer science and had lived in the U.S.A. for more than 30 years. The other had a doctoral degree in economics and had lived the
U.S.A. for 15 years.
Second. Validation of New Generated Chinese instruments
Once back-translations were done, all three translators got together to review and compare both drafts of the new Chinese versions and the second English versions. When there were inconsistencies between the two different languages, discussions among the translators and the two back-translators were conducted to verify meaning and wording between the two languages. Several translations and back-translations were done until agreement of translation among the translators and the two back-translators was obtained.
Third. Bilingual Techmque for Evaluating the Adequacy of Translation
When an agreement was obtained, another two Chinese bilinguals were invited to review and compare both the revised Chinese versions and the original English versions to evaluate the adequacy of translation. One of these two bilinguals was a doctoral nursing student who had lived in the Um'ted States for eight years. The other was a nursing cluiician, with a master degree and had lived in the U.S- for about four years. She was 123 also an experienced editor and worked on school journals and with the Chinese student association. Concept equivalence was asserted; no items were assessed as mistranslated.
Two items in the SF-36 Health Survey Instrument were reworded as suggested by one expert to improve face validity. Other than that, face validity was established as the two reviewers commented that the instruments were well written.
Finally, another Chinese bilingual, an assodate professor of medidne, who had lived in the U.S. about 20 years, was invited to review both English and Chinese versions of the instruments. This evaluator also commented that instruments were written well.
Fourth. Empirical Test for Establishing Face Validitv
After the processes of translation and validation of translarion were completed, a pilot study was conducted with six Chinese persons with various educational levels. Face validity was estabUshed with comments that the items were understandable and readable.
Plan of Data Analysis
Data analysis involved three phases: (1) description of sample profile, (2) psychometric testing for selected mstruments, and (3) testing and evaluating of model adequacy. The Statistical Package for Sodal Science (SPSS 9.0 for window) was used for describing sample profile and psychometric testmg for selected instruments. The EQS computer program on the mainfirame of the University of Arizona was used for model testing and model re-specification. In the following section, activities for each phase are addressed. 124
Phase One - Description of Sample Profile
In this phase, three steps were involved. First, before data entry, data for each questionnaire for each participant were examined for missing data. In general, very minor random missing data occurred except for seven individuals' data that had consistent missing data on one or another item. Data from these seven cases were omitted from analysis. Second, after data entry, two procedures were used to clean data. First, the raw data on spreadsheets were printed and visually checked with the original questionnaires for each individual. Second, frequencies for variables were examined to identify out-of-range values. Third, after data were cleaned, instrument items were reverse coded and weighted according to instructions for each scale. To assure accuracy, frequencies were rerun to identify and correct out of range values. Missing data were also evaluated. Once all procedures for data preparation were done, descriptive analysis for all demographic information was conducted. Frequendes were calculated for categorical data. Central tendency and variability were described for continuous data.
Phase Two - Psvchometric Testing
The reliability and validity of selected instruments were evaluated by conducting internal consistency and factor analysis. For the reliability test, descriptive stadstics and internal consistency were performed and evaluated. Each item was reviewed for its contribution to instrument variance by evaluating the means, ratio of standard deviation to item mean, mter-item correlations and item-total correlations. Standardized coefficient alphas were used as indicators of internal consistency (Cronbach, 1951). The criterion level was .70 for a new psychosodal scale and .80 for a mature scale (Edward, 1970; 125
Nunnally, 1978). The standard deviation for each item was expected to be one half the means. Inter-item correlations were expected rangmg from .30 to .70 for more than half of the items (Gordon, 1968; Keriinger, 1973). The minimum criterion for the item-total correlations was set at .40 or over (Anastasi, 1976: Gordon, 1968).
Construct validity was assessed by factor analysis with principle component extraction and obhque rotation techmques. By factor analysis, scale items were evaluated and identified for whether they were sufficiently intercorrelated to be considered as a factor. The principal component method, a most common factor extraction method, was
used because it assumes all measurement errors are random. The oblique rotation was
used to result in correlated factors because items in a particular scale were assumed to be correlated with one another (PoUt, 1996).
For the determination of number of factors, inspection of scree plots and
eigenvalues were used. A scree plot is a reh'able index with high factor loading (Polit).
The eigenvalues are an index of how much variance in the factor solution is explained by a
given factor. A factor with an eigenvalue greater than one is considered important in
accounting for variance (Long, 1983). Nunnally (1978) suggests factor loadings of
greater than .30 are acceptable for a valid indicator. Keriinger (1973) recommends factor
loading greater than .40 for a valid indicator.
Phase Three - Testing and Evaluating of Model Adequacy
Li this phase, the model was tested by performing the statistical technique of
structural equation modelmg (SEM). Before model testing, assumptions required for 126
SEM and radices used for model evaluation are evaluated and presented. Procedures for model testing are presented.
Requirements for Structural Equation Modeling CSEM)
Requirements for SEM include (I) theoretical and (2) statistical aspects.
Theoretically, three requirements for causation including temporal ordering, measurable relationships among variables, and nonspuriousness are necessary for theory building.
These three requirements have been described earlier and they are not repeated here.
Other requirement for the statistical aspect included (1) normality and linearity of data, (2) multiple indicators of latent variables and (3) adequate sample size. If any of the statistical necessities are violated, the model fit may be affected (Muthen, 1993; West, Finch, &
Curran, 1995).
Normality and Linearity Data
Normality and linearity of data is expected for SEM (Benlter, L995). Since SEM is an alternative form of linear multiple regression analysis, the same assumptions required
for linear multiple regression are required for SEM. Use of nonlinear data such as categorical data increases the tendency to reject the model chi-square test (West, Rnch &
Curran, 1995). Furthermore, according to statistical theory, a normal distribution is need
for estimation methods in computer programs (Benlter). Methods used to estimate the
parameters in SEM vary with different distributions of data. With the most commonly
used estimation method. Maximum likelihood (ML), normality of data is expected
(Benlter). In this study, ordinal or Likert scales were used for measuring all study
variables. Total scores (continuous data) were employed for model testing. The 127 distributions for all variables were examined by conducting descriptive statistics before model testing.
Multiple Indicators of Latent Variables
Ideally, at least three indicators are required for a latent variable to be "just identified" for a measurement model (Byrne, 1995). The indicators should not be redundant in order to avoid linear dependency in model testing (Bollen, 1989). In other words, the indicators should be multidimensional. When items in each indicator are highly correlated, there is tendency for linear dependendes, which cause the measurement model to be statistically underidentiRed (Benlter, 1995). This requirement is also associated with use of estimation methods. With Maximum likelihood (ML) estimation, it is assumed that no single variable can perfectly explain another variable (Bollen). This requirement was partiaEy met in this study because a nonstandard full latent model with four latent and two manifest variables was proposed. In the four latent variables proposed in this study, only one latent variable had three indicators. The other three had two indicators each. When this requirement is violated, it increases the probability in rejecting the model chi-square test (Bentler). Therefore, for this study, the tendency to reject the model chi-square test was anricipated.
Adequate Sample Size
According to the literature review, there is no clear guideline in determining minimum adequate sample size for causal model testing. While BoUen (1985) suggests "at least several cases per free parameter" (p. 286), others recommend several cases per variable (Anindell & Vanderende, 1985; Ferketich & Verran, 1990). For those who 128 suggest uses of variables as the rule for detenriining adequate sample size, Arrindell and
Vanderende (1985) suggest a minimum of 20 cases per variable. Ferketich and Verran beheve a minimum of 10 cases per variable plus 50 is adequate.
Boomsam (1983) indicates that a sample size of 100 or more is required for a model testing with the Maximum Likelihood method (ML). A smaller sample size can lead to reject a true model when the maximum likelihood estimation (ML) is used (West,
Finch & Curran, 1995). For this study, the criteria suggested by Boomsam (1983) and
Ferketich and Verran (1990) were used to determine the adequate sample size. Therefore, an attempt to recruit of a minimum of least 100 to 160 was made.
Indices for Evaluation of Model Adequacv
Unlike regression analysis, which calculates R' for each dependent variable to predict the proportion of true variance for the dependent variables individually, the model- fitting program generates several statistical indices (also called fit indices) to evaluate the overall fit of the model to the data (Rex, 1998). A. goodness of fit index is a test to examine whether the theoretical model is good or not 5t with the data (Bentler, 1995).
According to the Uterature, 5t indices differ with different statistical computer programs.
For this study, several indices used for HQS are reviewed.
The first fit index is the likelihood ratio also called the Pearson Chi-square (x"), which tests the significance of the difference in fit between the hypothesized model and the just identified model (independent) model (Bentler). A non-significant Chi-square value, for example, x" (10) = 29.88, p> .05, means the hypothesized model does not statistically significantly differ firom the just identified model m fitting with data. In contrast, a Chi- 129 square test of x" (10) = 29.88, p< .05 indicates the hypothesized model is significantly worse than the Just identified model in fitting with data. It also means that the hypothesized model is significantly worse than if it had 10 more paths in fitting to the data.
Ideally, a non-significant value of the x" is expected to accept the null hypothesis (H") which means that there is no difference between the data (the sample covariance matrix) and the model (the predicted covariance matrbc) (Byrne, 1995). Since the x' is designed to test the magm'tude of discrepancy between the residual covariance of the theoretical model and the observed data, however, the larger the sample size, the easier it is to have a larger magnitude of residual covariance to reject the null hypothesis of (x"). In other words, there is a tendency to reject a true fit model with a large sample using the x" index.
The second index is the relative Chi-squre also called the (x" I df) ratio.
To dimim'sh the sensitivity of the Chi-square to the sample size, another index, the normed fit index (NFI) was developed (Bentler, 1990; Bentler & Bonett, 1980). The NFI is a practical criterion to test the goodness of fit by examining the proportion of the x" that has been explained by theoretical model. In other words, it tests the proportion in the improvement of the overall fit of the theoretical model relative to the null model.
The formula for NFI is (y^-yV
3C'o
X"o is the X" for null or independent model
x\ is the x'for hypothesized or theoretical model.
Byren (1995) suggested the possible value for the NH ranges firom 0 to 1.0. A value of greater than .90 refers to as a good fit, which means that the hypothesized model 130 is 90% better than the null model (Bentler & Bonett, 1980). Although the NFI is appropriate for complex model fit, the NFI still tends to underestimate model fit in small sample (Bentler, 1995). Inevitably, the NFI still has the limitation that it tends rejecting a true model for a small sample (Bentler, 1990a).
To better adjust the problem with sample size, the nonnormed 6t index (NNFI), a parsimony fit index was developed (Bentler & Bonett).
The formula for NNFI is fyVrifnyf y\l dfr^
xVdfo
dfo is the degree of fireedom for nuH or independent model
dfr is the degree of freedom for hypothesized or theoretical model
Inescapably, NNFI is easier out of the range of 0 to 1.0 when the x\is smaller
than dfr and it will have a negative value when the x\/ dfr is greater than the xVd^
(Bentler, 1995). The NNFI may also overestimate the fit in a small sample as compared to
other indexes (Anderson & Gerbing, 1984).
Finally, the Comparative of fix index (CFI) was created to accommodate the
problem of sample size.
The formula for CFI is! rYVdf,f)-f Y>-dfT) !
X'o-dfo
The CFI takes care of the problem of sample size and is suggested as an
appropriate index for parsimony fit model testing for all sample sizes (Bentler, 1990a;
Byrne, 1995). It tests the absolute of x'over the degree of fireedom. The possible value 131 for CFI ranges from 0 to 1.0 and a value of greater than .90 is suggested as a good parsimony fit model (Bentler & Bonnett, 1980).
Summary
Currently, several fit indices were available for evaluating model adequacy. In addition to Chi-square index, several indices were developed to decrease the sensitive of
Chi-square to sample size. Since each index has different purposes, advantages and limitations, to evaluate the model fit from different perspectives, multiple indices were used for this study.
Procedures for Valuation of Model Adequacy
To assure the results were trustworthy, several important procedures were used for reviewing the output files. First, the covariance matrix was examined to make sure the data file had been correctly read. When data were not appropriately read in the program file or were inappropriately defined in data file, a value of 0 appeared in the covariance matrix. This can occur often when using an external data file for running the model file.
To avoid the problem, the format for each variable specified in model file was checked to assure they were consistent with the format specified in data file.
Second, any error message that indicated problematic specifications was checked.
For example, a warning message such as "user specified 11 variables, but only 9 variables
were found in the data file"; or "used specified 150 input cases, but only 100 cases were found in the data file". Such problems can stop programs and affect the results of the
model fit. To avoid the problem, the number of variables and cases specified in the
program file was checked to assure they were consistent with the data file. 132
Third, any condition code that indicated problematic estimations for parameters was checked, (Bentler, 1995). When all estimations for parameters are technically accepted, no condition code should appear and the message "parameter estimations appear in order and no special problems were encountered during optimization" is given in the output file (Bentler). Problematic estimations can be avoided by applying appropriate start vales (Bentler, 1995), which is an im'tial point that allows program to begin parameter estimation (Byrne, 1995). This strategy was used for this study.
Fourth, the residual covariance matrix was examined to identify any problematic specifications involved in any equation test (parameter). The residual covariance matrix
reflects the discrepancy between the observed covariance matrix and the hypothesized
covariance matrix (Byrne, 1995). The standardized version was used to recognize the
degree of misfit associated to any two variables. For a good fit model, all standardized
residuals for each pair of variables should be small and evenly distributed. Ideally,
standardized residuals are expected to range from -.10 to .10 (Byrne, 1994, p. 53). When
larger residuals are associated with any particular parameters, there are problematic
specifications involving those parameters. Further examination of individual variables or
modification procedures might be made based on the results.
Fifth, the iterative summary was evaluated to assure the convergence of
estimations of parameters (Bentler, 1995). The iterative summary indicates the degree to
which parameter estimation is modified as the program attempts to estimate the optimal fit
between data and model. IdeaEy, a low numbers of iteration is expected. Possible
problems associated with a large number of iterations could be theoretical 133
(underidentified) or practical (start values) (Bentler). On the iterative summary, the function values should decrease as the number of iterative increase. Generally, the convergence criterion defaults at maximum of 30 if the programmer does not change the default set up by the program. When iterations are over 30, a warning message about the meaning of any result is suspected due to large numbers of iterations occurs. Simflar to the problem of condition code, the problem of large numbers of iterations can be fixed by applying an appropriate start value (Bentler, 1995). This strategy was used for this study.
Sixth, goodness-of-fit indices were reviewed to evaluate model adequacy. The goodness of fit indices including Chi-square [x" (df)], relative Chi-squre (x' / df ratio),
Normed Fit Index (NFI), Bentler-Bonett Nonnormed Fit Index (NNFI) and Comparative
Fit Index (CFI) were used. The statistical significance for rejecting the Chi-square test was set at p< .05. A. value less than 3 for x"! df ratio < 3 and a value of .90 for NFI,
NNFI and CFI (Byrne, 1995) were used in this study.
Seventh, evidence of misfit was examined by evaluating the Wald and Lagrange
Multiplier tests to determine revisions for model fit improvement. The Wald test mdicates misfitting parameters that could be dropped firom the model. The Lagrange
Multiplier test detects potential parameters that could be added to improve model fit
(Bentler, 1993). Revisions for any parameter or variable are determined based on the prior theoretical knowledge and statistical judgment (Bentler, 1995).
Eighth, modification of the hypothesized model was done based on the evaluation of the above procedures and substantive theoretical knowledge. Normally, one parameter 134 was changed at a time to determine the eflfect on model fit. The changes in CFI and x I were used to determine whether the model fit improved (Byrne, 1995).
Nineth, standardized coeffidents including direct and indirect effects were examined to assess the direction and magnitude of parameters within the structural model for interpretation of the model (Byrne, 1995).
Summary
The design, methods, and procedures for data collection were described. The essential components, overall purposes, and specific procedures for validation of this cross-cultural study were provided and discussed. The overall plan for data analysis was outlined. 135
CHAPTER4
DESCRIPTION OF SAMPLE AND RESULTS
For this chapter, the description of the sample profile, and survey results related to caregivers* feelings about role requirements are presented. Psychometric properties of measures are addressed. A description of study variables including correlations and descriptive statistics is presented. Results of the model testing are reported.
Description of the Sample Profile
For this study, a total of 210 survey packages were mailed to potential subjects identified by the investigator and by the recruiting representatives; 30 were identified by the investigator and 180 were identified by the recruiting representatives. There were seven cities including Atlanta, Boston, Los Angeles, San Francisco, Washington District of Colombia, Phoenix and Tucson involved as study sites. Of the seven study sites, the investigator was responsible for recruiting from Tucson. The six recruiting representatives recruited in the cities where they lived. Finally, one hundred and forty-four participants returned the surveys. This is a 69% response rate.
En this study, there were very minor random missing data for all subjects who returned packages, except seven participants who consistently had missing data on one or another instrument. These seven packages were unusable due to incomplete data for the model variables. Therefore, 137 subjects were used to describe the demographic characteristics and to test the model. There was less than .05% missing data for all responses. With a small amount of missing data,, the mean, replacement procedure was used, for the scale scores (Bollen, 1989). 136
Characteristics of Caregivers
The caregivers' ages ranged from 21 to 70 with a mean of 41 (SD=10.6)- The caregivers had been living in the United States from 2 years to 58 years with an average of
19 years. The education level of the sample ranged from 9 to 22 with a mean of 16.1
(SD=2.7) (Table 2). There were 96 (70.1%) female and 41 (29.9%) male caregivers.
The sample was predominantly forieng bom immigrants (N=127,89.2%). Only only 14 (10.2%) were bom in the United States. Of the 127 immigrants, 55 (40.1%) were
from Mainland China and 44 (32.1%) were from Taiwan. Eighteen (13.1%) were from
Hong Kong. Three (2.2%) were from Indonesia. Two (1.5%) were from Vietnam and only one (0.7%) was from the Philippines (Table 3). Half of the participants (N=69,
50.4%) identified themselves as Christian, 23 (16.8%) were Buddhist, 6 (4.4%) were
Catholic, 2 (2.2%) were Taoist and 35 (25.5%) indicated no religious preference (Table
3). The majority of participants (N=116,85.4%) were in the middle to upper income
bracket (from $40,0(X) to $149,999 per year) and only 15 (10.9%) reported an annual
income below $ 39,000 (Table 3). Most caregivers (N=99,67.9%) were fiiU-time
employees. Twelve (8.8%) were part time employees and 32 (23.4%) were homemakers
(Table 4). i?
137 Table 2.
Age. Length of Time in the U.S. A. and Educational Level of the Caregivers (N=137)
M SD Range
Age 41 10.60 21 -70 Length of time in the U. S. A 19 12.00 2 -58 Level of Education 16.1 2.7 9 -22
Table 3-
Characteristics of the Caregivers (N=I37)
Description N Valid %
Gender Feniale 96 70.1 Male 41 29.9 Missing 0 0 Total 137 100
Birth place Mainland China 55 40.1 Taiwan 44 32.1 Hong Kong 18 13.1 United States 14 10.2 Indonesia 3 2.2 Philippine 2 1.5 Vietnam 1 0.7 Total 137 100
Religion beliefs Christian 69 50.4 Buddhist 36 26.3 Catholic 6 4.4 Taoist 2 1.5 None 23 16.8 Missing 1 .7 Total 137 100 138 Table 4.
House Income and Types of Occupation of the Caregivers (N=137)
Description N Valid %
House income Less than $10,000 1 .7 $20,000-$29,000 3 2.2 $30,000-$39,000 LI 8 $40,000-$49,999 20 14.6 $50,000-$59,999 15 10.9 $60,000-$69,999 19 13.9 $70,000-$79,999 18 13.1 $80,000-$89,999 5 3.6 $90,000-$99,999 12 8.8 $100,000-$124,999 14 10.2 $125,000-$144,999 3 2.2 $150,000-above 10 7.3 Missing 6 4.4 Total 137 100
Type of job Full time 93 67.9 Part time 12 8.8 Housekeepers 32 23.4 Missing 0 0 Total 137 100 139
Among the 137 participants, 40 (29.2%) were recruited from San Francisco; 28
(2L4%) were recruited from Tucson from Atlanta, respectively. Nineteen (13.9%) were recruited, from Boston, 11 (8%) were cruited from Los Angeles; 6 (4.4%) were recruited
from Washington D. C., and 5 (3.7%)were recruited from Phoenix (Table 5).
Characteristics of Care Recipients
The age of care recipients ranged from 55 to 95 years old with a mean age of
74.37 (SD=9.05) with 75 (54.7%) females and 62 (45.3%) nudes. The majority (N=83,
60.6%) was alert; 34 (24.8%) were forgetfiil; 14 (10.2%) were disoriented; and 6 (4.4%)
were confused. In terms of mobility, 59 (43.1) were ambulatory; 25 (18.2%) used a cane or a walker; 30 (21.9%) needed assistance from others; and 21 (15.4%) were bed bound
(Table 6). Table 7 Usts the number and percentage of health problems among the care-
recipients. There were 119 (86.9%) caregivers who reported that their care redpients to
have one or more health problems and 15 (10.9%) reported that their care recipients had
no health problem. Ten health problems were identified (Table 8). Of the ten health
problems, hypertension and/or heart diseases (N=66,49.3%) were the most common
problems, followed by diabetes (N=34,25.4%). Other problems included stroke (N=33,
24.6%), arthritis (N=32,23.9%), demenua (N=30,22.4%), cancer (N=30,22.4%),
chronic obstruction pulmonary diseases (COPD) (N=20,15.7%), renal failure (N=14,
10.4%), hip fracture (N=10,15%), and Parkinson's disease (N=6,4J%). Table 5.
Geographic Location of the Caregivers (N=137)
CiUes N VaUd %
San Francisco 40 29.2 Tucson 28 20.4 Atlanta 28 20.4 Boston 19 13.9 Los Angeles 11 8.0 Washington DC 6 4.4 Phoenix 5 3.7 Total 137 100.00 14L
Table 6.
Characteristics of the Care Recipients (N = 137)
N VaUd%
Gender Female 75 54.7 Male 62 45.3 Missing 0 0 Total 137 100
Conscious status Alert 83 60.6 Forgetful 34 24.8 Disorientation 14 10.2 Confuse 6 4.4 Missing 0 0 Total 137 100
Mobility Ambulatory 59 43.1 Using cane or walker 25 18.2 Need assistance from others 30 21.9 Bed bound 21 15.3 Missing 0 0 Total 137 100 Table?.
Summary of Health Prnhlems among the Care Recipients (N=137)
# of Health Problems N %
L 26 19.0 2 26 19.0 3 30 21.9 4 18 13.1 5 12 8.8 6 6 4.4 7 1 .7 None 15 10.9 Missing 3 2.2
Total 137 100.0 143
Table 8.
Types of Health Problems among Care Recipients (N=L34)
Description Yes Valid % No Missing Total
Hypertension/Heart disease 66 (49.3%) 68 3 137 Diabetes 34 (25.4%) 100 3 137 Stroke 33 (24.6%) 101 3 137 Arthritis 32 (23.9%) 102 3 137 Dementia 30 (22.4%) 104 3 137 Cancer 30 (22.4%) 104 3 137 COPD 21 (15.7%) 113 3 137 Renal failure 14 (10.4%) 120 3 137 Hip fracture 10 ( 7.5%) 124 3 137 Parkinson 6 ( 4.5%) 134 3 137 144
Kinship Between the Caregivers and the Care Receipients
Table 9 presents types of kinship relationships between the caregivers and the care receivers with the number and valid percentage. There were 62 (45.3%) daughter caregivers. Of them, 35 (25,6%) cared for mothers and 27 (9.7%) took care of fathers.
There was a similar number of sons (N=35,25.6%) and daughters-in-law (N=34,24.8%) providing care. Of the 35 sons, 22 (16.1%) cared for fathers and 13 (9.5%) took care of mothers. Among the 34 (24.8%) daughters-in-law, 23 (16,8%) were taking care of mothers-in-law and 11 (8.0%) cared for fathers-in-law. There were six (4.4%) sons-in- law in this study; three (2.2%) provided care to fathers-in-law and the other three cared for mothers-in-law.
Caregiving Involvement
AH caregivers provided assistance with at least three activities of daily living
(ADL). Some provided assistance with up to 16 activities. The mean number of activities caregivers were involved in was 9.6 (SD= 3.5). Table 10 summarizes types of assistance,
which caregivers were involved. Among the 16 activities, transportation (N=134,97.8%)
was the most frequent assistance provided by the caregivers; then, shopping (N=130,
94.9%), and handling mail & bills (N=126,92%). More than half of the caregivers
provided help with cleaning (N=106,77.4%), cooking (N=104,75%), admuustration of
medications (N=99,72.3%), answering the phone (N=98,71.5%), Bnancial support
(N=96,70.1%), and dressing (N=85,62%). Other activities were laundry (N= 67,
48.9%), bathing (N=61,4t.5%), toileting (N=52,38.0%), walking (N=51,37.2%), 145 moving/ changing position (N=36,26J%), feeding (N=37,27.1%), and care after soiling
(N=36,26.3%).
Summary
The sample was predominately were foreign-bom immigrants. Women, either daughters or daughters-in-law were the primary caregivers of relatives. Most participants were Christian, well-educated, full time employees and in the middle to upper income.
The majority of care redpients was also females and had an average age of 74.4. Most of them were alert with one or more than one chronic illnesses.
The average number of activities of daily living (ADL) caregivers provided was
9.6. More than half of the caregivers provided the following types of assistance (from the most frequent): (I) transportation, (2) shopping, (3) handling mails & bills, (4) cleaning, cooking, (5) administration of medications, (6) answering the phone, (7) finandal support and (8) dressing. 146
Table 9.
Kinship between the Caregivers and the Care Recipients (N=137)
Description Yes Valid %
Daughters/mothers 35 25.6 Daughters/fathers 27 19J Sons/fathers 22 16.1 Sons/mothers 13 9.5 Daughters-in-law/mothers-in-law 23 16.8 Daughters-in-law/fathers-in-law 11 8 Son-in-Iaw/fathers-in-Iaw 3 2.2 Son-in-law/mothers-in-law 3 2.2 Total 137 100 L47
Table 10.
Types of Assistance Provided by the Caregivers (N=137)
Description Yes valid % No Missing Total
Transportation 134 (97.8%) 3 0 137 Shopping 130 (94.9%) 7 0 137 Handling mails & bills 126 (92.0%) 11 0 137 Room clean 106 (77.5%) 21 0 137 Cooking 104 (75.9%) 31 0 137 Medication 99 (72.3%) 38 0 137 Answering phone 98 (71.5%) 39 0 137 Fmanciai support 96 (70.1%) 41 0 137 Dressing 86 (69.0%) 51 0 137 Laundry 67 (48.9%) 70 0 137 Bathing 61 (44.5%) 76 0 137 Toileting 52 (38.0%) 85 0 137 Walking 51 (37.2%) 86 0 137 Feeding 37 (27.2%) 100 0 137 Moving/ Change position 36 (26.5%) 100 I 137 Care after Soiling 36 (26.3%) 101 0 137 148
Survey Results Related to Caregivers' Feelings about Role Requirements
Tn this study, seven items from the Poulshock and Demiling's (1984) burden inventory (BI) were used to measure the concept of subjective feelings about role requirements. The seven items in the BI are (I) bathing, (2) dressing, (3) toileting, (4) moving, (5) care after soiling, (6) feeding, and (7) grooming. Descriptive statistics, instead of psychometrics, are described in this section. Table LI summarizes the descriptive statistics including mean, standard deviation and range of scores from the highest to lowest.
The seven item means from highest to lowest were bathing (M= .87, SD= 1.09), dressing (M= .65, SD= .81), grooming (M= .63, SD= .70), toileting (M= .62, SD= .99),
and moving (M= .61, SD= .88). Both item 5, care for wetting and soiling and item 6,
feeding had the same mean but there was a large difference in the standard deviations and
ranges of scores between two items. Item 5, care for wetting and soiling had a mean of
.52 (SD= .97) with range from 0 to 3 and item 6, feeding had a mean of .52 (SD= 128)
with scores ranging from 0 to 9.
In table 12, data associated with each specific item (task) are presented. The
numbers and percentages of caregivers involved and not involved are described. The
numbers and percentages of caregivers' feeling about involvement are also addressed. Table iL
Descriptive Statistics of Subiective Feelings about Role Requirements (^=1371
Item No. Description M SD Range
Iteml Bathing .87 1.09 0-6 Iteni2 Dressing .65 .81 0-4 Item7 Grooming .63 .70 0-3 Item3 Toileting .62 .99 0-5 Item4 Moving .61 .88 0-3 ItemS Care for wetting & soiling .52 1.28 0-3 Item6 Feeding .52 .97 0-9 150
Table 12.
Frequencies of Subjective Feelings about Role Requtrements (N=L37')
Description N Valid %
Groonaing Involved 71 51.8 Not involved 66 48.2 137 100.0 Involved (N=7I) No problem 57 80.3 Tired It 15.5 Difficulty 2 2.8 Missing I L4
Total 71 100.0
Bathing Involved 66 48.2 Not involved 71 51.8 137 100.0 Involved (N=66) No problem 26 39.4 No problem and/or Tired I 1.5 Tired 31 47.0 Difficulty 7 10.6 Angry 1 1.5
Total 66 100
Dressing Involved 65 47.4 Not involved 72 52.6 137 100.0 Involved (N=65) No problem 45 69.2 Tired 17 26.2 DifBcuIty 2 3.1 Angry t L.5
Total 65 100.0 151
Table 12. (Continued)
Description N Valid %
Toileting Involved 52 38.0 Not involved 85 62.0 137 100. Involved (N=52) No problem 26 50.0 Tired 21 40.4 Difficulty 5 9.6
Total 52 100.0
Moving/Position change Involved 49 35.8 Not involved 88 64.2 137 100.0 Involved (N=49) No problem 23 47.0 No problem and/or Tired 1 2.0 Tired 19 38.8 Difficulty 5 10.2 Angry 1 2.0
Total 49 100.0
Feeding Involved 47 34.3 Not involved W 65J 137 100.0 Involved (N=47) No problem 33 67.4 No problem and/or Tired 1 2.0 Tired 10 20.4 Difficulty 2 4.2 Angry I 2.0
47 100.0 152
Table 12. (Continued)
Description N Valid %
Care after Soiling Involved 38 27J Not involved 99 72.3 137 100.0 Involved (N=38) No problem 20 52.6 No problem and/or Tired 1 2.6 Tired 10 26.4 Difficulty 6 15.8 Angry 1 2.6
38 lOO.O 153
Among the 137 participants, 71 (51.8%) were involved in grooming care and the remaining 66 (48.2%) were not. Of the 71 caregivers, who involved in groonung care, the majority (N=57,80.3%) claimed no problem at all. About 11 (15.5%) reported it was tiring and two (2.8%) confirmed it was difficult. In contrast, there were 66 (48.2%) participants helping with bathing and 71 (51.8%) who did not. Of the 66 involved caregivers, 26 (39.4%) claimed no problem at all. One (1.5%) indicated it was sometimes no problem and sometimes tiring. Thirty-one (47%) found it tiring. Seven (10.6%) experienced difficulty with bathing. One (1.5%) reported assistance made her/him upset.
For assistance with dressing, 65 (47.4%) were involved and 72 (52.6%) were not.
For the 65 involved caregivers, 45 (69.2%) indicated no problem at all. Seventeen (26.2
%) found it was tiring, two (3.1%) indicated it was difficult and one (1.5%) confessed it was upsetting. In terms of toileting care, about 52 (38%) caregivers were involved and 85
(64.2%) denied giving help. Among the 52 involved caregivers, 25 (50%) claimed it was no problem at all. Twenty (40.4%) complained it was tiring and five (9.6%) confirmed it was difficult.
There were 49 (35.8%) caregivers who provided assistance with moving and/or positioning change and 88 (64.2%) who did not. Among the 49 involved caregivers, 23
(47 %) denied any problem. One (2.0%) indicated it was sometimes no problem and sometimes tiring. Nineteen (38.8%) complamed it was tiring, five (10.2%) confirmed it was difficult and one (2.0%) confessed it was upsettmg. For assistance with feeding, 47
(34.3%) were involved, and. 90 (65.7%) were not. Of the 47 involved partidpants, 33
(67.4%) indicated it was no problem at all. One (2.0%) described sometimes no problem 154 and sometimes tiring. Ten (20.4%) claimed it was tiring, two (4.2%) claimed it was difficult and one (2.0%) confirmed it was upsetting.
For care after soiKng, only 38 (27.7%) were involved but the nujority (N=99,
72.3%) were not. Among the 38 who provided care after soiling, 20 (52.6%) claimed it was no problem at all. One (2.6%) indicated sometimes no problem and sometimes tiring.
Ten (26.3%) described it was tiring, six (15.8%) reported it was difficult and one (2.6%) reported it was upsetting.
Summary
Among the seven tasks, while grooming care was the most firequent assistance provided by the adult Chinese caregivers, care after soiling was the least firequent activity the participants performed. In general, most caregivers felt their assistance was no problem at all, except for bathing, and moving or position changes, tofleting, and care after soiling.
Psvchometric Properties of Measures
To assure the reliabihty and validity of measures for model variables, the psychometrics of the scales used in this study were evaluated and are described in this section. The reliabib'ty was evaluated by assessing internal consistency including the standardized Cronbach's alpha coefficients, the relationship between the item means and standard deviations, the inter-item correlations and the item to total correlations.
The criterion for the standardized coefficient a is .70 for a new psychosodal scale and .80 for a mature scale (Edwards, 1970; Nunnally, 1978). The standard deviatioa for each item was expected to be less than half the item mean (Nunnally). The mter-item 155 correlations were expected to be m the range from .30 to .70 for more than half of the items (Gordon, 1968; Kerlinger, 1973). The minimum criterion for the item-total correlation was .40 or over (Anastasi, 1976; Gordon, 1968).
Construct validity was assessed by factor analysis using principal component extraction and oblique rotation techniques. Nunnally (1978) suggests the criterion for factor loadings is greater than .30 for acceptable. Kerlinger (1973) recommends that the factor loadings need to be .40 or over. The criterion of greater than JO was applied because scales used in this study were either newly translated or revised. Table 13
presents the reliability each scale. Table 14 summarizes validity with the range of factor
loadings, eigenvalues and percentage of explained variance for each scale.
The Revised Yang's Hlial Vales Subscale fYFVS-R^
The YFVS-R was used to measure the concept of filial values. The average item
mean of the YFVS-R was 2.22 (SD= .74) with scores ranging from 1.68 to 2.45, on a 4-
point scale from 0 to 3. The Cronbach's alpha coefficients of the YFVS-R reached the
criterion of a good reliable scale with a standardized alpha of .89. All item standard
deviations were less than half their item means, except item 8, "I feel that I should say
something that may save my parent's face as needed". The inter-item correlations also met
the criterion with a range of .31 to .74. The item-total correlations ranged from .58 to
.72, which exceeded the minimum criterion of .40. Table 13.
Summary of Reliabilitv Assessment for All Measures
Description of Scale Item (M> SD Range Reliability Number of Subjects Filial Value Subscale (FVS) 2,22 .74 1.67- 2.45 a = .89 137 Filial Affection Subscale (FAS) 2.46 .62 2.12- 2.78 a = .81 137
Personal & Social Cost Subscale (PSCS) 2.32 .73 2,01 - 2.52 a = .83 137 Physical Emotional Cost Subscale (PECS) 2,20 .74 1.84- 2,40 a = .81 137 Economic Cost Subscale (ECS) 2.00 .60 1,38- 2,36 a = .70 137
Picot's Caregiver Rewards Scale (PCRS) 1.59 1,22 .26- 2,53 a = .89 137
Behavioral-Oriented Subscale (BOS) 1.48 .82 .77- 1.79 .85 137 Cognitive-Oriented Subscale (COS) 1.48 .86 .82- 1.85 a = .82 137
Physical Component Subscale (PCS) 3.00 .60 1.64- 4.83 a = .89 137 Mental Component Subscale (MCS) 3.64 .84 1.77- 5.31 a = .86 137
Note. Cronbach's Alpha "a" used for reliability Table 14.
Summary of The Range of Factor Loadings for All Measures ('N=137'>
Description of Scale Subscale Range of Factor Igenvalue Variance % Loadings Yang's Filial Obligation Scale Filial Value Subscale .67- .81 4.48 56.04 Affection Subscale .72 - .87 3.35 41.91
Cost of Care Inde^ Personal & Social Cost Subscale .77 - .88 2.67 66.85 Physical Emotional Cost Subscale .76 - .83 2.56 64.01 Economic Cost Subscale .75-.81 2.48 62.06
Picot's Reward Scale Overall Scale .16-.78 7,30 29.18
Ways of Coping Scale Behavioral-Oriented Subscale .42 - .65 5.08 28.25 Cognitive-Oriented Subscale .18- .67 4.70 26.10
SF-36 Health Survey Scale Physical Component Subscale .34 - .71 6.85 32.61 Mental Component Subscale .24 - .75 5.21 37.19 158
With principal components factor analysis, a scree plot proved one factor with an eigenvalue of 4.48, explaining 56.04% of the variance. All loadings were over the criterion of .30.
The Revised Yang's Filial Affection Subscale CYFAS-R)
The YFAS-R was used to test the concept of filial affection. The average item mean was 2.46 (SD= ,62) with scores ranging from 2.12 to 2.78, on a 4-point scale from 0 to 3. The Cronfaach's alpha coefffdents of the YFAS-R also reached the criterion of a
mature scale with a standardized alpha of .81. The criterion of the standard deviation for
each item expected to be less than half of the item means was met for all items. More than
half of item's inter-item correlations ranged within the criterion of .30 to .70. The
mim'mum criterion of .40 for item-total correlations was satisfied with a range of .43 to .66
was obtained.
The principal components factor analysis showed one factor on the scree plot with
an eigenvalue of 3.35, accounting for4l.9l % of the variance. All factor loadings were over the criterion of .30.
Personal-Social Cost Subscale (TSCSl
The concept of personal-social cost was measured by the PSCS. The average item
mean was 2.32 (SD= .73) with scores ranging from 2.01 to 2.52 on a 4-point scale from 1
to 4. The internal consistency coefficients also met the criterion for a mature scale with a
standardized alpha of .83. Other criteria for reliability were met. The standard deviations
for each item were less than half the item means. The inter-itemcorrelations ranged from
.49 to .70. Item-total correlations ranged from .61 to .76. 159
The principal components factor analysis revealed one factor with an. eigenvalue of
2.67, explainmg for 66.85 % of the variance. The criterion of factor loadings .30 was met with a range of .77 to .88.
Phvsical-Emotional Cost Subscale (PECS)
The PECS was used to measure the concept of physical and emotional burden.
The PECS had an average item mean of 2.20 (SD= .74) with scores ranging from 1.84 to
2.40 on a 4-point scale from I to 4. The internal consistency coefficient achieved the criterion for a mature scale with a standardized alpha of 8 L The standard deviations were less than half of the item means for all items. The criterion of more than half of items' inter-item correlation ranging between .30 to .70 was satisfied. Item-total correlations ranged from .58 to .68, which satisfied the criterion of .40.
The principal components factor analysis confirmed one factor with an eigenvalue of 2.56, explaining 64.01 % of the variance. The criterion of factor loadings over JO was met with a range of .76 to .83.
Economic Cost Subscale (ECS)
The ECS was used to test the concept of economic burden. The average item mean was 2.00 (SD=^ .60) with scores ranging from 1.83 to 2.36 on a 4-point scale from 1 to 4. The ECS had an internal consistency alpha of .70 that met the minimum criterion of acceptability. The standard deviations for each item were less than half of the mean. The inter-item correlations ranged from .40 to .62. The item-total correlations ranged from
.61 to .76. 160
The principal components factor analysis revealed one factor with an eigenvalue of
2.48,^ accounting for 62.06 % of the variance. The criterion of factor loadings over .30 was met with a range of .75 to .81.
The Picot's Caregiver Rewards Scale fPCRSI
The PCRS was used to test the concept of perceived, reward. The PCRS had an average item mean of 1.59 (SD= 1.22) with scores ranging from .26 to 2.53 on a 5-point scale from 0 to 4. The internal consistency coefficient reached the criterion for a mature scale with a standardized alpha of .89. However, only one item met the criterion of item standard deviation of less than half the item mean. This item was item 11, "I have a closer
relationship with my relative". The criterion of more than half inter-item correlations
within a range of .30 to .70 was not met. All items met the mmimum criterion of .40 for
item-total correlations, except for the following five items. Item 5, 'T feel that my relative
will remember me in his/her will for my care". Item 7, "I feel nurses, doctors and social
workers work harder to care for my relative too". Item 8, 'T feel that placing my relative in
a nursing home will be avoided". Item 19, 'T feel I do not need to hold a job". Item 24, "I
feel having my relative live with me means added money coming into the house".
A principal components factor analysis with a one-factor solution was performed.
The scree plot showed one factor with an eigenvalue of 7.30, accounting for 29.18% of
the variance. All factor loadings were over the criterion of .30 except three items. These
three items were item 5, 'T feel that my relative will remember me in his/her will for my
care"; item 19, 'T feel I do not need to hold a Job"; and item 24," I feel having my relative
live with me means added money coming mto house". 161
Behavioral-Oriented Subscale (BOS')
For this study, to reduce the length and ensure that items were culturally relevant to Chinese caregivers, original items with factor loadings greater than .35 were retained.
The meaning of each item was also examined to ensure items were culturally applicable to
Chinese caregivers. Eighteen items were selected for the Behavioral-oriented Subscale
(BOS).
The item in the BOS had a mean of L48 (SD= .82) ranging from .77 to 1.79 on a
4-point scale from 0 to 3. The internal consistency coefficient was fairly good with a standardized alpha of .85. The criterion of standard deviation less than half of the item mean was not met. The criterion for inter-item correlation within .30 to .70 for more than half items was not met. All items met the criterion of .40 for item-total correlation, except for four items. These tour items are as follows. Item 34," I take a big chance or do something very risky thing to solve the problem". Item 39," I change something so things
would turn out all right". Item 45, "I talk to someone about how I am feeling". Item 48,"
I draw on my past experience; I have been in a similar situation before".
A principal components factor analysis with a one-factor solution was performed.
A scree plot revealed one factor with an eigenvalue of 5.08, explaining 28.25% of the
variance. All factor loadings exceeded the criterion of .30 with a range of .42 to .65.
Cogmtive-Oriented Subscale (COS)
The COS was also derived from the revised 66-item Ways of coping scale (WOC-
R) (Folkman. & Lazarus, 1988) and used to measure the concept of usmg cognitive-
oriented coping strategies. Eighteen items were selected for the COS by two-factor 162 limited extraction with factor loadings greater than 30. As mentioned above, the meaning of each item was also examined to ensure that selected items were theoretically relevant to the concept of behavioral cognitive coping strategies.
The item means were 1.48 (SD= .86) ranging from .82 to 1.85 on a 4-point scale of 0 to 4. The internal consistency coefficient reached the criterion for a mature scale with a standardized alpha of .83. However, only four items met the criterion of standard deviations less than half of the item means. The criterion of more than half of items' inter- item correlation ranging between .30 to .70 was not satisfied. Only eleven items met the minimum criterion of .40 for item-total correlation.
A principal components factor analysis with one-factor solution was performed.
The scree plot showed one factor with an eigenvalue of 4.70, accounting for 26.10 % of the variance. All factor loadings met the criterion of .30, except for item 16," I sleep more than usual".
Phvsical Component Summary Subscale (PCSS)
The PCSS is one of two subscales from the 36-item Short-Form Health Survey
(SF-36) (Ware et al., 1993). It was used to measure the concept of the physical health.
The PCSS had a mean of 3.00 (SD= .60) ranging from 1.64 to 4.83 on a scale of 1 to 5.
The internal consistency coefficients for the PCSS reached the criterion for a good reliable scale with a standardized alpha of .89. All items' standard deviations were less than half the item mean. However, the criterion of inter-item correlations within JO to .70 for more than half items was not met. The criterion of .40 for item-total correlations for all items was also not satisfied. L63
A principal components factor analysis with a one-factor solution was performed.
The scree plot showed one factor with an eigenvalue of 6.85, accounting for 32.6 L% of the variance. All factor loadings met the criterion of .30 with a range of .34 to .68.
Mental Component Summary Subscale fMCSS)
The MCSS is the other subscale from the 36-item Short-Form Health Survey (SF-
36) (Ware et al., 1993). It was used to measure the concept of mental health. The item mean was 3.64 (SD= .84) ranging from 1.77 to 5.31 on a scare range of 1 to 6. The internal consistency coefficient also reached the criterion for a good reliable scale with a standardized alpha of .86. All items' standard deviations were less than half the item means. The criterion of half inter-item correlation of .30 to .70 was not satisfied. All items exceeded the criterion of .40 for item-total correlations, except for item 5_a and item 5_c. Item 5_a," Have you cut down the amount of time you spend on work or other activities as a result of any emotional problem?" Item 5_c, "Have you had the problem of doing work or other activities as carefully as usual as a result of any emotional problem?"
A principal components factor analysis with one-factor solution was performed.
Scree plot showed one factor with an eigenvalue of 5.21, accounting for 37.19% of the variance. AH factor loadings met the criterion of .30 except item 5_c "Have you had the problem of didn't do work or other activities as carefiiUy as usual as a result of any emotional problem?).
Summary
In general, all scales were identified as psychometricaUy sound measures for model testing. All scales achieved the criterion of good reliability for a mature scale with a 164 standardized alpha value above .80, except for the economic cost subscale, which had a standardized alpha of .70. Most factor loadings exceeded the criteria of .30; exceptions include some items from the perceived, reward scale, the cogm'tive-oriented coping scale and the mental component subscale.
Description of Studv Variables
According to Cohen (1988), correlation estimation is helpful in identifying meaningful variables for model testing. Minimum correlation coefficients of .20 between two variables are expected for model testing (Cohen). In the following section, the correlation estimations are described. A summary of the correlation estimations for each study variable is presented in table 15. Table 15.
Zero-Order Correlations of Study Variables (N= 137)
Indicators/Variables VI V2 V3 V4 V5 V6 V7 V8 V9 VIO Vll VI. Filial Values - .53** ,00 -.22* -.30** -,28** .28** .03 .09 ,19* .33** V2. Filial Affection " .15 -.11 -.18* -,21* ,21* .12 ,16 .09 .27** V3, Feelings about Role Requirements — ,29** .31** ,18* .19* ,15 ,11 -.19* -,19* V4, Personal-social Cost — 87** ,76** -,12 .18* .08 -.28** -38** V5, Physical-emotional Cost ,69** -.12 .14 ,06 -.32** -.41** V6. Economic Cost ~ -.05 .15 ,08 -.29** -.32** V7. Perceived Rewards ,34** .03 ,06 V8. Behavioral-oriented coping — ,80** -.08 -.11 V9. Cognitive-oriented coping — -.01 -.06
ViO, Physical Health - VI1, Mental Health
Note. •* Correlation is significant at the 0,05 level (2-tailed), ** Correlation is significant at the 0,01 level (2-tailed), 166
Correlation Rstimations
The correlation among variables ranged from r = .00 to .80. Filial values were significantly, positively related to the filial affection (r= .53, p < .01), perceived rewards
.28, p < .01), physical health (r= .19, p < .05) and mental health (r= .33, p < .01).
Filial values were significantly, negatively associated with the personal-social burden
(r= -.22, p < .05), physical-emotional burden (f= -.30, p < .01) and economic burden
(p=-.28, p < .01). FiUal affection was also significantly, positively correlated to the
perceived rewards (r= .21, p < .05) and mental health (p= .27, p < .01) and it was
significantly, negatively related to physical-emotional burden (r= -.18, p < .05) and
econonuc burden (r= -.21, p < .05).
Subjective feelings about role requirements was significantly, positively assodated
with personal-sodal burden (p= .29, p < .01), physical-emotional burden .31, p < .01),
economic burden (n=.18, p < .05) and perceived rewards (r= .19, p < .01). Subjective
feelings about role requirements was significantly, negatively correlated to physical health
(r= -.19, p < .05) and mental health -.19, p < .05).
Personal-social burden had a significant positive relationship with the physical-
emotional burden .87, p < .01), economic burden (r= .76, p < .01) and behavioral-
oriented coping (r= .18, p < .05). It had a significant, negative relationship with physical
health (n= -.28, p < .01) and mental health (r= -.38, p < .01). Physical-emotional burden
was significantly, positively correlated to economic burden .69, p < .01). In addition,
it was significantly, negatively related to physical health (r= -.32, p < .01) and mental 167 health (r= -41, p < .01). Economic burden had a significant, negative relationship with physical health -.29, p < .01) and mental health (i^ -.32, p < .01).
The perceived reward was significantly, positively correlated to behavioral- oriented coping (n= .34, p < .01) and cogm'tive-oriented coping (r= .44, p < .01). There was a significant, positive relationship between the behavioral-oriented coping and cognitive-oriented coping (p= .80, p < .01). There was also a significant, positive relationship between physical health and mental health (r= .49, p < .01).
Descriptive Statistics of Studv Variables
Descriptive statistics are another usefiil tool for evaluating appropriateness of variables for model testing (Muthen, 1993). Muthen indicates that when a variable had a skewness or kurtosis value above the range of -1 to 1, the results of model testing can be adversely influenced, especially the Chi-square goodness-of-fit test. In addition, fCIdebrand (1986) suggests when a variable has a skewness value above -.20 to .20, its distribution is severly skewed. Therefore, descriptive statistics of study variables were examined. The mean for the variable of fiUal values was 17.80 (SD=4.46) in a range of 7 to 24. The distribution of filial values was negatively skewed with a value of -.40. There was one peak of kurtosis -.79. The variable of fiUal affectioa had a mean of 19.66 (SD =
3.26) ranging firom 8 to 24. Its value of skewness was -.75. The value of kurtosis was
.45. Table 16 lists the descriptive statistics of study variables with means, standard deviations, possible range of scores, actual range of scores, numbers of peaks, skewness and kurtosis. Table 16.
Descriptive Statistics of Study Variables (N= 137)
Variables M Possible Actual #of Skewness Kurtosis Range Range peak VI, Filial Values 17.80 4,46 0-24 7-24 1 -.40 -,79 V2, Filial Affection 19.66 3.26 0-24 8-24 1 -.75 ,45 V3, Feelings about Role Requirements 1,13 1.85 0-21 0-7 1 1,64 1,82 V4. Personal-social Cost 9,23 2.39 4-24 4- 14 1 -.64 -,12 V5, Physical-emotional Cost 8,79 2.37 4-24 4- 14 1 -.53 .03 V6, Economic Cost 8.01 1.90 4-24 4- 13 1 .63 .48 V7. Perceived Rewards 39,82 15,91 0-92 4-92 1 -.11 -.45 V8, Behavioral-oriented coping 26,68 8.01 0-54 8-45 1 -.61 .39 V9. Cognitive-oriented coping 28.34 8.01 0-54 12-47 1 -.36 ,76 VIO, Physical Health 63,12 7.61 21 -75 41 -75 1 -.56 ,03 VII, Mental Health 51,01 7.34 14-70 30-70 1 -.14 -,15 169
For the variable of subjective feelings role requirements, the mean was 1.13 (SD=
1.85) ranging fromO to 7. It had a positive skew of 1.64. The value of kurtosis was 1.82, which reflects that its distribution was very flat. The mean for the variable of personal- social burden was 9.23 (SD= 2.39) ranging from 4 to 14. It had a single peak with a skew of -.64. The kurtosis was -.12, which indicates that it had a fairly normal distribution. A kurtosis of 0 indicates that a distribution has a normal curve (Hfldebrand, 1986). The mean for the variable of physical-emotional burden was 8.79 (SD= 2.37) ranging from 4 to
14. The skewness value was -.53. The kurtosis value was -.03, which is very close to 0 reflecting that it had an almost normal distribution. The variable of economic burden had a mean of 8.01 (SD= 1.90) ranging from4 to 13. It had value of -.63 for skewness and a value of .48 for kurtosis.
The variable of perceived rewards had a mean of 39.82 (SD= 15.91) ranging 4 from 92. It had a single peak with a skew of .11 and a kurtosis of .45. For the variable of behavioral-oriented coping, the mean was 26.68 (SD=8.01) ranging from 8 to 45. It had a negative skew of -.61 and a kurtosis of .39. The mean for the variable of cognitive oriented coping was 28.34 (SD=8.01) ranging from 12 to 47. It was negatively skewed with a value of -.36. Its kurtosis value was .76.
The variable of physical health had a mean of 63.12 (SD= 7.61) ranging from 41 to
75. It had a negative, skewed with value of -J^6. It had a kurtosis of .03, which reflects that it had an almost normal distribution. The mean for the variable of mental health was
51.01 (SD= 7.34) ranging from 30 to 70. It was slightly negative, skewed with a value of
-.14 and it had a kurtosis value of -.15. 170
Summary
Overall, all scales for study variable successfully met the criteria for the psychometric test. All scales had fairly to very good reliability and validity. The correlarion estimations revealed that each variable had a significant relationship with at least one or more variables in this study.
Two variables, filial values and the mental health, have significant relationships with 2/3 of the rest of the variables. There were significantly assodated relationships, ether positively or negatively, between mental health and all study variables except for perceived rewards, behavioral-oriented coping and cognitive-oriented coping.
The values of skewness for all variables were from -.11 to 1.64. About 4/5 of variables were severely skewed because their skewness values were out range of -.20 to
.20 suggested by Hildebrand (1986). The kurtosis values ranged from -.03 to 1.82, which indicate that some variables had almost normal, and some had rather flat distributions.
Since the variable of subjective feelings about rote requirements had a severe skewed value of 1.64 and a very flat kurtosis value of 1.82, it was anticipated to be a problematic variable for model testing according to the criterion suggested by Muthen (1993); when skewness or kurtosis value is out the range of-I to 1, the results of model testing can be adversely afiected, especially the Chi-square goodness-of-fit test.
Results of the Evaluation of Model Adequacy
Figure! (page 101) presents all direct and indirect causal effects proposed within the structural model m the original modeL The 11 measurement regression paths (F —»•
V; the factor loadings) and 11 error variances (Es) within measurement models are also 171 presented. In this section, the direct and indirect effects specified in hypotheses are described.
Direct Effects
HI. Caregivers' health is directly positively affected by subjective beliefs of traditional filial obligation, perceived rewards and use of coping, and directly negatively impacted by perceived burdens.
H2. Use of coping is directly positively influenced by perceived burdens.
ED. Perceived burdens are directly affected by subjective beliefs of traditional filial obligation.
H4. Perceived rewards are positively influenced by subjective beliefs of traditional fiUal obligation and negatively affected by subjective feelings about role requirements.
H5. Subjective feelings about role requirements is negatively influenced by subjective beliefs of traditional filial obligation
Indirect Effects
HI. Caregivers' health is indirectly affected by subjective beliefs of traditional fiUal obligation through the mediated effect of perceived burdens.
H2. Caregivers' health is indkectly affected by subjective beliefs of traditional filial obligation through the mediated effects of perceived burdens and use of coping.
H3. Caregivers' health is mdirectly affected by subjective beliefs of traditional filial obhgation through the mediated effect of perceived rewards. 172
H4. Caregivers' health is indirectly affected by subjective beliefs of traditional fiUal obligation through the mediated effects of perceived rewards and subjective feelings about role requnrements.
Results of the Original Model Testing
For the original model, the residual covariance matrix was used for assessing overall model fit because it reflects the discrepancy in fit between the model covariance and the data covariance. The average off-diagonal standardized residual value was .13, which suggests the model had a poor fit with the data. Ideally, a good 6t model should have small and evenly distributed residual covariances (Bentler, 1995; Byrne, 1995). In this study, there was 71.21% of residual covariance between - .10 to .10. About 10.61% of the residual variance fell between -.10 to -.40 and 18.19% fell between .10 to .40.
According Bentler (1995), a large residual between a particular pair of variables indicates that the hypothesized covariance does not provide a good explanation for the data for that particular pair of variables. A large residual adversely affects model fit
(Bentler). In the original model, the largest off-diagonal standardized residual values from the highest for the first 20 pa^ of variables (the computer printout only displays the furst
20 pairs by defaulted) were between .48 to .16. Among those problematic pairs of variables, V3 and VI1 were the most frequently occurring variables associated with other different variables, then, VIO followed. Since the targe covariance residual between variables is likely to adversely affect model fit, modifications of structural and/or correlates of those variables are expected. 173
In addition, the Chi-square for this model was [x' (41) = 168.88, p< .05]. The
Chi-square ratio (df)] was 4.1. The NFI, NNH and CFI were .75,72 and .80, respectively. All findings proved that the hypothesized model did not adequately explain the data. There was some degree of misfit in some way between the hypothesized model and the data, although it is close to the criterion of .90 or greater for an acceptable fit.
OveraU, the original model with 11 variables was determined to be a poor fit and did not explain the data well.
Therefore, model revisions were made based on an evaluation of model adequacy by applying statistical judgments and theoretical knowledge. Evidences for poor fit and the justification of modifications are illustrated in the next section.
Evaluation of Poor Fit and Justification of Modifications
From a statistical judgment perspective, the variable of subjective feeling about providing assistance (V3) was an unfavorable variable in the original model, according to the evidence found in the descriptive statistics of study variables mentioned earlier in this chapter and the above findings for the original model. First, V3 had a large value of skewness and kurtosis that tended to adversely aSect a good model fit (Muthen, 1993).
Second, V3 had an extremely small variance, as compared to other variables in the model, which indicates that the V3 was not a meaningfiil variable for the model testing. Third, large covariance residuals between V3 and other variables consistently occurred.
Besides, there were some problematic specifications for VIO and VLl in the original model. First, the upper limit of largest standardized residual was .48 for VIO and
Vll, which reflects model misfit associated with VIO and Vll. Second, the Lagrange 174
MultipUer suggested adding a parameter between F1 and F3. The Wald test indicated that three parameters could be removed without losing information. These three parameters were VIO, F4; Vll, F4; and V3, FL According to, Bentler (1995), the Wald test can be used to evaluate which problematic specification can be dropped without losing information. The Wald test is also use for parsimony considerations, in which the model has more explanatory power with fewer numbers of parameters (Byrne, 1995).
From a theoretical perspective, with parsimony considerations, V3 could be onnitted without handicapping the whole theory. In addition, two indicators for F4 (VIO and Vll) could be tested separately without losing information.
Therefore, two revised models were developed and tested. For the first revised model, V3 was eliminated. The parameter between VIO and F4 was keep but the parameter between VI1 and F4 was removed (Figure 4). For the second revised model, similarly, V3 was dropped. In contrast, the parameter between VIO and F4 was removed but the parameter between Vll and F4 was keep (Figure 5). The causal hypothesized correlations and pathways for both revised models were described as follows. Figure 4 The First Revised Model + Y
-P
E2—p. V2
E3 E7 Note. Fl; Subjective Beliefs of Traditional Filial Obligation VI: Filial Values V2: Filial Affection F2: Perceived Burdens V4: Personal-social Cost V5: Physical-einoiional Cost V6: Economic Cost F3: Use of Coping V8: Behavioral-oriented Coping V9: Cognitive-oriented Coping V7: Perceived Rewards VIO: Physical Health Omitted: Figure 5. The Second Revised Model + Y
E2 ^ V2
E3 E7 Note. Fi: Subjective Beliers of Traditional Filial Obligation Vl; Filial Values V2; Filial Affection F2: Perceived Burdens V4: Personal-social Cost V5: Physical-emotional Cost V6: Economic Cost F3: Use of Coping V8: Behavioral-oriented Coping V9: Cognitive-oriented Coping V7: Perceived Rewards VIO: Mental Health Omitted; Ill
Causal Hypothesized Effects for the First Revised Model
For the first revised model, the outcome variable was changed from health status
to perceived physical health. The direct and three indirect effects are re-specified in the
hypotheses as follows.
Direct Effects
HI. Caregivers' perceived physical health is directly positively affected by
subjective beliefs of traditional filial obligation, use of coping and perceived rewards, and
directly negatively impacted by perceived burdens.
H2. Use of coping is directly positively influenced by perceived burdens.
H3. Perceived burdens are dnrectly affected by subjective beliefs of traditional
filial obligation.
H4. Perceived rewards are positively influenced by subjective beliefs of
traditional fiHal obhgation.
Indirect Effects
HI. Caregivers' perceived physical health is indirectly affected by subjective
beUefs traditional filial obligation through the mediated effect of perceived burdens.
H2. Caregivers' perceived physical health is indirectly affected by subjective
beliefs of traditional fiUal obbgation through the mediated effects of perceived burdens and
use of coping.
H3. Caregivers' perceived physical health is mdirectly affected by subjective
behefs and. values of traditional filial obligation through the mediated effect of perceived
rewards. 178
Causal Hypothesized Paths for the Second Revised Model
Similarly, all hypothesized effects were the same as those specified in the first revised mode, except for the outcome variable was perceived mental health instead of perceived physical health. The direct and three indirect effects are re-specified in the hypotheses as follows.
Direct Effects
HI. Caregivers' perceived mental health is directly positively affected by subjective beliefs of traditional filial obligation, use of coping and perceived rewards, and directly negatively impacted by use of coping and perceived burdens.
H2. Use of coping is directly positively influenced by perceived burdens.
H3. Perceived burdens are directly affected by subjective beliefs of traditional filial obligation.
H4. Perceived rewards are positively influenced by subjective beliefs of traditional filial obligation.
Indirect Effects
HI. Caregivers' perceived mental health is indirectly affected by subjective beliefs of traditional filial obligation through the mediated effect of perceived burdens.
H2. Caregivers' perceived physical mental is indirectly affected by subjective beliefs of traditional filial obligation through the mediated effects of perceived burdens and use of coping.
ED. Caregivers' perceived mental health is indirectly affected by subjective beliefs of traditional filial obligation through the mediated effect of perceived rewards. 179
Results of Model Testing: for the First and Second Revised Models
Results indicate that the above changes did improve the fit of the models with the data. For the first revised model, the Chi-square was decreased to (jf 24 = 70.47, p< .05) and the CFI improved to .91. Thex'/df ratio deaeasedfrom4.1to 2.9 forthe first revised model. For the second revised model, the Chi-square was decreased to (x"24 =
84.89, p< .05) and the CFI improved to .90, which is minimumly acceptable. The x" ^ df ratio decreased fi:om4.1 to 3.5. Since the Chi-square test is very sensitive to sample size, a model might be interpreted as a good fit with data when the CFI equals .90 or greater and the X" / df ratio is less than 3, even if the Chi-square test is significant (p< .05)
(Munro, 1997). In addition, Bentler (1995) believes that the CFI is the most reliabily index for model test. In this study, the CFI was used as the final decision for good model
6t. Overall, findings suggest that the first revised model (CFI= .91 and the ^ / df ratio =
2.9) had a better fit with data than second model (CFI= .90 and and the x" / df ratio = 3.5).
Table 17 hsts the average off-diagonal absolute standardfeed residuals forthe original and two revised models. Table 18 summarizes the goodness-of-fit indices including Chi- square, X" ratio, NFI, NNFI, and CFI for the original and two revised, models. Table 17.
Summary of Average Off-diagonal Absolute Standardized Residual Covariance for Each Model
Average off-diagonal absolute standardized residual covariance Original Model with 11 variables (Figure 2) .13
The First Revised Model with 9 variables (Figure .08 6) Drop V3 and VII but keep V10
The Second Revised Model with 9 variables (Figure 7) Drop V3 and VIO but keep vl 1
Note. Fl: Subjective Beliefs of Traditional Filial Obligation VI: Filial Values V2: Filial Affection F2: Perceived Burdens V4: Personal-social Cost V5: Physical-emotional Cost V6: Economic Cost F3: Use of Coping V8: Behavioral-oriented Coping V9: Cognitive-oriented Coping F4: Caregivers'Health Status VIO: Physical Health VI1: Mental Health V3: Subjective Feelings about Role Requirements V7: Perceived Rewards Table 18,
Summary of Goodness-of-Fit Indices for the Original and Revised Models (N= 137)
X (dO X" /df Ratio NFI NNFI CFI
Original Model with 11 variables (Figure 3) 168.88(41) 4.1 ,75 .72 .80 p<,05
The First Revised Model with 9 variables 70.47 (24) 2.9 .88 .87 .91 (Figure 6) Drop V3 and V11 but keep V10 p<.05
The Second Revised Model with 9 variables 84.89 (24) 3.5 .86 .84 .90 (Figure 7) Drop V3 and VIO but keep VI1 p< .05
Note. Good fit: x" (^0 statistic significant at p< .05 x^/dfratio<3 NFI greater than ,90 NNFI greater than ,90 CFI greater than .90
Fl: Subjective Beliefs of Traditional Filial Obligation VI: Filial Values V2: Filial Affection F2: Perceived Burdens V4: Personal-social Cost V5: Physical-emotional Cost V6: Economic Cost F3: Use of Coping V8: Behavioral-oriented Coping V9: Cognitive-oriented Coping F4: Caregivers'Health Status VIO: Physical Health VI1: Mental Health V3; Subjective Feelings about Role Requirements V7: Perceived Rewards 182
Direct Standardized Effects for the First Revised Model
Rgure 6 is the complete first revisedd model with path coefficients. Table 19 presents the standardized direct, indirect and total effect values for the first revised model.
For the furst hypothesis, caregivers' physical health was directly positively influenced by beh'efs of traditional fiUal obligation (p= .08) and perceived rewards
(P= .27). In contrast, caregivers' physical health was directly negatively affected by perceived burden (P= - .03) and use of copmg (P= - .18). For the second hypothesis, caregivers' perceived burdens had a direct positive effect on use of coping (P= .24). For the third hypothesis, caregivers' subjective beliefs of traditional fiUal obligation had a direct positive effect on use of coping (P= -.30). For the fourth hypothesis, caregivers' subjective beliefs in traditional filial obligation had a direct positive effect on perceived rewards (p= .37)
Indirect Standardized Effects for the First Revised Model
For the first hypothesis, caregivers' subjective beliefs of traditional fiUal obligation had an indirect positive effect on physical health through the mediated effect of use of coping and perceived burdens. Findings confirmed that there was a positive indirect relationship between caregivers' subjective befefs of traditional filial obligation and physical health. Total indirect effects were (- .30 x .24 x - .18 = .01). Figure 6 The First Revised Model with Standardized Path Effects
.081
F2 -.025 ^2984 F3 F1
.853' V4 V5 V6 -.176^ E2—• j V2 V8 V9 .369' E4 E5 E6 E8 E9 .27:
ElO t E7
Note. Fl; Subjective Beliefs of Traditional Filial Obligation VI: Filial Values V2: Filial Affection F2; Perceived Burdens V4; Personal-social Cost V5; Physical-emotional Cost V6; Economic Cost F3: Use of Coping V8: Behavioral-oriented Coping V9: Cognitive-oriented Coping V7: Perceived Rewards VIO: Mental Health Table 19,
Standardized Direct. Indirect and Total Effects for the First Revised Model
Note. Fl; Subjective Beliefs of Traditional Filial Obligation VI; Filial Values V2: Filial Affection F2: Perceived Burdens V4: Personal-social Cost V5: Physical-emotional Cost V6: Economic Cost F3: Use of Coping V8: Behavioral-oriented Coping V9; Cognitive-oriented Coping V7*. Perceived Rewards VIO: Physical Health 185
For the second hypothesis, caregivers' beliefs in traditional Glial obligation had an indirect positive effect on physical health through the mediated effect of perceived burdens. There was also a positive indirect relationship between caregivers' subjective
beliefs of traditional filial obligation and physical health. The total indirect effects were
(-.30x-.03=.0L).
For third hypothesis, caregivers' subjective beliefs of traditional filial obligation, had an indirect positive effect on physical health through the mediated effect of perceived rewards. A positive indirect relationship between caregivers' subjective beUefs of
traditional filial obligation and physical health was found. The total indirect effects were
(.37 X .27 = .1).
Direct Standardized Effects for the Second Revised Model
Figure 7 is the complete second revised model with path coefficients. Table 20
summarizes the standardized, direct, indkect and total effect values for the first revised
model.
For the first hypothesis, caregivers' mental health was directly positively influenced
by subjective beliefs of traditional filial obligation (P= .07) and perceived rewards (P= .27).
In contrast, caregivers' mental health was directly negatively affected by perceived burdens
(P= - .03) and use of coping (P== - .21). For the second hypothesis, caregivers' perceived
burdens had a direct positive effect oa use of copmg (P= .24). Figure 7 Tiie Second Revised Model with Standardized Path EfTects
.073
F2 -.034
.791'i F3 F1 .642' .932J .857' •Mf V5 V6 E2 • \2 V4
.372' E4 E5 E6
.262
Ell
E7
Note. Fl: Subjective Beliefs of Traditional Filial Obligation VI: Filial Values V2: Filial Affection F2; Perceived Burdens V4.* Personal-social Cost V5: Physical-emotional Cost V6: Economic Cost F3; Use of Coping V8: Behavioral-oriented Coping V9: Cognitive-oriented Coping V7: Perceived Rewards VII: Mental Health Table 20.
Standardized Direct. Indirect and Total Effects for the Second Revised Model
'mmmmw VI ,795* .795*
V4 .963* -,290 -,290 ,963»
V6 ,781 • -,235 -.235 ,781*
V8 .932* .157 .217 .157 .217 .932»
vii .073 -.034 -.212* .262 .063 -.045 .073 -.034 -.212» .262 m F3 .24!• .241"
Note. Fl; Subjective Beliefs of Traditional Filial Obligation VI: Filial Values V2: Filial Affection F2; Perceived Burdens V4: Personal-social Cost V5: Physical-emotional Cost V6: Economic Cost V7: Perceived Rewards F3: Use of Coping V8: Behavioral-oriented Coping V9: Cognitive-oriented Coping V11: Mental Health L88
For the third hypothesis, caregivers' subjective beliefs of traditional fiUal obligation had a direct negative effect on perceived burdens (P= - .30). For the fourth hypothesis, caregivers' subjective beliefs of tradiaonal fflial obligation had a direct positive effect on perceived rewards (|3= .37).
Indirect Standardized Effects for the Second Revised Model
For the first hypothesis, caregivers' subjective beliefs of traditional filial obligation had an indirect positive effect on mental health through the mediated effect of use of coping and perceived burdens. Findings support that there was a positive indirect relationship between caregivers' beliefs in traditional filial obligation and physical health.
Total indirect effects were (- .30 x .24 x - .21 = .02).
For the second hypothesis, caregivers' subjective beliefs of traditional fiUal obligation had an indirect positive effect on mental health through the mediated effect of perceived burdens. There was also a positive indirect relationship between caregivers' subjective beliefs of traditional filial obligation and mental health. The total indirect effects were (- .30 x - .03 = .01).
For third hypothesis, caregivers' subjective beliefs of traditional filial obligation had an mdirect positive effect on mental health through the medicated effect of perceived rewards. A positive indirect relationship between caregivers' subjective beliefs of traditional filial obligation and mental health was found. The total indirect effects were
(.27x.27 = .l). 189
Summary
Findings for the original model proved that it was a poor fitting model, however, it
provided usefiil information for modification of model. Findings suggest that two revised
models were acceptable good fit models based on the most reliable CFI. However, the
hypothesized correlations and pathways were empirically supported for the first revised
model because the second revised model failed, to meet the criterion of less than 3 for the
X" /df ratio. 190
CHAPER5
DISCUSSIONS. IMPLICATIONS, LIMITATIONS,
AND RECOMMENDATIONS
In this chapter, a brief discussion of the sample proffle is presented. The results of model testing are interpreted and discussed, foQowed by a brief discussion of findings in relation to theoretical orientations. Implications for nursing are addressed. Limitations of the study and recommendations for fixture research are described. Rnally the strengths of the study and conclusions are presented.
Brief Discussion of Sample Profile
The main purpose of this study was to identify causal relationships among factors that contribute to caregivers' health status. Specifically the study examined the impact of subjective beliefs of traditional filial obligation on caregivers' health through the mediated effects of perceived burden, perceived reward and use of coping options in Chinese caregivers.
The majority of the participants (89.2%) were foreign-bom immigrants. This
profile is consistent with the report given by the LF-S. Census in 1990. The U.S. Census
reported that nearly 80% of adult Chinese immigrants were foreign-bom. Family cargiving research has shown the caregivers are predominately female, either daughters or
daughters-in-law (Chou, LaMontagne, & Hepworth, 1999; Ehrlick, 1992). As expected,
nearly seventy percent of participants in this study were female, either daughters or
daughters-in-law, which was consistent with the literature. The predominance of women
reflects the fact that family caregiving is an extension of a woman's role, that is rooted in 19 L
Chinese culture (Ho, 1996). Literature also indicates that most caregivers are middle aged
(Robinson, 1997; SchoSeld, Herrman, & Bloch, 1997) and work either full-time or part-
time (Dautzenberg, Diederiks, Philipsen, &Tan, 1999; Spitze, Logan, Joseph, & Lee,
1994). These were also consistent with this study. The mean age of caregivers was 41
years old and more then two thirds were either full-time (67.9%) or part-time (8.8%)
employees. In contrast to some other caregiving research, the caregivers in this study
were well educated and had middle to upper incomes. This is a typical characteristic for
Asian immigrants reported in the U.S. Census (1990).
In terms of care recipients, in this study, the average age of care recipients was 74
and most care recipients (54.7%) were female, which is congruent with Houde's (1998)
and Lee, Brennan and Daly's (2001) findings. In a study of 240 caregivers of older adults,
Lee, Brennan and Daly found that most care recipients were female and over 70 years old.
Houde (1998) indicates there are 6.6 million elders over the age of 65 with some kind of
chronic health problems receiving long-term care; most of this care is being provided by
family caregivers in the home. In the current study, more than eighty percent of care
redpients had at least one chronic illness.
Summary
The sample in this study was similar to those in other studies in the literature.
Overall, the sample profile in ths study was that the majority of the caregivers were mid
life, working daughters or daughters-6i-law who provided care to firail elders with some
type of chronic illnesses. Most caregivers were well educated and had a moderate to high
economic status. 192
Inteqjretation and Discussion of Findings of Model Testing
Based on the Chi-square test and fit indices, the original proposed model did not adequately fit the data. Although the original model did not prove to adequately explain the data, it did provide useful information for model revision. Two models were developed based on results firom the original model testing. One model had physical health as the outcome. In the other, the outcome variable was mental health. The first revised model had good fit as indicated by the CFI= .91 and /df ratio = 2.9. The second revised model had a minimumly acceptable good fit as indicated by the CF[= .90 and the
/df ratio = 3.5.
The overall findings suggested that data provided a better explanation for physical health than for mental health as indicated by fit indictes. However, the coefficients for each individual path effect in both revised models are the same (Rgure 6, page L88; Rgure
7, page 192). The results of models were technically the same but the fit indicies being abit different may be explained as follows. There is evidence that people, like Chinese, firom a collective oriented-cluture show a higher reluctance to self-disclose their emotional feelings (Chen, 1995) or mental conditions (Parish, 2(KX)) due to concerns about the consequences of their behaviors, as compared to people firom individualistic clutures, such as Americans (Triandis, Brislin & Hui, 1988). In addition, there is a tendency for Chmese to deny or neglect to recognize their mental problems due to the effect of social stigma
(Parish, 2000). It is likely there is a higher sodal desirability bias for mental health than physical health measurement among the Chinese caregivers. 193
Since the results of both, models are technically tha same, findings of both models are taken and interpreted as a whole in the following section. The term "physical and mental health" has been used to refer to caregivers' health status in interpretation of the findings in the following section. In both models, caregivers' subjective beliefs of traditional filial obligation was hypothesized to have direct effect and indurect effects on physical or mental health through mediators such as perceived burdens, use of coping and perceived rewards. Each hypothesized path within the context of both revised structural models represent specific hypotheses (Rgure 6, page 183 & Rgure 7, page 187). A total of seven direct and three indirect causal effects were tested, respectively. Both direct and indirect causal effects are interpreted and discussed as follows.
Direct Causal Effects
HI. Caregivers' physical and mental health is directly positively affected by subjective beliefs of traditional filial obligation, use of coping and perceived rewards, and negatively affected by perceived burdens.
The first hypothesis was partially supported; all direct effects were supported, except the negative relationship between caregivers' physical and mental health and use of coping (Rgure 3, page L07). As predicted, findings showed that caregivers' physical and mental health was directly and positively affected by subjective beliefs of traditional filial obligation. The direct positive relationship between beliefs of traditional filial obligation and physical helath (P= .08) or mental health (P= .07), respectively, indicates that a stronger sense of beliefs ui traditional fib'al obligation leads to better physical or mental
health. According to the literature review, no research has specificaEy addressed the 194 direct effect of subjective beliefs of traditional filial obligation on caregivers' physical and mental health. This is the first study to examine that the direct effect.
Another direct positive relationship was found between perceived rewards and physical helath (P= .27) and mental health (P= 26), repectively. This expected relationship suggests that a stronger sense of perceived rewards leads to better physical or mental health. This finding is consistent with a previous study (Stephens, Franks, &
Townsend, 1994), suggesting that caregivers' physical well-being was predicted by sense of rewards. In their study, a positive association between sense of rewards and physical well-being was showed in a sample of 95 women with multiple roles including caregivers, mothers and wives.
There was a direct negative relationship between perceived burdens and physical health (P= - .03) and mental health (P= - .03), respectively. This indicates that a stronger sense of perceived burdens leads to poorer physical health. The adverse impact of caregiving strains on caregivers' physical or mental health has long been recognized in literature (Clybum, Stones, hadijstavropoulos, &. Tuokko, 2000; Dew et al, 1998; Hale et al., 1987; Pratt, Wright, & Schmall, 1987). In a study of 613 caregivers of relatives with
Alzheimer's disease, caregivers' well-being was mediated by appraisal of burdens (Clybum,
Stones, Hadijstavropoulos, & Tuokko). Di another sample of 133 caregivers of heart transplant recipients. Dew and colleagues (1998) found caregivers' physical and mental well-being declined as perceived burdens increased. The negative relationship between caregivmg strains and health was also supported m a study comparing caregivers to non- caregivers (Haley et al, 1987). Haley and colleagues found caregivers with, a higher sense 195 of burden had a poorer health status than did non-caregivers. Pratt, Wright and Schmall
(1987) also reported a poorer health status among caregivers with a higher sense of caregiving strains.
Surprisingly, a direct negative relationship between use of coping and physical health (p= -.18) or mental health (P= -.21), respectively, was found. This negative relationship was in contrast to the direction predicted in both revised models. The negative relationship between use of coping strategies and physical and mental health implies that the use of coping was correlated with declining physical and mental health. It can be interpreted that caregivers, who use coping strategies more often, either behavioral-oriented orcogm'tive-oriented strategies, may have worse physical and mental health than those who use coping strategies less often.
The unexpected relationship between use of coping strategies and physical and mental health can be explained in three ways. First, it may be suggested that the statistical control for indicators of the use of coping was less than perfect to the extent that the measurement model may have omitted important indicators or items. According to the literature review, although the ways of coping has been widely used in social psychological literature and grouped into two categories, it has been criticized for inconsistencies in factor structure, which results ui problems of interpretation due to unstable construct vaHdity (Atkinson & Violato, 1993; Ayers, 2001). In this study, since only one measure, ways of coping, was used, the construct validity may have been limited due to an unstable factor structure, bi addition, some items cross-loaded on more than one factor. It is likely that other clear and comprehensive coping measures are needed. 196
Second, it is also possible that mediated factors in the study did not include all antecedent factors that naight affect the direction of relation betweea use of coping and physical health. Studies have shown that demands of care (Leavitt et al, 1999; Eitzner, &
BCnutzen, 1998), and types of stressor (Folkman & Lazarus, 1980; Folkman et al, 1986;
1987; Vedhara, Shanks, Anderson, & Lightman, 2000) seem to be important influences on coping choice, which were omitted in this study. Folkman and colleagues (1987) found coping options were associated with different types of stressors perceived by different age populations. While older adults perceived health problems as stressful, nuddle-aged adults cited family activities. Eitzner and BCnutzen (1998) reported that caregivers experienced significant stress related to the demands of care, which determined coping options.
Third, the two indicators, behavioral-oriented and cognitive-oriented coping, for use of coping might be also considered here. According to Lazarus and Folkman (1984), different coping responses may have different effects on health perceptions in different situations. For example, the coping strategy of escape-avoidance is an appropriate coping response for crisis situations to decrease the negative effect of anxiety reacuon, but it is not a good coping strategy for situations requmng healthy behavior. Haley and assodates found that use of coping had mediating effects on family caregiving, however, the impact of use of coping strategies on caregivers' health varied from study to study. While some studies discover certain types of coping strategies can relieve the unpact of caregiving burden on caregivers' health (Haley, 1997; Henkle, 1994; Lee, Brennan & Daly, 2000;
Teel, Duncan, & Lai, 2001), Neundorfer (1991) reported there was no significant relationship between use of coping and physical health. Some studies have even found 197 that over-use of some types of copmg strategies adversely effect health (Billings et al.,
2000; Dew et al., L998). In a longitudinal study of 253 caregivers involved in AIDS- related caregiving^ caregivers' physical health was indirectly affected by high use of cogm'tive coping strategies such as avoidance. In a stud.y of 133 caregivers of heart transplant recipients. Dew and associates (1998) found a poorer physical health among caregivers, who used avoidance coping strategy more frequently, [n addition, Picot
(1991) suggests that use of coping strategies is influenced by individuals' ethmc backgrounds. Picot indicates that while Africa American caregivers tend to use cognitive- oriented coping for their psychological and physical needs, Anglos prefer behavioral- oriented coping. Use of specific strategies is shaped by individuals' cultural influences.
This may provide an explanation for why use of coping strategies did not positively relate to physical health.
H2. Use of coping is directly positively influenced, by perceived burdens.
The second hypothesis was supported as a positive relationship was in agreement with the theoretical basis proposed in both revised model. The relationship between perceived burden and use of coping was the same (P= .241) for both physical health and mental health models, which suggest that as caregivers' perception of burden mcreased, they tended to use copmg strategies more frequently. This &iding implies that use of coping is triggered by perceived burden, which is suggested in Lazarus and Folkman's
(1984) stress-coping theory. Lazarus and FoDcman defiled coping as protective
psychological and/or behavioral processes mduced by stressfiil environmental demands.
This is empirically supported by Campbell's (1997) study. In a sample of 80 spouse 198 caregivers of dementia, Campbell found that use of coping strategies was associated with stressM appraisals.
H3. Perceived burdens are directly, negatively affected by subjective beliefs of
traditional Mai obUgation.
This hypothesis was also supported. The relationship between subjective beliefs of
traditional filial obUgation and perceived burdens was also the same (P= -J30) for both
physical health and mental health models, which suggest that caregivers who have a
weaker belief about traditional filial obb'gation have a greater sense of caregiving burdens.
This finding is supported in literature (Chou, LaMontagne and Hepworth, 1999; England,
1990). By using SEM on a sample of 150 Chinese caregivers of relatives with dementia in
Taiwan, Chou, Lamontagne and Hepworth (1999) reported that filial obligation had a
significant direct negative effect on caregiving burden. In a study of 168 adult children
caring for demented parents, England (1990) indicates that adult child caregivers with
stronger beliefs about traditional filial obligation had less caregiving burden than did those
with weaker beliefs about traditional filial obligation
H4. Perceived rewards are positively influenced by subjective beliefs of
traditional fiUal obligation.
As predicted, this hypothesis was also supponed. The relationship between
subjective beliefs in traditional filial obligation and perceived rewards was the same
(P= .37) for physical health and mental health models, which suggest that stronger beliefs
about traditional filial obligation lead to a stronger sense of perceived rewards. The
findmg is supported by Hsueh (1998)) in a qualitative pilot study of two informants in 199
Taiwan. These two informants believed that care for dependent elderly family members was an unquestionable responsibility for adult children (Hsueh). They agreed that sense of duty in providing care to dependent elderly family members motivated them to provide family caregiving. One informant indicated she felt satisfied with the caregiving she provided. The other informant indicated that she felt proud of getting through the journey of caregiving. Jones (1996) reported a similar finding in another qualitative study of 20
Asian caregivers of elderly family members.
Indirect Causal Effects
In this section, the total effect values for each indirect causal path are interpreted based upon model predictions for possible explanation(s) and followed by a brief discussion.
HI. Caregivers' physical and mental health is affected by subjective beliefs of traditional fih'al obligation through the negative mediated effect of perceived burdens.
The total for the indirect effect of beliefs of traditional fiKal obligation on physical health and mental health were the same (-.30 x -.03 = .01). Findings suggest an indirect
positive relationship between filial obligation and physical and mental health, although the
effect is very small. Findings imply that caregivers with stronger behefs about traditional
filial obligation tend to perceive less caregiving burden, which leads to better physical and
mental health. Findings support the negative mediating effect of perceived burdens
reported by Jones (1995; L996) in a qualitative study of 20 Asian caregivers of elderly
family members in the United States. Findings are also supported by another and by a
study of 148 caregivers of partners with cancer (Nijboer et al., 2000). 200
H2. Caregivers' subjective beliefs of traditional filial obligation has an indirect effect on physical and mental health through the mediated effect of perceived burdens and use of coping.
An indirect effect between caregivers' beliefs of traditional filial obUgation and physical and mental health was found. The total effects were (-.30 x .24x -.18 = .01) for the indirect path between beliefs of traditional fib'al obligation and physical helath, and
(-.30 X .24x -.21 = .02) for the indirect path between beliefs of traditional filial obligation and mental health. This suggests, consequently, there is an indirect relationship between fiUal obligation and physical and mental health, respectively, through two mediating effects, perceived burden and use of coping. According to the literature review, the relationships of beliefs of filial obligation and perceived burdens (Chou, LaMontagne, &
Hepworth, 1999; Cicirelli, 1993; Raveis, Karus, Pretter, 1999), perceived burden and use of coping (Blankfeld, Deborach, Holahan, & Charles, 1999; Saldana, 1999), and use of coping and physical health (Lee, Brennan & Daly, 2000; Teel, Duncan, & Lai, 2001) have been documented. However, the underlying processes among those variables are still poorly understood due to inconsistendes in findings relating to these variables. No literature explores the relationships among filial obEgation, perceived burden, use of coping and health status by using structural equation modeling (SEM), which this study did. The strength of SEM was to identify the essential process of effects among a set of variables by confirming mterrelationships and mediator patterns among them. By using
SEM, the underlying process of the complexity of the caregiving experience and caregivers' physical and mental health was uncovered. 201
In re-examining this indirect causal path, taken as a whole, some light is shed on the seemingly contradictory findings in the relationship between use of coping and physical health, and between use of coping and mental health. It suggests that caregivers, who had stronger beliefs about traditional filial obligation tended to have less sense of caregivng burden, which led to use of coping strategies, either behavioral-oriented or cognitive- oriented, less frequently. As caregivers used coping strategies less frequently, they had better physical and mental health. The overall indirect effects suggest that caregivers' beliefs about traditional filial obligation had a positive indirect effect on physical health
(-.30 X .24 X -.18 = .01) and mental health (-.30 x .24x -.21 = .02), respectively, although the magm'tude was small.
When only the direct effects between variables were reviewed, they seem to contradict each other and conflict with previous findings in literature. When the whole path is examined, the underlying processes clarify previous contradictory findings about the relationship between fiUal obligation and physical and mental health. This indirect path suggests that perceived burden and use of coping served as mediating factors to counterbalance the effect of caregivers' beliefs of filial obligation on their physical and mental health. Furthermore, it also provides an explanation for how caregivers' health was indirectly affected by cultural mfluences that might also influence other mediating factors such as perceived burden and use of coping options.
H3. Caregivers' physical or mental health is positively mfluenced by beliefs of
filial obligation through a positive mediated effect of perceived rewards. 202
The total was (.37 x .27 = .10) for the mdirect path between beliefs of traditional fih'al obligation and physical health, and (.37 x .26 = .10) for the indirect path between beliefs of traditional filial obligation and mental health. As predicted, this path supports the positive effect of subjective beliefs in traditional filial obligation on physical and mental health. This finding indicates caregivers who have a stronger sense of beliefs in traditional fiUal obligation also perceive a greater sense of reward; and this leads to better physical and mental health.
Although increasing attention has been paid to the positive aspects of family caregiving (Carruth, 1996; Carruth, Holland, & Larsen, 2000; Lundh, 1999; Picot, 1991) rarely has caregiving research focused on the positive aspects of filial obUgation (Jones,
1995; 1996). Jones reported that traditional filial obligation plays an important role in motivating and enabling Asian caregivers to engage in a long-term family caregiving in a qualitative study on 20 Asian caregivers of elderly. She also indicated that caregivers who have a stronger sense of behef in filial obh'gation tend to feel satisfied with family caregiving. Phillips (1988) suggested that subjective beliefs and values have a crucial role in influencing the caregivers' perceptions, actions and reactions to family caregving. BQein
(1989) indicates that caregivers' positive reactions to family caregivmg are connected to thenr positive sense of obligation.
The total effects for this path, reinforce confidence m tests of the previous hypotheses that there is a positive relationship between cultural mfluences and health. The total effects also provide an explanation for how caregivers' health was indirectly affected by cultural mfluences that might also affect the mediatmg factor of perceived burden. 203
Results not only provide aa explanation for how caregivers' health is affected by cultural influences but also suggest caregivers' cultural beliefs have significant contributions to their physical well-being.
Summary
The results of this study support the hypothesized direct and indirect effects of subjective beliefs of traditional filial obligation on physical and mental health proposed in the both revised models, except for the direction of the effect between use of coping and physical and mental health. For the unexpected relationship between use of coping and physical and menatl health, three explanations are possible including measurement issues related to onutting important indicators or items, unstable theoretical basis due to omitting antecedent factors that might affect the direction of relation between use of coping and physical and mental health, and measurement issues related to cultural differences.
Findings suggest that caregivers' beliefs about traditional fiUal obh'gation is a primary predictor for caregivers' physical and mental health. Caregivers' subjective beliefs of traditional filial obligation affect caregivers' physical or mental health not only directly but also indirectly. Caregivers' beUefs of traditional filial obligation affected physical or mental health through mediators including perceived burdens, use of coping and perceived rewards m three ways: (I) through the chain of indirect effects leading fi:om subjective beliefs of traditional fib'al obligation to perceived burdens to use of coping to physical and mental health, (2) through the chain of mdirect effects leading fi:om caregivers' beliefs of traditional filial obligation to perceived burdens to physical and health, and (3) through the 204 chain of indirect effects leading from caregivers* beliefs of traditional fflial obligation to perceived rewards to physical and mental health.
While Caregivers' beliefs of traditional fflial obhgation affect their physical health
(P= .08) or mental health (P= .07) directly, the overall total indirect effects of subjective beliefs of traditional filial obligation on physical health are [(- .30 x .24x - .18 = .01) +
(-.30 X -.03 = .01) + (.37 X .27 = .10)] = .12. The overall total indirect effects of subjective beliefs of traditional Sial obligation on mental health are [(-.30 x .24x - .21 =
.01) + (-.30 X -.03 = .01) + (.37 X .26 = .10)] = .13. Since the indirect effects of subjective beliefs of traditional filial obligation are stronger than the direct effects, the magm'tude of subjective behefs of traditional filial obhgation effects on physical and mental health appear to be indirect. The results also support the mediated effects of the causal processes.
Findings in Relation to Theoretical Orientations
In this section, findings in relation to three theoretical perspectives including social exchange theory, role theory and stree-copmg theory for the conceptual framework are discussed.
Social Exchange Theory
Findings are consistent with sodal exchange theory and support that the exchanges among family menibers are voluntary or mvoluntary depending upon the family members' perceptions of each other's important needs (Robson &Thumher, 1979), which are trigged by caregivers' cultural beliefs. The predicted relationships between cultural beliefs, and. perceived burdens and perceived rewards support that family caregiving is a 205 reciprocal relationship that is perceived as having costs (burden) as well as rewards (Rossi
& Rossi, 1990). Findings alos suggest the essential role of sodal exchange in understanding the process of family interaction. Although the perceptions of needs between caregivers and care recipients were not test in this study, findings support that perceived burdens and rewards are possible responses of the redprocal interaction, family caregiving.
Role Theorv
Results are consistent with role theory, which assumes during a cargiving interaction the caregiver has a moral and social obligation directing the caregiving role that is determined, by cultural beliefs (Klein, 1989). Findings are also consistent with the conceptual framework that cultural beliefs have a positive, dnrect effect on perceived rewards and consequence of family caregiving, respectively, among Chinese caregivers.
The positive relationship between cultural beliefs and perceived rewards suggest that an
individual's role behaviors and responses tend to blend sodal expectations with self- expectations to situate ongoing interactions (Ktryker & Statham, 1985). To meet moral
and social obligations, a caregiver's reactions are formed in the images of her/himself in
fulfilling role responsibilities ui relation to family caregiving. The individuals' subjective
cultural beliefs not only provide the avenue to enable an individual to engage in redprocal
activities but also determine the responses and consequences of family caregiving role
taking (Phillips, 1988). 206
Stress-coping Theory
The overall findings are consistent with stress-coping theory and suggest that personal factors (i.e., beliefs) play important roles in influencing an individual's perceptions of and reactions to stressful situation (Lazarus & Folkman, 1980) and the consequence of family caregiving are determined by how a person appraises it (Haley et al., 1987). The variation in interpretations of and reactions to stressful situations is determinded by the cogm'tive appraisal, which is defined as the buffering system in mediating the encounter and the reaction (Lawton, 1989).
As predicted, there was a positive relationship between perceived burden and use of coping and suggests that use of coping was triggered by perceived burdens. The negative relationship between perceived burdens and caregivers' health suggests that when individuals experienced environmental demands that resulted in negative emotional responses, their physical and psychological well-being were impacted also (Pearlin et al.,
1990). In constrast to expectation, a negative relationship between use of coping and caregivers' health status was found. Given these results, more careful exploration of the relationship between perceived burdens and use of coping and between use of coping and health is needed.
Implications
This study is an initial testing of relationships between beliefs in traditional fiKal obligation, perceived burdens, perceived rewards, use of coping strategies and physical health to determme the impact on caregivers' health status. The results of this study have several implications for nursmg. 207
For Nursing Research
Findings showed caregivers' subjective beliefs of traditional filial obligation had a direct influence oa perceived burdens, perceived rewards and physical health. Whfle subjective beliefs of traditional filial obligation had both direct and indirect impacts on caregivers' physical health, use of coping had a negative mfluence on caregivers^ physical health, which was in contrast to the relationship proposed in the revised model In addition, use of coping has been identified as an important mediator in counterbalancing the appraisal and consequence of family caregiving, future studies should explore types of coping preferences among Chinese caregivers and how they relate to perceived burdens, perceived rewards and health status.
Since research is vital to discover the overall outcome of those impacts, intervention focus research should be conducted. For example, developing effective coping strategies, teaching helpful self-care method or relaxation strategies to help caregivers to cope and/or decrease physical and mental stress results of caregiving.
Future studies should also explore the relationship of subjective beliefs of traditional filial obligation, perceived burdens, perceived rewards, use of coping and health status in different populations, such as America-bom Chinese or Anglo caregivers. In addition, as society becomes characterized by multicultural diversity, comparing different ethnic groups for understanding the discrepancies of the impact of filial obligation on perceived bxirden, perceived reward, use of coping and health status are suggested. 208
For Nfursing Practice
Findings suggest that caregivers' cultural influence should be considered and included in nursing care. Data presented in this study indicate that cultural influences, deeply rooted in Chinese culture, play an important role in affecting perceived burdens and caregivers' health. Since the cultural influence of filial beliefs had significant contributions to alleviating perceived burdens which lead to perceived rewards, the assessment of subjective cultural beliefs should be an integral part of the nursing care plan for Chinese caregivers. In this study, use of coping seemed to be an important mediator as well as a
potential intervening factor influencing caregivers' health. As more knowledge about
coping preferences and effective coping options for Chinese caregivers are explored,
nurses can help to provide better and more comprehensive care. Knowledge of
caregivers' cultural influence on appraisal of and responses to the caregiving role provides
important information on which to base culturally appropriate care. BCnowing the
caregiving experience among the Chinese caregivers encourages nurse to provide
culturally sensitive care, which could also improve the quality of care and quality of life for
both caregiver and care-receivers.
Clinically, this study of the family caregiving experience and health status suggests
important implication for nursing interventions because it is a theoretical explanation to
account for the process of change in family caregivng. This model should provide usefiil
information for nursing inlervention because important antecedent variables, underlying
processes and outcome were tested. This model not only predicted outcomes but also
identified underlying processes, which mediate outcome. Identification of underlying 209 mediated patterns between filial obligation and physical health provides an explanation for conflicting findings in previous studies to guide nursing practice.
For Nursing Education
Current nursing education emphasizes cultural sensitivity and cultural diversity.
Information obtained firom this study provides preliminary understandings regarding the
Chinese family caregiving experience that may contribute to substantive knowledge building, as more empirical data are available in the future. To ensure culturally sensitive nursing skills in the areas of family care giving, nurses should be aware of the cultural factors in predicting or mediating the consequences of family caregiving. Findings suggest that culturally sensitive care to Chinese caregivers or other ethnic groups should be encouraged. Findings on how Chinese culture affects caregivers' health provides an example to help nurses develop and implement culturally sensitive care plans for Chinese caregivers and other ethnic groups.
For Theorv Construction
The study employed a structural equation modeling approach to identify the underlying essential processes and mediated patterns of family caregivers experience among the Chinese. The predicted and unpredicted findings in this study are usefiil for further theory development and testing. Needs for modifications of the prelimmary model for Chinese family caregiving experience and health status were suggested in the data, specifically the variable of use of coping. Findmgs firom this study may help clarify critical antecedents and mediators in predicting outcome. 210
According the literature review, the family caregiving experience may vary across individuals as well as cultural groups. The findings in this study have demonstrated that beliefs about traditional Glial obligation, a cultural influence, plays a crucial role in predictmg the appraisals and consequence of family caregiving among the Chinese. The theoretical model identified in this study helps to communicate the cultural domain for nursing practice for understanding and assessing cultural specific interventions for the
Chinese caregivers.
Finally, the process of constructing the model, testing the parameters, and revising the model based on statistical Judgment and theoretical knowledge led to a greater understanding of the essential components of family caregiving and procedures for model testing. The structure of the model provides a firamework for expanding the picture and filling in the nussing pieces as more research findings become available.
To Policy Making
Findings suggest insights about health care policy needs for Chinese caregivers.
Data suggest that policy makers should pay attention to the needs of Chinese caregivers in terms of their own health as well as family cultural needs. Although the current health care system emphasizes multi-cultural competency, it still needs to pay more attenuon to addressing different cultural perspectives in. care. Policy must reflect the same support for all ethnic groups. Hndings also provide usefiil mformation implications for policy makers to develop health care service programs that could benefit Asian family caregivers and develop effective programs to assist family caregivers firam other cultures. For example, a
home-based mterventiott program^ focused on balancmg the cultural influence on Chmese 211 caregivers and cultural expectation of dependent elderly family members may help Chinese caregivers with their caregiving role. Furthermore, it would be helpfiil and meaningful to
Chinese caregivers to provide a home-based long-term care program, with financial support and educating on problems oriented to enhance care rather than developing
institutional long-term care facilities.
limitations and Recommendations
A revised model based on statistical judgment and theoretical knowledge was
tested in this study. Two best-fit revised models were developed based on the CFI and
/ df ratio for a goodness-of-fit. The first revised model was identified as the best fitting
model for this study as indicated the CFI = .91 and x' ^ df ratio = 2.9. The second revised
model only yielded an acceptable value SO for the CFI but its / df ratio test was 3.5,
which failed to meet the criterion of less than 3. Therefore, before developing conclusions
for this study, limitations need to be acknowledged.
Limitation of Design
The design of this study has some impact on internal and external vaUdity of the
study fundings. The major limitation of this study related to its design. The study was
nonexperimental and applied a cross-sectional approach with a convemence sample using
self-report measures. In the future, a longitudinal study is suggested to add confidence for
makmg causal inferences 6:om the data.
There were also several important limitations for generalizing the findings related
to sampling bias. One potential important limitation is that the sample ia this study was
predommately foreign-bom Chinese immigrants. The ftidings may not be representative 212 for those who are American-bom Chinese. Therefore, the generaiizability of study results is limited. Another potential important limitation is that most of subjects identified themselves as Christians. The findings in this study may not represent the picture for those who were non-Christians or Buddhists orTaoists, which are historically dominant religions among the Chinese. The other limitation is that the majority of subjects in this study were well educated and had middle to upper incomes. The caregiving experience among this population may differ from those who are less educated and have a lower social status.
Limitations of Interpretation of Model Fit
In this study, the sample size was rather small and it would be premature to apply this empirically derived theory without fiirther testing on a larger sample. Furthermore, the interpretation of a model, which fits the data to the model adequately based on the fit indices but the Chi-square test significantly rejects , is still debatable. The fit indices were partially satisfied in this study; only x' /df ratio and the CFI provided preliminary evidence of the validity of the model. Confidence about the adequacy of model is limited.
Re-specifying parameters for identifying a more adequate theory to better explain the data may be needed.
Another limitation is that some possiblely important independent variables or mediating factors were omitted due to too small variance or were not collected in this study. For example, the variable of subjective feelmgs about providmg assistance was omitted from model testing due to too smaE variances. Factors such as caregiving demand and caregiving involvement may need to be included, but were not. 213
Consequently, the results in the current study would, represent a conservative assessment of effect.
The other limitation is that the criterion of three indicators for each latent variable was not met in this study. AE latent variables were underidentified; this means that less than three indicators were used to measure aH latent variables, except for the latent variable of perceived burden. This limitation may weaken the results of model fit testing.
Perhaps more factors in a larger sample are required to explore the overall impact of being a caregiver of a dependent elderly. Future studies should explore the relationship of subjective beliefs of traditional filial obligation, types of coping options and health status to identify effective coping strategies for health improvement. In addition, since only one coping measure was used, data indicate that a more stable model with more indicators for coping measure for ftiture studies is needed. To reduce the bias due to fallible measures, more work is required to select appropriate indicators for each latent variable.
Conclusions
Originally, this study attempted to explain the relationships among fiUal obUgation, subjective feelings about providing assistance, perceived burden, perceived reward, use of copmg, and health status. The original model did not have a good fit with data.
Two revised models were developed based on the information given in the results of the original model testmg. For the new revised models, the variable of subjective feelings about providing assistance was omitted and the parameters for health status were re-specified. The two indicators for health status were tested separately. Among the two revised models, the first model, which deftied physical health as the outcome variable, was 214 determined as the best fitting model as indicated by CFI= .91 and the / df ratio = 2.9.
The second model, which tested mental health as outcome, had a minimumly acceptable good fit with data as indicated by CFI= .90 and the / df ratio test = 3.5, which was failed to meet the criterion of less than 3.
All hypotheses were supported except for one. All direct and indirect path effects between fifa'al obligation and physical and between fiUal obligation and mental health were confirmed, except for an unexpected negative direct effect between use of coping, and
physical and mental health. Results suggest filial obligation is the primary predictor for
perceived burden, perceived reward, use of coping strategies and physical and mental
health. Filial obligation had a direct and four indirect effects on physical and mental
health. Perceived burdens, perceived rewards and use of coping strategies were mediators
between filial obligation and physical and mental health. APPENDIX A
HUMAN SUBJECTS APPROVAL 216
MUNwaiarfOF tbiuii Subjects Coamiiuee ARIZONA./llXl^^zlvve 1622E.Mib(lSd((t PO Bo* 245137 HEAUHSCIlNasCENia Biam.AZ8Sn<.5U7 (}20)62M721
19 June 2000
Kuei-Hsiang Hsueh, Ph.D. Candidate Advison LindaPiullips, Ph.D. College of Nuising POBOX210203 •M
RE: FAMILY CAREGIVING EXPERIENCE AND HEALTH STATUS AMONG CHINESE CAREGIVERS INTHE UNITED STATES
Dear Vb. Hsueh:
We have received documents conceningyour above lefmnced project Regulations published by^ the U.S. Department of Health and Humaa Services [45 CFRPart 46.101 (b) (2)1 exemptthis^ oftesearch&omreviewbyoutComfflittee.Note:Eagliih/CIiincscvenioBsofaUqacstioiiiMiits must be submitted for revieir pinr to me. Please be advised that ckarance fron officU authorities for site(s) where proposed recruitment is to be conducted must be obtained prior tomitutionatcachsite.
Thankyou for ufotmug us of your wodL youhave any questions concemingthe above, please contactthisofBce.
Smcerely,
David G. Jbhnson^W Chaomaa Huniaa Subjects Committee
DGJ4s cc: Department^ColIege ReviewCommittee 2L7
MllNivasiirop
HmuD Subjcds Commiuce ARIZONA. 1622&MibelSiRa B0.taZfiU7 HEAUHSdQKESCENm 1ticiaa.AZ83724-SU7 (S20)S2Mm
19 June 2000
Kuei-Hsiang Hsuefa, Phi). Candidate Advisor Linda Phillips, PhD. College of Nusing POBOX210203 sK-d
RE; FAMILY CAREGIVING EXPERIENCE AND HEALTH STATUS AMONG CHINESE CAREGIVERS IN THE UNITEO STATES
DearMs.Ibueb:
We have received documents concern^ your above Rfeesced project Regulations public by the U.S.DepartnientofHealthandHuniaaSetvices [45 CFRPait 46.101 (b) (2)] exempt this type ofreseaich&om review by our Committee. Note: Englnh/Ch&icie versnis of allqaestioMuuts must be submitted for review prior to uie> Please be advised that ekanoce from oflktil authorities for sitc(s) where proposed recmitmeit b to be coidieted must be obtained prior tomitiatioaateachsitc.
Hunk you foriiifonnihg us of your wotk. If you have any questions concent the above, please contactthiso£5^ce.
Sincerely,
David G. niflfrman Human Subjects Committee
DGJ/js cc: Department/College ReviewCommittee APPENDIX B
SUBJECTS DISCLAIMER FORM 219
UNIVERSITY OF ARIZONA HEALTH SCIENCE CENTER SUBJECTS DISCLAIMER FORM
Family Caregiving Experience and Health Status among Chinese Caregivers in the United States
You are being asiced to voluntarily participate in a study to explore the family caregiving experience and health status of Chinese caregivers in the United States. With your permission, I would like you to answer a survey questionnaire. It takes you about 25 to 35 minutes to finish. There are no known risks iavolving in your participation. You may choose to withdraw fix)m the study at any time with no consequence at all. By responding the survey questionnaire, you will be giving your consent to participate in the study. Your identity will not be revealed and your confidentiality will be maintained in all reports of this project. You will receive a thank-you card with a package of 10 stamps for your contribution to this study if you return the survey package. Your contribution will help health care professionals to understand cultural factors that influence health problems of Chinese family caregivers. The results of this study will advance nursing knowledge for improving the quality of care and quality of life for both caregivers and care-receiver. For further information, please call me, Kuei-Hsiang (feueh, at 520-320-1578. If you have questions concerm'ng your rights as a research subject, you may call the Human Subject Committee office at 626-6721.
Thank you
Investigator Date
Telephone Number 220
APPENDIX C
PERMISSION LETTERS 221
]aK23,2QQ0
THE UNIVERSITY OF AtABAMA rCtC V(FO*lt KnofUttlbKii U04Ka«]rAte,AcclB T«aaii,AZ 15719 DorKiottui^
I am hocb^^ pving jott petoMkn to ad to aaiubtt iiao CliMM, iqr Go* o( Gwlndg^pC^tfattmiioaaiybeawBaiiedtfaeronwipmcwoCQieliKlB.
I m abo Kadmg to yoii theOQwib KonoK aid oodug uMnaiaiu.
Good kckoaToiirduMftaioiL Snotre^,
UMeM taRBBM TiKiiMii.«
Terms of Agreemcatfor Uie of the Pkot Circghrer Reward Scale
The user of tiie Picot Caregiver Reward Scale ^CRS) must agree to the foUowiog conditions:
I. Whea using the PCRS, you must indicate that Sandra L Fulton Picot is the owner of the scale's copyright by placement of the foilowmgcopyri^tnoticeonall copies ofdie' • scale:
• cl995bySandraJ.FultonPicot
Picot her approval. The translated PCRS must include the copyright notice.
3. The PCRS may not be distributed for sale.
4. No alterations may be made to PCRS without the prior written permissioa of Sandra J. FaltoaPicot,PhD.
5. U^completioaofAeptopo8eduse,acopyofyourfindingsandpsychomdrics should be sent to Sandra J. Fulton Kcot, PhD, RN, FAAN, Uhiversity^ ofli^land. School of Nursing, 655 West Lombard, Baltnnor^ Muyland 21201.
To indicate youragreemoitto the above conditions, please sign and date both copies of the contract and send one signed copy to Dr. Picot Keep the second signed copy for your personal files. Yoa m^ reach Dr. Picot by E-mail at sn)[email protected], telephone: 216>368>5980, or fine 216-368-3542.
Signature & d^ of the PCRS usa Signature & dAte of the copyri^t owner
M Of mm. 223 CEKIEFORAlDSPREVEhrnONmJDlES TEl;415»79i5i 74NEVMOKrGOMERY,SUIIE6i)0 FAX>415»7«U SANFRANCISCO,CAlJFOIim94I05 UftlLSFOUMANgR&UCSFIDU
SUSANFOIXMAN.PRO. PROFESSOROFMEIHCINEAND CODIRECm
DearCoileagae: He Ways of Coping that was revised in 198S is in the public domain and you do not need special pennission to use it In 1988 the Consulting Psychologists Pi^ made minor modifications to a fewitems. Their version is copyrighted, andif you wish to use their version and/ortheir scoring service, you'llneedtheirpennission. Youcanreachthemat PO Box 10096, Palo Alto, CA 94303 (6S0/969-8901). Youmight also want the manual for the Ways ofCopingJt is available throughthesamepublisher. Sincerely,
SusanFolkman,Ph.D. Professor ofMedicine 224
filiilRdy Scale Plgelofl
h»: ]fUi(SIIIP:]fidi01f|ifejiiicudMw]
Tr. Uiliiueb9Bunfli|iinoiiidi
Cte
Sh^ HiilKeijrSade Sol: (l21/l)07;tfPI( Impntiice: NionDil
Dear Miss Hsueh, [ am Kuaog-hui Yeh, (he other author of yourresquest for use thefilMpie^ scale that developed by Professor K. S. Yang and me. I am glad you are mteiested iDthfiscaieandtiytoputitintoyourdissertatioaieseardi iVowE email you the paper that introducing the sacle by attached file, but thepaperwaswrittenbyChinese. Ihopeyoucanopenandieadii Can you tell me where shall [send you the materials if you wantacopy of the scale with scoring instruction?
Best regards
^g'huiYeh Institute of Eihnolo^ AcademiaSinica
D?w????.doc APPENDIX D
INSTRUMENTS 226
Demographic Information
Dear Friends: I would like to invite you to answer this anonymous survey. This questionnaire is designed to better understand your experiences and feelings about providing care to your dependent elderly, your current health and beliefs about traditional filial piety. Your contribution will help healthcare professionals and social workers to further understand your feelings, health and needs. This study may also provide information to other caregivers. Your identity will not be revealed and your confidentiality wilt be maintained in all reports of this project. You will receive a thank- you card with a package of ten 33 cents stamps for your time and contribution to this study. If you have questions about this survey, please call me at 520-320-1578 or write e-mail to [email protected]. Thank a million again for your time and help.
Kuei-Hsiang Hsueh, Ph.Dc, R.N. University of Arizona College of Nursing ID L Caregiver Gender Male=I Female=2 Age: (in years) Where were you bom?_ _ (If you can not find answer from the following list, please write down the answer in the blank). American =l Canada =2 China =3 HongKong=»t Indonesia =5 Malaysia =6 Singapore =7 Taiwan =8 What is your religion?_ (If you can not find your answer fixim the following list, please write the your own religion in the blank).
None=l Christian=2 Buddhist =3 Taoist =4 Jewish=5 Mormon=6 Presbyterian=7 What is your education? (in years) What is your occupation now? Clerical/Sales (private organization) Clerical/Sales (public organization) Manager (private organization) Manager (public organization) Teacher (middle or high school) Instructor Professional Self Employment _ Housewife Never work. Retired Unemployment Other Would you say your yearly household income is between: Less than S10,000 _ 510,000-519,999 _ 520,000 - 529399. S30.000 -539,999 . 540,000-549,999. 550,000 - 559,999. S60.000 -569,999 . 570,000 -579,999. 580,000 - 589,999. 590,000-599,999. 5100,000 - 5124,999. 5125,000-5149,000. 5150,000 or above. How long have you been lived in the United States: ycan(s). Month(s). How long have you been a caregiven year(s)_ Month(s). What is your relau'onship with care receiver? Father/ daughter Mother/ daughter. Father/son. Mother/son. Father-in-law / daughter-in-law Mother-in-law / daughter-in-law. Father-in-law / son-in-law Mother-in-law / son-in-law Other n. Care Recipient Genden Male=l Female=2 Age: What is your care receiver's education?. .(in years) What type(s) of illness does yourcare receiver has? None Diabetes Arthritis. Kidney disease Cancer. Stroke Parkinson's disease Heanng problems. Hypertension/Heart disease. Dementia. Hip ftacnire COPD. Eye problems Other Would you describe yourcare receiver's cognidve status (most of the ume) as: Alert Forgetful Disorientation Confuse. Would you describe yourcare receiver's mobility (Most of the time)as: Ambulatory. Using cane or walker. Need assistance ftom other person. On wheelchair Whatcaregiving take(s) do you provide to yourcare receiver (indicate as many as you provide)? Cooking Feeding Laundry Bathing Walking Moving orchange position Toileting Soiling Shopping Cleaning room. Answering phone call(s) Medicadon Handling mail(s) or biU(s). Transportation assistance. Other 228
Poulshock & Deimling's Burdea Inventory ntA. Being a caregiver, you are expected to providing some instrumental help to your dependent relative, please indicate whether you find any of the following tiring, difficult and/or upsetting. PUT AN (X) ON AS MANY LINES AS APPLY TO YOU.
Help Needed No Help But No Needed Problem Tiring Difficult Upsetting Bathing
IDtessing
Toileting
Moving
Care after wetting & soiling Feeding
Grooming (comb hair. brush teeth, shave) 229
Cost ofcare Index
As a caregiver, have you bad the foilowing experience? Please rate the extent to which you agree with the following statements describe the potential and/or actual feelings of caring for you care receiver. Choose only one answer for each statement. Note: the term he or she refers to the care receiver.
WA'Sce^DS »^iSiSangR^
IKelEtfiatKaimpSc^IffiSsnelpMRMfKj^^pitS^K^ m tss*1 2 I feel that as a result of caring for he/she t do not (will not) I 2 3 4 have enough tinie for myselfl ;If.^|tBifeCTDmg^j^Be^caus5a^ng^Bg^ gpi
4 I feel that caring for he/she has negatively (will negatively) I 2 3 4 affected my appetite. 5 LfeetthaccaiuigfociEe/shepuifCwilLputXarstraiaoa - t 2 3- fairalvreratronshrpsi- ; •. 6 I feel that my family and I must give up (will have to give up) l 2 3 4 Qecessities because of the expense to care for he/she. C fe^^^dLring^t Be/sEe;d^ my^ routine _ t 2 S 4
8 [ feel that caring for he/she has caused me (will cause me) to 1 2 3 4 be physically fatigued. iI:fe^tIiac;iinrfaimIy[and!rcaiin^C^ m 7 4-". 1 tfi&ejn^oiS^beraa^St^li^tpeiie^^K^ 10 I feel that caring for he/she has caused me (will cause me) I 2 3 4 tobecomc anxious. 11 li^Ltbaccamgjpcp^shesmte^i^Cwipm^ . * ' t- 3" , : • ffiendis'bn5rends6rmyfSmi n t feel that caring for he/she is (will be) too expensive. I 2 3 4 230
Picot's Perceived Reward Scale
m-C. The following statements are about some of the ways people feel about caring for another person. Please tell me how you feel now about caring for your dependent lamily member. Choose one and only one answer that best describes your feeling for each statement Note: the term he or she refers to the care receiver.
to:«i3ate«>»in.aawiaa^aiawyiat'faaK«BisdaB^^ 1 I feel better about myself. ••EfMlilfiavcBsoomeSstKimritblorahcmiaS^ : iiiperson:ar(Hind3DCtsonjs,wtKsicEnMs:ocBMiiIii(ap^^ tin 4 [ f^l having others say that taking care of my relative is the right thing to do is important. lfeeE^tmjEtcIatry§;^IfreTO 2I3 focr^fCarKS:?^!:;; f 6 I feel someone wilt take care of me when 1need it. j'tfeeEriurses^doctOTsmdisddaJavbrie^^ •:.-r:^carefbi;in®ieIatwe:tooi^'i!;3;.sS:ia'^»!s^iVsi£;t3fe;rS'^^^^ 8 [ feel that placing my relative in a nursing home will be avoided.
;• 9 rfeettjmtdaaci^nttts^an eveiStfittiJe^outiniytdativersclmcs^ - „V ilHIiiiiiilllii 10 I feel receiving a smile, touch, eye contact &om my relative is important. 11 CfMEIhayeadpser relatippsfaip wftfMxrelative^ : W rj 12 I feel I have an opportunity to repay my relative for a past debt. i^ffMtreaiVinga^fnianfcyo^
14 I feel I have become a better person by teaming new information to care for the elderly. :ERetrEtaveBeCTnie;xbettecpersoitb»leaming^CT^w^3^^ ^ --•tacafftfetMeiifeiscrifiiSssSS^ 16 I fral that I have made many new fnends
IS Lfeel [ have the freedom to make decisions that matter. ^rUN: i;.. ••. -EfralilTdo noCneedta^ 20 r fral receiving praise and adtmrabon for my efforts from doctors, nurses, and social worker is important. i-^lifiMtJlrcaEncHiii^laiinn^&vimEkiMiligeadtfdmcixss'lif
22 I frcl happier now than 1 did before I started caring for mv relative my relative^ ]rES«Piat?cainif5gnwnJatmsI^n^^ ';:rafaM^arawifiliwSc"c^«fij^aS;4^tS%ii#Slrl2i5^ 24 I f«l my family members now took up to me because of my e£forts under difHcult circumstances. '^r^E^EBavmemjitcIatEvcslwcmtKmetnieiuisadded'inoneysr^?^^:^^^ fe3CommEmtO!Bouset'isl>!gKi£:»iig^fet^<2^^ 231
Ways of Coping
III-D. In general, when people have something happen to them or when they have a problem, they will use some ways to handle their problem. Being a caregiver, inevitably, you must encounter one or some problems in your life related to family caregiving. Please read each item below and indicate to what extent you use it in the situation you just encounter.
SlJSBMnro
alp^cbncent^dm^BacIBaitadajiaiiMBeniKB 2 I try to analyze the problem in order to understand it better. 0 I 2 3 3<^igSS IS WT J 2fw'^u;::dLU!i3E.i:j*r,i:L4ne''VlitSiZSjiilC.i': 4 1 feel that time will have make a difTerence- the only thing is 0 I 2 3 to wait. CbaisaihroECompromisc tagecsomethmg BOsitiVe SboLtGer i: Q ^ L Z. J situaiion^ . 6 [do somedung (hat I didn't ihink would work, but at least I 0 I 2 3 am doing something. ftrytOi^ttfifcDeESOttresponaBrttd'cBaiiKBiKdEfiecmnfe^^ StfJsi 8 [ Calk to someone to find out more about the simation. 0 1 2 3 EctiticoiebErectatemy^E. ' ,0 : L 3 10 Try not to bum my bridges, but leave things open somewhat. 0 I 2 3 SsOJ^'ir,^ S-i-ejJiSc-L 12 [ go along with fate; sometime I just have bad luck. 0 I 2 3 13 EgponasifhcthihgiBadihap^ned- • ; " ^' IsprlH Z - 14 [ try to keep my feelings to myself. 0 I 2 3
ntBiiis 16 L sleep more than usual. 0 I 2 3 sLE:':Emie^ani^toft&£:MhonCsy.wBci^ IS 1 accept sympathy and understanding &om someone^ 0 I 2 3 |isi SIS 20 1 am inspired to do something creative about the problem. 0 I 2 3 fsiii 22 1 get professional help. 0 t 2 3 sflislpa smm 24 1 wait to see what would happen before doing anything. 0 1 2 3
lifeslfe 26 [ make a plaa of actioa and follow it. Q I 2 3 m ^^8 232
Never Use Use quite Use a use somewhat a bit great deal iCIeEniKfeamKoucsomefibwilJrs^S im: Sw*Ki;'r^ 29 [ realize Chat [ had brought the problem on myself. 0 I 2 3 f3m EEe^vepamnowbiett^thSiEv^ ia®®-IISiB 31 [ calk to someone who can do something concrete about the 0 I 2 3 problem. 0- L 2 S. 33 [ tried to make myself feel better by eating, drinking, 0 I 2 3 smoking, using dmgs. or medication, etc. 5|taEeii^cffijp^|d^iTO^^igg^|tii^
35 [ try not to act coo hastily or follow my first haunch. 0 I 2 3 i:m ~:V:~Trv^ •\frX:Jv;':nrLr':>rr a•^^£5=^5-^ 37 1 maintained my pride and kept a sti^ upper lip. 0 I 2 3 3S •Etedii5CbvepwBat;isinipo^efcli^^TS':E :;X|3||R 39 [ change something so things would turn out all right. 0 I 2 3 40 Eayor£6errigwith.peopIet 0 I •:>:."::Sv:S. 41 I don't let it get Co me; 1 refused to think too much about tt. 0 1 2 3 •:-42?_ ; Easfcadyi 53 I accept Che situation, since nothing can be done. 0 I 2 3 asttstgr 5fStlslS 55 [wish that I could change what is happening or how [ feel. 0 I 2 3 T^tSiaumg^MfDet^gaEfbuiii^^ 'iifc- !;!«Sji2sS-iSj4 57 r daydream or imagine a better time or place than the one L 0 I 1 3 am in. S» irwtsEtfiaiit^s{MtiMvradId3i^i.v^ccs 59 [ have fantasies or wisli about bow things might turn out. 0 I 2 3 233 Never Use Use Usea use somewhat quite a great bit deal fift 61 I prepare myselffor the worst, 0 L 2 3 63 [think about how a persoa [admire would handle this 0 1 2 3 situation and used that as a model. ^^':aa,'U,»'^.'tL 65 [remind myself how much worse things could be. 0 L 2 3 6 SF-3^ Health Survey m-E. The roUowfng questions ask your views about your health. Please read eadi item below and indicate how you reel about your healths Choose one a^ only one answer that best describes your feeiiags for each statement. I. In general, would yoa say your health is; Excellent Very good Good Fair Poor 2. Compared to one year ago. how would you rate your health in general now? Much better Somewhat About Somewhat Much worse now better now the same worse now now 3. The following quesa'ons are about activiu'es you nughc do during a typical day. Does your health now limit you in these activities? If so. how much? Yes, Yes, No, Limited Limited not limited a lot a little at all 3-1 Vigorous activities, such as running, lifting heavy obfects. partictpatinK in strenuous sports. 3-2 Moderate activities, such as moving a table, pushing a vacuum cleaner, bowling, or playing golf. 3-3 Lifting or carrying groceries 3-4 Climbing several flights of stairs 3-5 Climbing one flight of stairs 3-6 Bending, kneeling, orstooping 3-7 Walking more than a mile 3-8 Walking several blocks 3-9 Walking one block 3-10 Bathing or dressing yourself 4. During the past 4 weeks, have you had any of the following problems with your work or other regular daily activities as a result of your physical health? Yes No 4-1 Cut down on the amount of time you spent on work or other activities 4-2 Accomplished less than you would like 4-3 Were limited in the kind of work or other activities 4-4 Had difficulty performing the work or other activities (for example, it took extra effort) 235 S. During the past 4 weeks, have you had any of the following problems with your work or other regulardaily activia'es as a result of any emotional problems (such as feeling depressed or anxious)? Yes No 5-1 Cut dowa on the amount of u'me you spent oa work or other activides 5-2 Accomplished less than you would like 5-3 Didn'tdo work or other activities as carefully as usual 6. During the past 4 weeks, to what extent has your physical health or emotional problems interfered with your normal social activities with family, fnends, neighbors, or group? Not at all Slfghtly Mbderaldy Quite a bit Extremely 7, How much bodily pain have you had during the past 4 weeks? None Very mild Mild Moderate Severe Very severe 8. During the past 4 weeks, how much did pain interfere with your normal work (including both work outside the home and housework]? Not at all A little bit Moderately Quite a bit Extremely 9. These quesuons are about how you feel and how things have been with you during the past 4 weeks. For each question, please give the one answer that comes closest to the way you have b«n reeling. AUoT Most Agood Some AlitUe None of the of the bitot of the of the the time time time the time time tinie 9-t Did you feel fiill of pep? 9-2 Have you been a very nervous person? 9-3 Have you felt so down in the dumps that nothing could checryouup? 9-4 Have you calm and peaceful? 9-5 Did you have a lotof ener^? 9-6 Have you felt downhearted and blue? 9-7 Did you ftel worn out? 9-8 Have you been a happy person? 9-9 Did you feet tired? 236 10. During the past 4 weeks, how much of the time has your physical health or emotional problems interfered with your sodal activities (like visiting friends, relative, etc)? Alter Most of Some of A little of None of the time the time the time the time the time 11. How TRUE or FALSE is each ofthe following statements foryou? Ddlnitdy Mostly Don't Mostly Definitely true true laiow false false in [ seem to get sick a little easier than other people 11-2 I am as health as anybody 1 know 11-3 I expect my health to get worse lt-4 My health is excellent 237 Yang's Revised FUial Belief Scale ni-F. Everyone has different belieb about bow to treat parents or elderly. Now, I would like you to tell me your belief about how you should treat your parents or elderly. Please read each item below and indicate to what extent you agree with. Please choose one and only one answer that best describes your reeling for each statement. 2 1 feel that E should show my parents that they are 0 I 2 3 tmpoctanttome. rl^'^TECeeltGillEsfioardlpwattentioi^^ S;0£S£ HisnS:IJSiffigSEiss'jaSS 4 [ feel that 1should nuke scarifies to take care of my 0 I 2 3 parents. - • tieatinentsiasEcatflocmKpannts^^io^aiesidK^ PB 6 r ftel that Lshould provide living expenses to myparents 0 I 2 3 to make them as comfortable as possible. mm 8 I feel that I should say something for my parents in order 0 I 2 3 to save thetr face. ni-G. Everyone has varying degrees of affection toward hiis/her parents. Now, I would like you to tell me about your affection toward you parents. Please read each item below and indicate to what extent you agree with. Please choose one and only one answer that best describes your feeling for each statement. mst mssmmBBms mSB^M 1 niw m^saSmSS^ MmBusm 2 Would you describe your feeling toward your parent as 0 1 2 3 detached? ::^WbutiIvQiic.£icti6qvoi^fiietm£ T" Would you describe your filing toward your parents 0 1 2 3 as levins? W: T Would you describe your filing toward your parents 0 [ 2 3 as solicitous? 5^S7'PVoaia^i£i^ailKgydacie3Sgl£mn£iw w UCL^Ktibiut *r^SiiiSrfeSli^ T" Would you describe your feeling about parent as 0 I 2 3 emotional cold? 238 l««Xd9IE»S<. (+!i®«33 cents Sll)s . |[jffir«Bil520-320-1578 . SftSe-mailS [email protected] IllCfi S«llll(MB99F$eM1«±9F5e^ -V : m: (1 )S (2)^ «»¥»: (S*) ^CtR4tXB« .iia :4»Bv S*. Kfftltt. m. HB..) 38$5^Sa9iii±i6S6:__(»iBSB« .«!in:fBv JUJlEffiSv ®ig. MBv v) c^m ¥ M 239 (i)M_ (2)««»_ (4)«ifc_ (5)jift_ ) «»«*««: ¥(^SD 14¥v . V ) (1)ifcA'^^IIJl_ (2)ittj[f«li«*_ (3)»llfit«AJl_ (5)iaii?«iiiia_ immtSLmmmm_ (9)SS±» (iO)^Mxi^_ (11)5^* (12) (13)^ft6 (HatiJ ) C2)¥« (3)/«X<^ (l)-*ll7cfi(.T (2)-M=*ll7C (3)-M=jril^ (4)=M0*ll7C (5)HM£*ll7C (6)2SAKIl7C (7)AM-fcrJlllx (8)-bMAMIl7c (9)AMAirJt7c (10)AStiril7c (11)tE-h-«ll7C (12)+^MtH*ll7C ^(13)tia*ll7cJ5t± ^ (2)«A: (3)5^^- (4)#? {6)««« _{7)3t AiC* {8)3t«»i« _(9)« »*«*: 'h»»ji*6¥v =f»*sji9¥v . : m um^ *«3sus M «a/MaifiL itIfllB/'C.^•fll nf iRii«iflr(a e m (»* m: ) : {i)?is_ (2)«^ _ (3)i!«iii _ (4)Si. _ tSIIJIJSj : (1)SI6g » _(2)3lfflffia_ (3)li^Afiti$_ (4)ifell« itft ««: 3feic m» wot. m Jzm jt»SM «««& «33i(®s#»eirv ±»M. wm^ma^fsm RfSdiatii: ) 241 •Sl^ mm. 9lSi ±mm Ammm m± mmmmm. msf-. siiff) B. A'A-fiiiH#. «iitsn^T9\iBm ? ii«cii3ieadiui, laff X«ll (l^«Sft). m: Tms « M nt m 1^ X 1. 4. liM^/ifi«il[SddRnSS. 5. Ai^mwrn/ib, iKSftiKi5$AddRi!iii». 6. ABmwm/iA. 8. 242 MffX (llXSK)o B: T$ 6, S]lffiff3iE»S«1iSI|, #YAII«S. 7. ssmer. 9. sumairv 10. %mi&/mnmRvm±iSim. »sft«sxa9. 11. 12. ii*iiaf6/ttas««#iiiifl^/ttigsa9 jRiff. i3.s)ii«iiBM^tt»-S''ttH". fisftftasa^. u.avaiifadiiaAa. ssftSBUsnrs^. i5.»mifad!ia;^^. 16. l7.»Stl»SftUS. 18. ssffSYttu^n s«. 19.SSff«^Tfta±9E. 2o.«xiia9siaav •±ia^xAAa9BiisRB, fiswi. 21. 22,«S9%^MR$6jia^tti¥. wsm. 23. iniaK«&/tta«f!±«««fi£WiK«-SKUi[^tt. 24. , KsasaAaas. 25. ssi«si£^i6asfl^£-&ii:. it«ad^3tE}iKitt;!n mmms:. 243 * IT n « m w $ m # ift 1- aRM^ST-i^SiS^aCttte. 2. aK#jS-«f|i!*3t5E7*1i3l, 5. aK*Jstsiaw:j?5C^#«-!£Eisiif3i, 6. aft«tT-«aR»aTKft.MS«aiicii?«t7-!»¥o 7.9.itnx^tmimis.m. 8. a;&K*«A»t. 3i£E7»if3ta 9. a#iastifiiit9i[SBo 10. aK#;^iBffias;^», ii-a«a#-ffl««a{?i. 12. . aft«i(ii«. 13. a««^i!r^. u.aKirfiifl&iBddss. i5.aK»«iffa9Ms«ffi. le.atb^xffrsffR^, 17. a#»jf«nsdd AiE^'C^«adiiis. 18. a#aiS3i^3^Addi^it2itt3iii. 19. a#iafts-^iigmii&aits»'-fta9¥. 20. attiui^-ii»nii«0[»ddir. 2i.aK*:S#-«o 22.a*#}g«iiA±adis«r. 23- a)»#a»36«ft«S-fflA, 24. mmmm, 25. a*;&-«itsi, n « * n M V $ M B| # fl m 27. . SS/Xffteo 28.S*KS»«JI«FIflF». 29.S3fllilSl&g5«inHo 30. SSSltl:llfM$6l$adlE«t^. 31. A»o 32. ie#i»{rAi£st3i(ff^iitnnH-To 33.s*ii«e m. 34. Mt3tSlilJinHo 35. a#=JSlil^^To 36.a«SJIiffll'i>, 37. 38. 4o.a-tt3tifeaA«-iB, 41. Btt, 42. 43. 44. 45.a#«Aiii£'C.^«ad«K. 46.a*si$i5d95t«. 47. A*±, 48. mTA^. 49. a^aiaa^sitftfhs, miiiiftmisiimmmo 50. 53. , a#SISiff3£(%K). 245 n * n « ir $ m M w # 54. 55. $#iMg3[se»^n*iesM«]i. 56.iigB#-eajc», 67. 58. $#aiff>:Eiiia^ifiK^Sii. 59. 60 64. 65. mmst i Bviff 66.S1i%lfiXtt. 246 1. -ttJtEie, m» » itw (^») 2. a-^iKTik. «K»«Ei l(r»-tt«*sttaisnw? »-K -« m-u 3. mu ? m irfiE -» mi tfe^ 3-l.^jbfi^l. mm. ftH«v #Alll^;tiill. 3-2. cfITKjtjtiliSl. jOHBS'f. )IIIRSII. 3-3. 3-4.IRIIff«V»K«. 3-5.JR-l«itfl9(i», 3-&.SIK»Tv Hievs. 3-7.^-ll^»a 3-d.£«l«l%II£liad«F. 3—9-^.0 247 ft. lasikmnm^o ft 5 4-1 4-4 Bb£^iB#KKSiddK», 5. ^nii^d^Q ftffl. 3a:ftSB9;Dtffit:^aiddii^«. ^stT^iinHS^ ? mjkiinmjk, m ft 5 5-1. 5-2. A£^iBiffit;^sa9ft 6. BiiBRfli«r:)5-®»»k. fts^«a«iE*-ittttXjf 16. aasA, m. -It *- 1fe^ fttt 7. , «fti?W(ir$^)*s±»0iff (jnaffv mritv nm. ?««)? -It *. *. M i^tss mmn m <^991 248 8. »«««. nm. s^mm). -IT *- m tfe^ BIT $ $ m « *• ft — * m t w ft $ m lb It It it 9-1.3a5*S*«ii5tJtl3EJt? 9-2. ? 9-4. ^&SVa(SI¥l»ID?$fidS(l ? 9-5.3K*5t«**fi^||jtl? 9-6.:«jtl?ir*B«t. BHdditJI? 9-8. «*§»«* S A ? 9-9.:3!jiil1ta(iJftfl|? 10. iiit;^iia9nK, ? Br* ^mm 9i'> -U mm 249 11. m * fi m ¥ » ¥ & ft at ft ft 11-1- SftL^tfcKffiW AESim 11-2. ii-3.sffns»«ii*sii. ii-4.a»«jii^*». 5. 6. SXffSjiKK»5^$, . 8. ai7«K^«»ir'f. 250 — * 11 u — n * Us, it 1. 2. 3. 4. 5. 6. «»^^«ftS/C.^#l&±? 7. 8. 251 ilEFERENCES Aboud^F. (1988). Children and Prejudice. New York: Basil Blackwell. Aldwin, C. M., & Revenson, T. A. (1978). Does coping help? A reexamination of the relationship between coping and mental health. Journal of Personality and Social Psychology. 53.337-348. Alley, J. M. (1988). Family careeiving: Family strains, cooing response pattern, and caregivers burden. Unpublished doctoral dissertation, Virginia Polytechnic Listitute and State University. Allen-Holmes, L. M. (1998). The role of physical health, mental health, and help orientation in the health promotion behaviors of informal careigvers of imparted elders. Dissertation Abstracts International Section B. the Sciences & Engineering 58 (lO-B). Almberg, B. Grafstrom, M., & Winblad, B. (1997). Major strain and coping strategies as reported by fanuly members who care for aged demented relatives. Journal of Advanced Nursing. 26 r4). 683-91. Amirkhan, J. H. (1994). Criterion validity of a coping measure. Journal of Personality Assessment. 62 (2). 242-261. Amirkhan, J. H. (1990). A factor analytically derived measure of coping: The coping strategy indicator. Journal of Personality and Sodal Psychology. 59 (5). 1066- 1075. Aneshensel, C.S., Rutter, C.M., & Lackenbruch, P.A. (1991). Social structure, stress, and mental health. American Sociological Review. 56 (2). 167-178. Anonymous. (1997). On the nature and origin of psychosomatic symptoms. Scandinavian Journal of Work. Environment & Health. 23 SUDPI 3.105-12. Arbuckle, J. L. (1997). Amos user's guide. Chicago: Samall Waters. Archbold, P. G., Stewart, B.,Gteenlick,M.,&Haryath,T. (1990). Mutuality and preparedness as predictors of caregiver role strain. Research in Nursing & Health. 13. 375-384. Aranda, M. P., & Knight, B. G. (1997). The influence of ethnicity and culture on the caregiver stress and coping process: a sociocultural review and analysis. Gerontologist. 37 (3). 342-54. 252 Atrindell, W., & Vanderende, J. (1985). An empricial test of the utility of the observations-to-variables ratio in factor and components analysis. Aootied Psychological Measurement. 9 (2). 165-187. Atkinson, M, & Violato, C. (1993). A factor analysis of the Ways of Coping Questionnaire based on data from saddening experiences. Psychology Reports. 72.1159- 1164. Ayers, S. (2001). Assessing stress and coping in pregnancy and postpartum. Journal of Psychosomatic Obstetrics & Gvecology. 22(1). 13-27. Ayers, S. (1999). Post-traumatic stress disorder following childbirth. Unpublished doctoral dissertation. University of London. Averill, J.R., Wisocki, P.A. (1981). Some observation on behavioral approaches to the treatment of grief among elderly. In Sobel, H. (Ed.), Behavior Therapy in Temu'nal Care: A Humanisric Approach. Cambridge. Mass: Ballinger. Azen, S. P., Palmer, J. M., Carlson, M., Mandel, D., Cherry, B. J., Fanchiang, S., Jackson, J., & Clark, F. (1999). Psychometric properties of a Chinese translation of the SF-36 Health Survey Questionnaire in the well elderly study. Journal of Aging and Health. 11 (2). 240-251. Baker, C. (1991). Cultural awareness in nursing. Canadian Nurse. 87(11). 38-40. Baines, E. M., & Oglesby, F. M. The elderly as caregiver of the elderly. Holistic N^ursing Practice. 7(1). 61-69. Banken, C. H. (1995). Alzheimer's patient caregivers: experimental drug treatment participation as a moderator of caregiver outcomes. Unpublished doctoral dissertation, Texas Technology University. Barber, C. E. (1988). Correlates of subjective burden among adult sons and daughters caring for aged parents. Journal of Aging Studies. 2. 133-144. Barusch, A.S.,&Spaid, W.M. (1989). Gender differences in caregiving: why do wives report greater burden? Gerontologist. 29 (5). 667-76. Batt-Leiba, M. I., Hills, G. A. (1998). Implications of coping strategies for spousal caregivers of elders with dementia. Topics in Geriatric Rehabilitation. 14 (1). 54- 62. Benner, P., & Wrubel, J. (1989). The primary of caring: Stress and coping in health and illness. 253 Berry, J, (L992). Methodological concerns. In L Berry, Y. Pootinga, M. Segal & R. Dasen OEds.), Cross-cultural Psychology, (pp. 210-246). Cambridge: Cambridge University Press. Berry, J. (1980). Introduction to methodology. Di H. Triandis & J. Berry (Eds.), Handbook of cross-cultural psychology (Vol 2, pp. 1-28). Boston: Allyn & Bacon. Berry, R. A., & Murphy, F. J. (1995). Wellbeing of caregivers of spouses with Parkinson's disease. Clinical Nursing Research. 4 (4). 373-86. Biegel, D. E., Milligan, S. E., Putnam, P. L., & Song, L. (1994). Predictors of burden among lower socioeconomic status caregivers of persons with chronic mental illness. Conrniunitv Mental Health Journal. 30 (5). 473-94. Billings, D. W., Folkman, S., Acree, M., & Moskowitz, J. T. (2000). Coping and physical health during caregiving: The roles of positive and negative affect. Journal of Personality & Social Psychology. 79 f 1). 131-142. Boland, D. L., & Sims, S. L. (1996). Family caregiving at home as a solitary journey. Image: Journal of Nursing Scholarship. 28.55-58. Bowers, B. J. (1987). Ditergenerational caregiving: Adult caregivers and their aging parents. Advances in Nursing Science. 9 (2). 20-31. Bradbur, N. M. (1969). The Structure of Psychological Well-beinjg. Chicago: Aldine. Braun, K. L., & Browne, C. V. (1998). Perceptions of dementia, caregiving, and help seeking among Asian and Pacific Islander Americans. Health & Social Work.. 23 262-274. Bricker, P. L., & Fleischer, C. G. (1993). Social support as experienced by Roman Catholic priests: the influence of vocationally imposed network restrictions. Issues in Mental Health Nursing. 14(21219-34. Brislin,R. (1980). Translation: Applications and Research (Ed.). New York: Wiley-Halsted. Brislin, R.W. (1970). Back-translation for cross-cultural research. Journal of Cross-cultural Psychology. 2 (3). 185-216. 254 Brody,E.M, (1981). Women ia the middle and family help to older people. The Gerontoloeist. 21 (5). 471-481. Browers, B. J. (1987). Intergenerational careigving: Adult caregivers and. their aging parents. Advances in Nursing Science. 9 (2). 20-31. Brown, P. L, Sloman, R., Brown, L.M., Mitchell, R. (1995). Coping strategies of caregivers of family members with dementia. Australian New Zealand Journal of Mental Health Nursing. 4 (2). 70-6. Brown, MXJD. (1990). Familv caregivers' and their spouses' wavs of coping with an elder parent in their home. Unpublished doctoral dissertation. University of Pittisberch. Bull, M.J. (1990). Factors influencing familycaregiverburden and health. Westem Journal of Nursing Research. 12 (6). 758-776. Bull, M. J., Maruyama, G., & Luo, D. (1997). Testing a model of family caregivers' perceptions of elder behavior two weeks posthospitalization on caregiver response and health. Scholarly Inquiry for Nursing ^ctice. 11(3). 231-55. Bush, R. C. (1970). Religion in Conmiunist China. Nashville, Teen: Abingdon Press. Cameron, C. F. (1990). An ethnonursing study of the influence of extended caregiving on the health status of elderly Anglo-Canadian wives caring for physically disabled husbands. Unpublished doctoral dissertation, Wayne State University. Campbell, J. C. (1997). The relationship ofcaregiverdemands, appraisal, and resources with spousal caregiver adaptation to dementia caregiving. Unpublished doctoral dissertation, Indiana University School of Nursing. Cantor, M.H. (1983). Strain among caregivers: A study of experience in the United States. The Gerontologist. 23 (6). 597-604. Carruth,A. BL (1996). Development and testing of the caregiver rectprocity scale. Nursing Research. 45.92-97. Carruth,A.K. (1996). Motivating factors, exchange patterns, and reciprodty among caregivers of parents with and without dementia. Research in Nursing & Health. 19 (5). 409-419. 255 Carruth, A K., Holland, C., & Larsen, L. (2000). Development and psychometric evaluation of the Caregiver Reciprocity Scale H. Journal of Nursing Measurement. 8 (2V 179-191. Caver, C.S., Scheier, MF., &Weintraub, J.K. (1989). Assessing coping strategies: A theoretical based approach. Journal of Personalitv and Social Psvcholoev. 56 (2). 267-283. Chambers, L. W., Sackett, D. L, Goldsmith, C. H., Macpherson, & McAuley, R. C. (1976). Development an application of an index of social function. Health Service Research. 11.430-441. Chan, R. C. K., Lee, P. W. H., & Lieh-Mak, F. (2(X)0). Coping with spinal cord injury: Person and marital adjustment ia the Hong Kong. Spinal Cord. 38(1 H. 687-696. Chang, B.L (1999). Cognitive-behavioral interventioa for homebound caregivers of persons with dementia. Nursing Research. 48 (3). 173-82. Chang, B. H., Noonan, A. E., & Tennstedt, S. L. (1998). The role of religion/spirituality in coping with caregiving for disabled elders. Gerontoloeist. 38 (4). 463-70. Chapman, D. W., & Carter, J. F. (1979). Translation procedures for the cross- cultural use of measurement. Educational Evaluation and Policv Analysis. 1 (3). 71-76. Chapman, K. J. & Pepler, C. (1998). Coping, hope, and anticipatory grief in family members in palliauve home care. Cancer Nursing. 21 (4). 226-234. Chenier, M. C. (1997). Review and analysis of caregiver burden and nursing home placement. Geriatric Nursing. 18 f3). 121-126. Chen, W.T. (1963). Analects: A source Book in Chinese Philosophy. Princeton, NT: Princeton University Press. Cicirelli, V. G. (1993). Attachment and obligation as daughters' motives for caregiving behavior and subsequent effect on subjective burden. Psychology and Aging. 8(2). 144-155. Clark, BL K., Bormann, C. A., Cropanzano, R. S., & James, K. (1995). Validation evidence for three coping. Journal of Personality Assessment. 65 (3). 434- 455. 256 Clark, M. S., Mills, J. (1979). Interpersonal attraction in exchange and sonununal relationships. Jouranl of Personality and Social Psychology. 37.12-24. Cleary, B. L, (1989). Alzheimer's disease: stressors and strategies associated with caregiving (Alzheimer disease). Unpublished doctoral dissertation. The University of Texas at Austin. Cochrane, J. J., Goering, P. N., & Rogers, J. M. (1997). The mental health of informal caregivers in Ontario: An epidemiological survey. American Journal of Public Health. 87 fl2). 2002-7. Cohen, S., Groves, L., Borden, W.,& Lazarus, L. (1989). In E. Light &B. Lebowitz (Eds.), Alzheimer's disease treatment and family stress: Directions for Research, pp. 50-105. Rockville, Maryland: national Institute of Mental Health. Confucius. (1982). The book of filial piety. In P. L. Mckee (Ed.). Philosophical Foundations of Gerontology, pp. 103-113. New York: Human Sciences Press. Coleman CK. Piles CL. Poggenpoel M. (1994). tofluence of caregiving on families of older adults. Journal of Gerontological Nursing. 20 (1H. 40-9.56. Connell, C. M., & Gibson, G. D. (1997). Racial, ethm'c, and cultural differences in dementia caregiving: review and analysis. Gerontologist. 37 (3). 355-64. Cook, K. S. (1987). Social Exchange Theory. Beverly Hills: Sage pulications. Chou, K. R., LaMontagne, L. L., & Hepworth, J. T. (1999). Burden experienced by caregivers of relatives with dementia in Taiwan. Nursing Research. 48(4). 206-14. Chou,K. (1998). Caregiver burden: structural equation modeling. Nursing Research fChina). 6(5). 358-371. Chu, H. Y. (1985). The impact ofdifferent religions on the Chinese family in Taiwan. In J. C. Hsieh 8c. Y. C. Chuang (Eds.), The Chinese Family and It^s Ritual Behavior, (pp. 129-154). Institute of Ethnology, Academia Sinfca, Monograph. Series B, 15 (Chinese). Choi, H. (1993). Cultural and non-cultural factors as deternunants of caregiver burden for the impaired elderly in South Korea. TheGerontologist. 33 (1). 8-15. 257 Dai, Y.T., &Dimond,M.T. (1998). Filial Piety: A cross-cultural comparison and its implication for the well-being of older parents. Journal of Gerontological Nursing. 24 (3). 13-18. Dai, Y. T., (1995). The Effects of Family Support. Expectation of Filial Peitv. and Stress on Health Consequences of Older Adults with Diabetes Mellitus. Unpublished doctoral dissertation, Unversity of Washington. D'Andrade, R., & Strauss, C. (1992). Human Motives and Cultural Models. Cambridge: Cambridge University Press. Day,J.C. (1996). Population projections oftheUriited States by age, sex, race, and Hispanic origin: 1995 to 2050. Current population reports, p 25-1130. Washington, DC: U.S. Gov. Print. Dellasega C. (1991). Caregiving stress among community caregivers for the elderly: does institutionalization make a difference? Journal of Community Health Nursing. 8 (4). 197-205. Dellasega, C. (1990). Coping with caregiving. Stress management for caregivers of the elderly. Journal of Psychosocial Nursing & Mental Health Services. 28 (1). 15- 6,19-22. Deimling,C.T., &Bass, D.M. (1986). Symptom of mental impairment among elderly adults and their effects on family caregivers. Journal of Gerontology. 42 (7). 778- 784. DeVries, H. M., Hamilton, D.W., Lovett, S., & Gallagher-Thompson, D. (1997). Patterns of coping preferences for male and female caregivers of frail older adults. Psychology & Aging. 12 (2). 263-7. Dew, M. A., Goycoolea, J. M., Stukas, A. A., Switzer, G. E., Simmons, R. G., Roth, L. H., & DiMartini, A. (1998). Temporal profiles of physical health in family members of heart transplant recipients: predictors of health change during caregiving. Health Psychology. 17 (2). 138-51. Dwyer, J. W. & Miller, M. K. (1990a). Difference s in caregiving network by area of residence: finpUcations for primary caregiver stress and burden. Family Relations, 39^27-37. Ehrlich, E. G. (1992). The nature of the experience of women caregivers for aged parents: a social problem. Unpublished doctoral cUssertation, Columbia University Teachers College. 258 Elliott, Kathryn Sabrena. Di Minno, Mariann. Lam, Darrick. Tu, Alicia Mei. (1996). Working with Chinese families in the context of dementia. Di Y., Gwen, D., Gallagher-Thompson et al. (Eds.), Ethnicity and the dementias, (pp. 89-108). Washington, DC: Taylor & Francis. England, M.C. (1990). Test of a model for caregiver strain. Unpublished doctoral dissertation: Case Western Reserve University Enyert, G., & Burman, M. E. (1999). A qualitative study of self-transcendence in caregivers of terminally ill patients. American Journal of Hospice & Palliative Care. 16 (2L 455-62. Farran, C. J., Keane-Hagerty, E., Salloway, S., Kupferer, S., & Wilken, C. S. (1991). Finding meaning: an alternative paradigm for Alzheimer's disease family caregivers. Gerontoloeist. 31 (4). 483-9. Fengler, A. P., & Goodrich, N. (1979). Wives of elderly disabled men: The hidden patients. The Gerontologist. 19 (2). 175-183. Ferketich, S., & Verran, J. (1990). Causal modeling. In Moody's Advances in Nursing Science. VH. Newbury Park, CA: Sage. Fillenbaum, G. G. (1979). Social context and self-assessments of health among the elderly. Journal of Health and Social Behavior. 20.45-51. Finch, J. (1989). Familv Obligation and Social Change. Polity Press: Cambridge. Finch, J., & Mason, J. (1990). Family obligation and kin support for elderly. Aging and Society. 10.151-175. Fingerman, K. L., Gallagher-Thompson, D., Lovett, S., & Rose, J. Ditemal resourcefulness, task demands, coping, and dysphoric affect among caregivers of the frail elderly. International Journal of Aging & Human Development. 42(3). 229-48.1996. Fink, S.V.W. (1992). Factors affecting the well-being of caregivtng families. Unpublished doctoral dissertation. University of Michigan. Finley, L. Y. (1998). The cultural context: families coping with severe mental illness. Psychiatric Rehabilitation Journal. 21 (3). 230-40. 259 Finley, N. L, Roberts, M. D., & Banahan, B. F. (1988). Motivator and inhibitors of attitudes of filial obligation toward aging parents. Gerontologist. 28(1'). 73-83. Eitzner, M. A., & knutzen, R. (1998). The impact of pituitary disease on the family caregiver and overall family fiinctioning. Psvchotheraov & Psvchosomatics. 67 (3), 181-188. Folkman,S. (1984). Personal control and stress and coping process: A theoretical analysis. Journal of Personality & Social Psychology. 46 (4). 839-852. Folkman, S., & Lazarus, R. S. (1988). Ways of Coping Questionnaire Sampler Set: Manual. Test Booklet. Scoring Key. Redwood, CA: Mind Garden. Folkman, S., & Lazarus, R. S. (1985). If it changes it must be a process: A study of emotion and coping during three statges of college examination. Journal of Personality and Social Psychology. 48.150-170. Folkman, S., & Lazarus, R. S. (1980). An analysis of coping in a middle-aged community sample. Journal of Health and Social behavior. 21.219-239. Folkman, S., Lazarus, R. S., Dunkel-Schetter, C., Delongis, A., & Omen, R. (1986). The dynamics of a stressful encounter: Cognitive appraisal, coping, and encounter outcomes. Journal of personality and Sodal Psychology. 50.992-1003. Folkman, S., Lazarus, R. S., Dunkel-Schetter, C., Delongis, A., &Gruen, R. (1986). The dynamics of a stressful encounten Cognitive appraisal, coping, and encounter outcomes. Journal of personality and Social Psychology. 50.992-1003. Folkman, S., Lazarus, R. S., Gruen, R., & Delongis, A. (1986). Appraisal, coping, health status, and psychological symptoms. Journal of Personality and Social Psychology. 50.571-179. Folkman, S., Lazarus, R. S., Pimley, S., & Novacek, J. (1987). Age differences in stress and coping processes. Psychology and Aging. 2.171-184. Freed, A. 0. (1990). How Japanese family cope with fragile elderly. Journal of Gerontological Social Work. 15 fl/2). 39-55. Fry,C. (1996). Age, aging and cultural, fii R. H., Binstock, L. K., George ^ds.). Handbook of Aging and the Social Science (4"^ ed.). (pp. 117-136). Fudge, H.,Neufeld, A., Harrison, M. J. (1997). Social networks of women caregivers. Public Health Nursing. 14 fl). 20-27. 260 Fuh, J. L, Wang, S. J., Liu, H. C., & Wang, H. C. (1999). The caregiving burden scale among Chinese caregivers of Alzheimer patients. Dementia & Geriatric Cognitive Disorders. 10 (3). 186-191. Gallagher, D., Rappaport, M., Benedict, A., Lovett, S., & Silver, D. (1985). Reliability of selected interview and self-report measures with fanu'lv caregivers. Paper presented at the Annual Scientific Meeting of the Gerontological Society of American, New Orleans. Gates, H. (1987). Chinese Working Class Lives: Getting by in Taiwan. Ithaca, Cornell University Press. Gaynor,S. E. (1990). The long haul: The effects of home caer on caregivers. Image: Journal of Nursing Scholarship. 22 (4). 208-212. Gelinas, D. F., O'Connor, P., Miller, R. G. (1998). Quality of life for ventilator- dependent ALS patients and their caregivers. Journal of the Neurological Sciences. 160 fSuDP n. S134-6. George, L. K, & Gwyther, L. P. (1986). Caregiver Well-being: A multidimensional examination of family caregivers of demented adults. The Gerontological sodetv of American. 26 (3). 253-259. Gelinas, D. P., O'Connor, P., Miller, R. G. (1998). Quality of life for ventilator- dependent ALS patients and their caregivers. Journal of the Neurological Sciences. 160 fSuPDl. n. S134-6. Gibbons, S. T. (1999). Physical health risks of caregiving... Part 13-part series. Issues on Aging. 22 (1). 6-7. Gilbar, O. (1999). Gender as a predictor of burden and psychological distress of elderly husbands and wives of cancer patients. Psvcho-Oncologv. 8 (4'>. 287-294. Given, C. W., Collins, C. E., & Given, B. A. (1988). Sources of stress among families caring for relatives with Alzheimer's disease. Nursing Clinics of North America. 23 (n. 69-82. Given, B. A., & Given, C. W. (1998). Health promotion for family caregivers of chronically ill elders. Annual Review of Nursing Research. 16.197-217. Given, B., BCing, S., Collins, C., Given, C. (1988). Family caregivers of the elderly: ftivolvement and reacuons to care. Archives of Psychiatric Nursing. 2 (5). 281- 288. 261 Given, C. W., Given, B. A., Stommel, M., Collins, C., King, S., & Franklin, S. (1992). The caregiver reaction assessment (CRA) for caregivers to persons with chronic physical and mental impairments. Research in Nursing & health. 15.27 L-283. Given, B. A., Stommel, M., Collins, C., BCing, S., & Given, C. W. (1990). Responses of elderly spouse caregivers. Research in Nursing & Health. 13.77-85. Given, B. A., Stommel, M., Collins, C., King, S., & Given, C. W. (1990). Responses of elderly spouse caregivers. Research in Nursing & Health. 13.77-85. Gonzalez, E. W. (1997). Resourcefulness, appraisals, and coping efforts of family caregivers. Issues in Mental Health Nursing. 18 (31.209-27. Goldschmidt, W. (1971). The comparative study of values. In W. Goldschmidt (Ed.), Exploring the wavs of manidnd (2°'' ed.). New York: Holt, Rinehart & Winston, Inc. Gonzalez, E. W. (1997). Resourcefulness, appraisals, and coping efforts of family caregivers. Issues in Mental Health Nursing. 18 (31 209-27. Goldschnudt, W. (1971). The comparative study of values. In W. Goldschmidt (Ed.), Exploring the wavs of mankind (2°*^ ed.). New York: Holt, Rinehart & Winston, Inc. Goodman, C. C. (1990). The caregving roles of Asian American women. Journal of Women and Aging. 2.109-120. Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement. American Socioloical Review. 25.161-178. Grad, I., & Sainsbury, P. (1963). Mental illness and the family. The Lancet. 7. 544-547. Grady, KX, Jalowiec, A., & White-Williams, C. (1998). Patient compliance at one year and two years after heart transplantation. Journal of Heart Lung Transplantation. 17 (4). 383-394. Graham, H. (1983). Caring: labour of love. In J, Finch, and D. Groves ^ds.), A Labour of Love. Women (pp. 13-30). London: Routledge and Kegan Paul. 262 Grant, G., Ramcharan, P., McGrath, M., Nolan, M., & Keady, J. (1998). Rewards and gratifications among family caregivers: towards a refined model of caring and coping. Journal of Intellectual Disability Research. 42(1). 58-71. Griggs, M. J. (1995). Home care following hospitalization of elderly rural patients with chronic diseases: an evaluaa'on of changes in perceived health status on fluctuations in utilization and cost of formal and informal services. Unpublished doctoral dissertation, University of Maryland at Baltimore. Grinker, R. R., & Spiegel, J. P. (1945). Men under Stress. New York: McGraw- ffill. Guest, C. (1986). An investigation of the dimensions of burden in family caregivers of Alzheimer's disease patient. Unpublished thesis. University of Mindsor at Ontario. Haley, W. E. (1997). The family caregiver's role in Alzheimer's disease. Neurology. 48(5 Supol 61S25-29. Haley, W. E., Levine, E. G., Brown, S. L, & Bartolucci, A. A. (1987). Stress, appraisal, coping, and sodal support as predictors of adaptational outcome among dementia caregivers. Psychology & Aging. 2 (4). 323-30. Haley, W. E., Roth, D. L., Coleton, M. I., Ford, G. R., West, C. A., Collins, R. P., & Isobe, T. L. (1996). Appraisal, coping, and social support as mediators of well-being in black and white family caregivers of patients with Alzheimer's disease. Journal of Consulting & Clinical Psychology. 64 (1). 121-9. Haley, W.E.,&Pardo,K.M. (1989). Relationship of severity of dementia to caregiving stressors. Psychology & Aging. 4 (4). 389-92. Hann, N. (1977). Cooing and defending: Processes of self-environment organization. New York: Academic Press. Hawkins, B. (1996). Daughters and caregiving: taking care of our own. AAOHN Journal. 44 (9). 433-7. Henkle,E.J. (1994). Hardiness, burden, stress, appraisal, cooing and well-being of family caregivers of homebound older adults. Unpublished doctoral dissertation, Indiana University School of Nursing. Heok, K. E. & Li, T. S. (1997). Stress of caregivers of dementia patients in the Singapore Chinese family. Litemational Journal of Geriatric Psychiatry. 12 (4). 466-469. 263 Hibbard, J., Neufeld, A., Harrison, M. J. (1996), Gender differences in the support networks of caregivers. Journal of Gerontological Nursing. 22 (9). 15-23. EDnrichsen, G. A. (1991). Adjustment of caregivers to depressed older adults. Psvchologv & Aging. 6(4). 631-9. Ho, D. Y. F. (1996). Filial and its psychological consequence, bi M. H. Bond (Ed.), The Handbook of Chinese Psychology (p.155-165). Hong Kong: Oxford. University Press. Ho, D. Y. F. (1994). Filial Piety, authoritarian moralism, and cognitive and cognitive conservatism in Chinese society. Genetic. Social. & General Psvchologv Mongraphs. 120 93). 349-365. Holroyd, E. A., & Mackenzie, A. E. (1995). A review of the historical and social processes contributing to care and caregiving in Chinese families. Journal of Advanced Nursing. 22 (3). 473-479. Hooker, K., Frazier, L. D., & Monahan, D. J. (1994). Personality and coping among caregivers of spouses with dementia. Gerontologist. 34 (3). 386-92. Horowitz, A. (1985). Sons and Daughters as caregivers to older parents: Difference in role performance and consequences. The Gerontologist. 23 (6). 612-^17. Houde, S. C. (1998). Predictors of elders' and family caregivers' use of formal home services. Research in Nursing «fc Health. 21 (6). 533-543. Hull, M. M. Coping strategies of family caregivers of dementia-related behavior problems. Oncology Nursing Forum. 19 (2). 1179-1187. Hundley, N. (1974). The Asian American: The Historical Experience ^ds.). Santa Barbara, CA: American Bibhographical Center- Clio Press, Inc. Ignatius, P. (1987). An analvsis of relationships among interdependence in family caregivers and the elderly, caregiver burden, and adaptation of homebound frail elderly. Unpublished doctoral dissertation. University of Catholic, dissertation. Ikels,C. (1983). The process of caretaker selection. Research on Apng. 4.491- 509. Ditrieri RC. Rapp SR. (1994). Self-control skillfiilness and caregiver burden among help-seeking elders. Journal of Gerontology. 49 fl). 19-23. 264 Ip, G. S. H., &. Mackenzie, A. E. (1998). Caring for relatives with serious mental illness at home: the experiences of family carers in Hong Kong. Archives of Psvchiatric Nursing. 12 fS). 288-94. Irvin, B.L. (1993). Social support, self-worth and hone as self-care resources for coping with caregiver stress. Unpublished doctoral dissertation. University of Texas at Austin. Irvin, B., & Acton, G. (1996). Stress mediation in caregivers in of cognitively impaired adults: Theoretical model testing. Nursing research. 45(30.160-166. Irvin, S. H., Carroll, W. K. (1980). Testing and assessment across cultures: Issues in methodology and theory. In H. D. Triandis & J. W. Berry (Eds.), Handbook of Cross-cultural Psvchologv: Methodologv (Vol. 2). pp.181-244. Boston: Allyn & Bacon. Jalowiec, A. N. (1988). Confirmatory factor analysis of the Jalowiec Coping Scale, C. Waltz & O. Strickland (Eds.), Measurement of Nursing Outcomes: Client Outcomes (PP. 287-308). New York: Springer Publishing Campany. Jalowiec, A., Murphy, S. P., & Powers, M. J. (1984). Psychometric assessment of the Jalowiec coping scale. Nursing Research. 33 G). 157-161. Jalowiec, A. N. (1979). Stress and Cooing in Hypertensive and Emerpencv Room Patients. Unpublished thesis. University of Illinois. Jalowiec, A. N., Murphy, S. P., & Powers, M. J. (1983). Psychometric assessment of the Jalowiec Coping Scale. Nursing Research. 33 (3). 157-161. Jivanjee,P. (1994). Enhancing the well-being of family caregivers to patients with Alzheimer's disease. Journal of Gerontological Sodal Work. 23 (1/2). 31-48. Jones, E. G., & Kay, M. (1992). Methodology concern: Distrumentation in cross- cultural research. Nursings Research. 41 (3"). 186-188. Jones, E. G. (1987). Translation of quantitative measures for use incross-cultural research. Nursing Research. 36 (5). 324-327 Jones, P. S. (1996). Asian American women caring for elderly parents. Journal of Family Nursing. 2(1). 56-75. Jones, P. S. (1995). Paying respect: care of elderly parents by Chinese and Rlipino American women. Health Care for Women fatemational. 16 (5). 385-398. 265 Kao, S., Lu, Z., Yeh, S., & lie, H. (1999). The relationship between family function, social support and caregiver burden for dependent elderly. Nursing Research (China) 7(2). L72-182. Kane,R.L. (1997). Understanding Health Care Outcome Research. Gaithersbury, MD: Aspen. Kenny, D. (1979). Correlation and causality. New York, Hohn Wiley & Sons. ECilleen, M. (1990). The influence of stress and coping on farruly caregivers' perceptions of health, hitemational Journal of Aging & Human Development. 30 (3). 197-211. King, A.Y.C., & Bond, M. H. (1985). The Confucian paradigm of man: A sociological view. In W. S., Tseng & D.Yil., Wu (Eds.), Chinese culture and mental health, pp. 29-46. New York, Academic Press. Kirkley, A., Griffin, S., McLintock, H., & Ng, L. (1998). The development and evaluation of a disease-specific quality of life measurement tool for shoulder instability: the Western Ontario Shoulder Distability Index (WOSI). American Journal of Sports Medidne. 26 (61.764-72. Klein, S. (1989). Caregiver burden and moral development. Image: Journal of Nursing Scholarship. 21 (2). 94-97. Kling, K. C., Seltzer, M. M., & Ryff, C. D. (1997). Distinctive late-life challenges: implications for coping and well-being. Psychology & Aging.l2 (2). 288-95. Knight, R. G., Devereux, R. C., & Godfrey, H. P. (1997). Psychosocial consequences of caring for a spouse with multiple sclerosis. Journal of Clinical & Experimental Neuropsychology. 19 (1). 7-19. Km"ght,B.G., Silyerstein,M.,McCallum,T.J., &Fox,L.S. (20(X)). A sociocultural stress and coping model for mental health outcomes among African American caregivers in southern California. Journal of Gerontology Series B- Psvchological Science & Social Sciences. 55B (3). 142-150. Kosberg, J., cairl, R. E., & Miller, D. M. (1990). Components of burden: Interventive imphcations. The Gerontologist. 30 (2). 236-242. Kosberg, J., & Cairl, R. (1986). The cost of care index: A case management tool for screening informal care providers. The GerontologisL 26 (3). 237-278. 266 Kramer, B. J. (1997). Differential predictors of strain and gain among husbands caring for wives with dementia. Gerontoloeist. 37 (2). 239-49. Land, O. (1946). Chinese family and society. New Haven, CT: Yale University Press. Lang, A., Brody, E. (1983). Characteristics of middle-aged daughters and help to their elderly mothers. Journal of Marriage and the family. 45.193-202. Laramie, P. C. and Wheatland, W. Y. (1998). Rural stroke caregivers: a qualitative study of the positive and negative response to the caregiver role. Topics in Stroke Rehabilitation. 5(3). 51-68. Lawrence, R. H., Tennstedt, S. L., & Almy, S. L. (1997). Subject-caregiyer response comparability on global health and functional status measures for African American, Puerto Rican, and Caucasian elders and their primary caregivers. Journals of Gerontology Series B-Psvchological Sciences & Social Sciences. 52Bf2). S103-11. Lawton, M. P., BCIeban, M. H., Moss, M., Rovine, M., & Clicksman, A. (1989). Measuring caregiving appraisal. Journal of Gerontology. 44 (3). 61-71. Lazrus, R.S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer. Lazarus, R.S. (1968). Emotions and adaptation: Conceptual and empirical relations. In Arnold, WJ (Ed.), Nebraska symposium on motivation. Lincoln: University of Nebraska Press. Leavitt, M., Martinson, L M., Liu, C. Y., Armstrong, V., Homberger, L., Zhang, J. Q., & Han, X. P. (1999). Common themes and ethnic differences in family caregiving the first year after diagnosis of childhood cancen Part H. Journal of Pediatric Nursing. 14 (2L. 110-22. Lee, H. S., Brennan, P. F., & Daly, B. J. (2(X)L). Relationship of empathy to appraisal, depression, life satisfaction, and physical health in informal caregivers of older adults. Research in Nursing & Health. 24 fl). 44-56. Lee, H., Kim, S., & You, K-S. (1997). Learning from other lands: Caring for elderly demented Koreans. Journal of Gerontological Nursing. 23 (9). 21-31. Lee, M. L, Lin, H. S., & Chang, M. C. (1995). Living arrangements of the elderly in Taiwan: Qualitative evidence. Journal of Cross-Cultural Gerontology. 10. 53- 78. 267 Lee, S., & Wu, S. (1998). Determinants of burden and depression among family caregivers. Nursing Research (China). 6(1). 57-68. Levine, N., Dastoor, D., & Grendron, C. (1983). Coping with dementia: A. pilot study. Journal of the American Geriatrics Society. 31 (I). 12-18. Li, L. W., Seltzer, M. M., & Greenberg, J. S. (1999). Change in depressive symptoms among daughter caregivers: an 18-month longitudinal study. Psychology & Aging. 14(2). 206-19. Lim, Y. M., Luna, I, Cromwell, S. L., Phillips, L. R., Russell, C. K., & Ardon, E. T. (1996). Toward a cross-Cultural undersanding of family caregiving burden. Western Journal of Nursing Research. 18 (3\ 256-266. Lin, M., Ku, N., Leu, J., Chen, J., & Lin, L. (1996). An exploration of the stress aspects, coping behaviors, health status and related aspects in family caregivers of hepatoma patients. Nursing Research (China). 4(2). 171-85. Liu, H. E. (1993). An exploration of the psychometrics of the quality of life index of Chinese family [Chinese]. Nursing Research (China).1(2). 127-36. Liu, H. (1991). Factors related to the quality of life of Chinese caregivers. Unpublished doctoral dissertation. University of Illinois at Chicago, Health Science Center. Lundh, U. (1999). Clinical. Family carers 2: Sources of satisfaction among Swedish carers. British Journal of Nursing. 8 (10). 647-652. Lynch-Sauer, J. (1990). When a family member has Alzheimer's disease: a phenomenological description of caregiving. Journal of Gerontological Nursing. 16 (9). 8-11. Mackenzie, A., Lee, D. T. F., Dudley-Brown, S., & Chin, T. M. (1998). Case management in Hong Kong: Evaluation of a pilot project in community nursing. Journal of Clinical Nursing. 7 (3). 291-291. Mackenzie, A., &. Holroyd, E. (1994). Study of caring and caregiving in Chinese families in Hong Kong... based on a paper presented at the First South-E^t Asian Nursing Research Conference organized by the Hong Kong Polytechnic, 5-7,1994. Hong Kong Nursing Journal. 68.8-11. 268 Mackenzie, A. E., & Holroyd, E. E. (1996). An exploration of the carers' perceptions of caregiving and caring responsibilities in Chinese families. International Journal of Nursing Studies. 33(1). 1-12. Maciejewsid, M. (1997). Generic Measures, R. L., Kane, Understanding Health Care Outcomes Research (Ed.). Gaithersburg, MD: An Aspen Pubh'cation. Mancini, J. A. (1989). Aping Parents and Adult Children. Massachusetts: Lexington Books. Maserang,J. W.S. (1992). A study of cooing strategies and feelings of burden among adults who are primary carepiyers for their chronically ill parents. Unpublished doctoral dissertation. Saint Louis Uniyersity. Martinson, L M., Caspers, G., & Muwaswes, M. (1998). Impact oyer time of Alzheimer's disease on the indiyidual and family members. Geriaction. 16 (2). 10-14. Martinson, L M., Chesla, C., & Muwaswes, M. (1993). Caregiying demands of patients with Alzheimer's disease. Journal of Community Health Nursing. 10 f4). 225-32. Martinson, I. M., Leavitt, M., Liu, C.Y., Armstrong, V., Homberger, L, Zhang, J. Q., & Han, X. P. (1999). Comparison of Chinese and Caucasian families caregiying to children with cancer at home: Part L Journal of Pediatric Nursing. 14 (2). 99-109. McCann,J.J. (1991). Effects of stress on spouse caregivers' psychological health and cellular immunity. Unpublished doctoral dissertation. Rush University, College of Nursing. McCarty, E. F. (1996). Caring for a parent with Alzheimer's disease: process of daughter caregiver stress. Journal of Advanced Nursing. 23 r4). 792-803. McCarty, E. F. (1993). Daughter caregiver stress assodated with the care of a parent with Alzheimer's Disease. Boston College. McCre,R. R. (1984). Situational determinants ofcoping responses: Loss, threat, and challenge. Journal of Personality and Social Psychology. 46(4). 919-929. McDonald, P£. (1995). Perceived health status and health promotion behaviors in Black and White informal caregivers of impaired elders. Unpublished doctoral dissertation. Case Western Reserve University. 269 McHorney, C. A., Ware, L E., Lu, L E, R., & Sherboume, C. D. (1994), The MOSt 36-item Short Form Health Survey (SF-36): HI tests of data quality, scaling assumptions and reliability across diverse patient groups. Medical Care. 32 (4). 40-66. McKee, K. J., Whittick, J. E., Gilhooly, M, M. L., Ballinger, B. R,, Gordon, D. S., Mutch, W. J., & Philp, I. (1999). The willingness to continue caring in family supporters of older people. Health & Social Care in the Community. 7(2). l(X)-108, Meleis,A.I. (1997). Theoretical Nursing: Development Sc Progress (3 rd.). New York: J. B. Lippincott. Miaskowski, C., Kragness, L., Dibble, S., & Wallhagen, M. (1997). Differences in mood states, health status, and caregiver strain between family caregivers of oncology outpatients with and without cancer-related pain. Journal of Pain &Svmptom Management. 13(3). 138-47. Milne, D. (1992), Assessment: A mental Helth Profile. Windsor Nfer-Nelson. Mishel, M., & Sotenson, D. (L983). Revision ofthe Ways of Coping checklist for a clinical population. Western Journal of Nursing Research. 15 (1). 59-76. Mixoguchi, R., lijma, S., Niino, N., & Orimo, H. (1995). Reliability and Validity of a Japanese version of the Cost of Care Index (Janpense). Nippo Ronen Igakka. Aasshi- Jaoanese Joumal of Geriatrics. 32 (6). 403-409. Mizoguchi, T., Tijima, S., Niino, N., Orimo, H. (1995). Reliability and validity of a Japanese version of the cost of care index. Nippon Ronen igakkai Zasshi - Japanese Joumal of Geriatric. 32(6). 403-409. Montgomery, R. J., Gonyea, J. G., & Hooyman, N. R. (1985). Caregiving and the experience of subjective and objective burden. Familv Relations. 34.19-26. Morris, R. G., Woods R. T., Davies, K. S., 8c Morris, L. W. (1991). Gender differences in carers of dementia sufferers. British Joumal of Psychiatry-Supplement. [101,69-74. Morrissey, E., Becker, J., & Rubert, M. P. (1990). Coping resources and depression in the caregiving spouses of Alzheimer patients. British Joumal of Medical Psychology. 63 f Pt 2). 161-71. Morse, S. R., & Fife, B. (1998). Coping with a partner's cancen Adjustment at four stages of the illness trajectory. Oncology Nursing Forum. 25 (4). 751-760. 270 Morycz, R. K., Malloy, J., Bozich, M., & Maitz, P. (1987). Radal differences in family burden: Clinical implication for social work» Journal of Gerontological Social Work. 10.133-154. Moos, R. H. (1995). Development and application of new measure of life stressor, social resources, and coping responses. European Journal of Psychological Assessment. 11 fI). I-I3. Moos, R. H. (1988). Life stressor and coping resources influence health and well-being. Evaluation Psicoloeica. 4 (2). 133-158. Moos, R. H. (1990). Coping Responses Inventory Manual. Califonua: Center for Health Care Evaluation, Stanford University. Moos, R. H., Brennan, Penny, L., Fondacro, M. R., & Bemice, S. (1990). Approach and avoidance coping responses among older problem and non-problem drinkers. Psychology & Aging. 5 (1). 31-40. Moss, M. S., Moss, S. Z., & Moles, E. L. (L985). The quality of relationships between elderly parents and their out-of-town children. Gerontoloeist. 25 (2). 134-140. Mui, A. C. (1995). Perceived health and functional status among spouse caregivers of frail older persons. Journal of Aging & Health. 7(2). 283-300. Mui, A C, (1992). Caregiver strain among Black and White daugter caregivers: A role theory perspective. The Gerontologist. 32.203-212. Munro, B., H., & page, E. B. (1993). Statistical Methods for Health Care Research. Philadephia: J. B. Lippincott Company. Murray, J., Schneider, J., Baneqee, S., & Mann, A. (1999). EUROCARE: a cross-national study of co-resident spouse carers for people with Alzheimer's disease: 11— A qualitative analysis of the experience of caregiving. Eitemational Journal of Geriatric Psychiatry. 14 (8). 662-667. Mutch, W. J., & Philp, L (1999). The willingness to continue caring in fanuly supporters of older people. Health & Social Care in the Community. 7(2). 100-108. Nagatomo, I., Akasaki, Y., Uchida, M., Tonunaga, M., Hashiguchi, W., & Takigawa, M. (1999). Gencter of demented patients and specific family relationship of caregivers to patients influence mental fatigue and burdens on relative as caregivers. htemational Journal of Geriatric Psychiatry. 14 (8). 618-625. Ill Neumann, P. J., Kuntz, K. M., Leon, J., Araki, S. S., Hermann, EL C., Hsu, M. A., & Weinstein, M. C. (1999). Health utilities and health status in Alzheimer's disease: a cross-sectionai study of subjects and caregivers. Abstract Book/Association for Health Services Research. 15.1999 Neundorfer, M. M. (1991). Coping and health outcomes in spouse caregivers of persons with dementia. Nursing Research. 40(5). 260-5. Neundorfer, M. M. (1989). The effect of coping on health outcomes in caregiving spouses of dementia patients. Case Western Reserve University. Ngan, R., & Cheng, I. C. K. (1992). The caring dilemma: Stress and needs of caregivers for the Chinese frail elderly. Hong Kong Journal of Gerontology. 6 (2). 34-41. Ngan, R., & Wong, W. (1995). Injustice in family care of the Chinese elderly in Hong Kong. Journal of Aging «fe Social Policy. 7(2). 77-94. Nolan, M,& Grant, G. (1989). Addressing the needs of informal carers: A neglected area of nursing practice. Journal of Advanced Nursing. 14.950-961. Novack, T. A., Bergquist, T. F., Bennett, G., & Gouvier, W. D. (1991). Primary caregiver distress following severe head injury. Journal of Head Trauma Rehabilitation. 6 {£^69-77. Nunnally, J. C. (1978). Psychometric Theory. New York: McGraw-Hill Book Company. Nydegger, C. N. (1985). Asymmetrical kin and the problematic son-in-law. In N. Datan & L. A. Greene (Eds.), Life-span Development Psychology: International Relations, pp. 99-123. Nydegger, C. N. (1985). Family ties of the aged in cross-cultural perspective. In B. B. Hess & W. E. Markson, (3 rd). Growing old in America: New Perspoectives on Old Age, pp. 71-85. New Brunswick, NJ: Transaction Publishers. O'Brien, M. T. (1993). Multiple sclerosis: stressors and coping strategies in spousal caregivers. Journal of Community Health Nursing. 10 (3). 123-35. Olson, P. (1993). Caregiving and long-term health care in the People's RepubUc of China. Journal of Aging & Social Policy. 5 (1-2). 91-110. Parks, A. H. & Pilisuk, M. (1991). Caregiver burden: Gender and the psychological costs of caregiving. American Journal of Orthopsychiatry. 61 (4). 501-509. 272 Patrick, D. L., Bush, J. W., & Chen, M. M. (1973). Methods for measuring levels of well-being for a health status index. Health Services Research. 8 (3). 228-245. Patrick, J. H.,& Hayden, L M. (1999). Neurotidsm, coping strategies, and negative well-being among caregivers. Psychology & Aging. 14 (2). 273-83. Patterson, T. L., Semple, S. J., Shaw, W. S., Yu, E., He, Y., Zhang, M. Y., & Wu, W. (1998). The cultural context of caregiving: a comparison of Alzheimer's caregivers in Shanghai, China and San Diego, California. Psychological Medicine.28 1071-84. Patterson, T. L., Semple, S. J., Temoshok, L. R., Atkinson, J. H., McCutehan, J. A., Straits-Troster, K. A., Chandler, J. L, & Grant, I. (1995). Stress and depressive symptoms prospectively predict immune change among HIV-seropositive men. Psychiatry Interpersonal and Biological Processes. 58.315-328. Pearlin, L. I., Mullan, J. T., Semple, S. J., Skaff, M. M. (1990). Caregiving and the stress process: An overview of concepts and their measures. The Gerontologist. 30. 583-594. Pearlin, L. I., Turner, H, & Semple, S. (1989). Coping and the mediation of caregiver stress. In E. Light & B. Lebowitz (Eds.), Alzheimer's distress treatment and family stress: Directions for Research, pp. 198-217. Rockville, MD: National Dnstitute of Mental Helath. Pearlin, L. L, & Zarit, S. H. (1993). Research into informal caregiving: Current perspectives and future directions. S. H, Zarit, L. L, Pearlin, & K. W., Schaie, Caregiving systems: Formal and Informal helpers. Peck, S. (1990). Forward. Pi H. Lewis. A question of values. SanFrandsco: Harper & Row. Pepin, J. L (1992). Family caring and caring in nursing, hnage: Joumal of Nursing Scholarship. 24 (2). 172-130. Periard, M. E., & Ames, B. D. (1993). Lifestyle changes and coping patterns among caregivers of stroke survivors. Public Health Nursing. 10 (4). 252-6. Pett, M.A., Caserta, M.S., Hutton, AJ., & Lund, D.A. (1988). bitergenerational conflict: middle-aged women caring for demented older relatives. American Joumal of Orthopsychiatry. 58 (3). 405-17. 273 Phillips, D. (1988). Accommodation for elderly persons in newly industricalised countries: Hong Kong experience. International Journal of Health Science. 18 (2). 255- 279. Phillips, L. R. (1986). Caring for the frail elderly at home: Toward a theoretical explanation of the dynanucs of poor quality family caregiving. Advances in Nursing Science. 4 (40.62-84. Philllips, L. R. (1988). The fit of elder abuse with the family violence paradigm, and the implications of a paradigm shift for clinical practive. Public Health Nursing. 5 £^222-229. Phillips, L. R., Luna, I., Russell, C. K., Lim, Y. M., Cromwell, S. L., & Torres de Ardon,T. (1996). Toward a cross-cultural perspective of family caregving. Western Journal of Nursing Research. 18 (3). 236-251. Phillips, L. R., Luna, L, & Torres de Ardon, E. (1994). Strategies for achieving cultural equivalence. Research in Nursing & Health. 17.149-154. Phillips, K. D., Thomas, S. P. (1996). Extrapunitive and intropunitive anger of HIV caregivers: nursing implications. Journal of the Association of Nurses in AIDS Care. 7 (2). 17-27. Picot, S.JJF. (1995). Choice and social exchange theory and the rewards of African American caregiver. Journal of National Black Nurse Association. 7 (2). 29-40. Picot, S.Ji^. (1995). Rewards,costs,andcopingof African American caregivers. Nursing Research. 44(3). 147-152. Picot, SJi^. (1991). The relationship between the rewards, costs, and coping strategies of Black family caregivers. Unpublished doctoral dissertation. University of Maryland at Baltimore. Picot, S. J. F., Youngblut, J., & Teller, R. (1997). Development and testing of a measure of perceived caregiver rewards in adults. Journal of Nursing Measurement, 5 0133-52. Pohl, J. M., Given, C. W., Collins, C. E., & Given, B. A. (1994). Sodal vulnerability and reactions to caregiving in daughters and daughters-in-law caring for disabled aging parents. Health Care for Women fotemational. 15 (5). 385-395. Polit, D. R, & Hungler, B. P. (1995). Nursing Research: Principles and Methods. Philadelphia, PA: J. B. Lippincott Company. 274 Polil, D. F. (1996). Data analysts & statistics for nursing research. Philadelphia, PA: L B. Lippincott Company. Pouishock, S.W. & Deimling, G. T. (1984). Families caring for elders in residence: Issues in the measurement of burden. Journal of Gerontology. 39(2). 230-239. Pratt, C., Wright, S., & Schmall, V. (1987). Burden, coping and health status: A comparison of family caregivers to community dwelling and institutionalized Alzheimer's patients. Journal of Gerontological Social Work. 10(1-2). 99-112. Pratt, C. C., Schmall, V. L„ Wright, S., & Clenland, M. (1985). Burden and coping strategies of careigvers to Alzheimer's patients. Family Relations. 34.27-33. Pratt, C. C., Wright, S., & Schmall, V. L. (1987). Pnichno, R., & Kleban, M. H. (1993). Caring for an institutionalized parent: the role of coping strategies. Psychology & Aging. 8 (1). 18-25. Pruchno, R. A., & Resch, N. L. (1989). Mental health of caregiving spouses: coping as mediator, moderator, or main effect? Psychology & Aging. 4 (4). 454-63. Quayhagen, M. P., Quayhagen, M. (1996). Discovering life quality in coping with dementia. Western Journal of Nursing Research. 18 (2). 120-35. Qureshi, H. (1986). Responsestodependency:Redprocity, affect, and poser in family relationship in old age. In G. Phillipson, M. Bernard, & R. Strange (Eds.). Dependency and interdependencv in old age: Theoretical Pespective in Policy Alternatives. London, Groom Helm. Qureshi, H., & Perring, A. (1989). The caring relationship between elderly people and their families. McMillian Basingstoke. Rapp, S. R., Shumaker, S., Schmidt, S., Naughton, M., Anderson, R. (1998). Social resourcefulness: its relationship to social support and wellbeing among caregivers of dementia victims. Aging & Mental Health. 2f 1). 40-8. Rappetoe, P. A., & Rogers, R. W. (1987). Effects of components of protection- motivation theory oa adaptive and maladaptive coping with a health threat. Journal of Personality and Social Psychology. 52 (3). 596-604. 275 Raveis, V. H., Karus, D., & Pretter, S. (1999). Correlates of anxiety among adult daughter caregivers to a parents with caner. Journal of Psychosocial Oncology. 17 (3). 1- 26. Ray, L. D., & Ritchie, J. A. (1993). Caringforchronically ill children at home: factors that influence parents' coping. Journal of Pediatric Nursing: Nursing Care of Children & Families. 8 (4). 217-25. Richardson, R. C., Sistler, A. B. (1999). The well-being of elderly Black caregivers and noncaregivers: a preliminary study. Journal of Gerontological Sodal Work. 1(1/21. 109-17. Richard, Holicky, M. A. (1996). Caring for the caregivers: The hidden victims of Illness and Disability. Rehabilitation Nursing, 21 (5), 247-252. Riedel, S. E., Fredman, L., & Langenberg, P. (1998). Associations among caregiving difficulties, burden, and rewards in caregivers to older post-rehabilitation patients. Journals of Gerontology. 53 (3). 165-174. Rex, B. BC. (1998). Principles and practice of structural equation modeling. New York: The guilford press. Roberts, J., Browne, G., Milne, C., Spooner, L., Gafni, A., Drummond-Young, M., LeGris, J., Watt, S., LeClair, K., Beaumont L., & Roberts, J. (1999). Problem- solving counseling for caregivers of the cognitively impaired: effective for whom? Nursing Research. 48 (3). 162-72. Robinson, B. C. (1983). Validation of a caregiver strain index. Journal of Gerontology. 38.344-348. Robinson, B. C., Thumher, M. (1979). Taking care of aged parents: A family cycle transition. Gerontologist. 19.586-593. Robinson, K. M. (1990a). Predictors of burden among wife caregivers. Scholarly inquiry for Nursing Practice: An international Journal. 4 (3). 189-203. Robinson, K. M. (1997). Family caregiving: Who provides the care and at what cost? Nursing Economics. 15 (5). 243-247. Robssi, A.,&Robssi,P. (1990). Of human bonding: Parent-child relations across the life course. New York: A. de Gryter. 276 Rook, K. (1987), Reciprocity of social exchange and social satisfaction among older women. Journal of Personality and Social Psychology. 52.145-154. Rosen, E. J. (1988). Ethnic and cultural dimensions of work with hospice families. The American Journal of Hospice Care. 4.16-21. Rosow, I., & Breslau, N. (1966). A Guttman health scale for the aged. Journal of Gerontology. 27 fl). 91-94. Rothbaum, E, & Xu, X. (1995). The theme of giying back to parents in Chinese and American songs. Journal of Cross-Cultural Psychology. 26 (6). 698-713. Rusinak, R. L., & Murphy, J. F. (1995). Elderly spousal caregiyers: knowledge of cancer care, perceptions of preparedness, and coping strategies. Journal of Gerontological Nursing. 21 (3). 33-41. Russell, C. A. (1991). Metropolitan yersus nonmetropolitan caregiyers of persons with dementia. Rush Uniyersity, College of Nursing. Sander, A. M., High, WM. J., Hannay, H. J., & Sherer, M. (1997). Predictors of psychological health in caregiyers of patients with closed head injury. Brain Injury. 11 1^235-249. Shaw, W.S., Patterson, T. L, Sempte, S. J., Grant, I., Yu, E. S., Zhang, M., He, Y. Y., &Wu, W.Y. (1997). A cross-cultural yalidation of coping strategies and their associations with carcgiying distress. Gerontologist. 37 (4). 490-504. Sayles-Cross, S. (1993). Perceptions of famiUal caregiyers of elder adults. Image - the Journal of Nursing Scholarship. 25 (2). 88-92. Sayles-Cross, S. (1989). Appraisal, social distance and the informal caregiyer's multidimensional cost of caring for an elder family member. Unpublished doctoral dissertation, Uniyersity of Texas at Austin. Schorr, A. L. (1960). Filial responsibility expectations and moral among aged parents. The Gerontologist. 17.492-299. Schumacker, K. L. (1995). Family caregiying role acquisition: Role-making through situated interaction. Image: Scholarship Inquiry for Nursing Practice. 9 (31.211- 226. Schumacher, K. L., Dodd, M. J., & Paul, S. M. (1993). The stress process in family caregiyers of persons receiying chemotherapy. Research in Nursing & Health. 16(6). 395-404. 277 Scott, D., Oberst, M., & Dropldn, M. (1982). A stress-coping model. InU.D. Sytterket & F. Dibbekkt (Eds), Coping with Stress: A nursing perspective. Maryland: Aspen. Scott, J. P., & Roberto, KL (1985). Use of informal ad formal support networks by rural elderly poor. TheGerontoIogist. 25 (6). 624-630. Scott, J. P., Roberto, K. A., & Hutton, J. T. (1986). Families of Alzheimer's victims. Family support to the caregivers. Journal of the American Geriatrics Society. 34 (^348-54. Seelbach, W. C., & Sauer, W. I. (1977). Filial responsibility expectations and morale among aged parents. Gerontoloeist. 17(6). 492-499. Segall, M., & Wykle, M. (1988). The black family's experience with dementia. Journal of Applied Social Sciences. 13 (1). 170-191. Seltzer, M. M. & Li, L. W. (1996). The transitions of caregiving: Subjective and objective definitions. The Gerontologist. 36(5), 614-626. Sevick, M. A., Sereika, S., Hoffman, L. A., Matthews, J. T., & Chen, G. J. (1997). A confirmatory factor analsysis of the Caregiving Appraisal Scale for caregivers of home-based ventilator-assisted individuals. Heart & Lung. 26 (6). 430-438. Sexton, D. L., Stephanie, R. N., Calcasola, S. L., Bottomley, S. R., & Funk, M. (1999). Adults' experience with asthma and their reported uncertainty and coping strategies. Clinical Nurse Specialist.13 (1). 8-14. Shanas, E. (1979). The family as a social support system in old age. The Gerontologist. 19.169-174. Shaw, W. S., Patterson, T. L., Semple, S. J., Grant, L, Yu, E. S., Shang, M., He, Y.Y., Wu, W. Y. (1997). A cross-cultural validation of coping strategies and their assodations with caregiving stress. The Gerontologist. 37 (4). 490-504. Shyu, Y. L., Archbold, P.G., Sc Inile, M. (1998). FincUng a balance point: a process central to understanding family caregiving ia Taiwanese families. Research in Nursing & Health. 21(3). 261-70. Shiau, S., Yang, C. Y., Yeh, Y. Y., Huang, J. Y., Chen, M. Y., & Wu, L. L. (1999). The study of family structure, illness symptom, and stress adaptation of psychotic patients. Kaouhsing Journal of Medical Sciences. 15(6). 359-371. 278 Sirapo-Ngam, Y. (1994). Stress, caregiving demands, and coping of spousal caregivers of Parkinson's patients. Unpublished doctoral dissertation. University of Alabama at Birmingham. Sisk,R.J. (2000). Caregivers burden and health promou'on. Ihtemational Journal of Nursing Studies. 37 fll 37-43. Sidle, A., Moos, R., Adams, J., & cady, P. (1969). Development of a coping scale. Achives of General Psvchiatrv. 20.226-232. Silliman, R. A., & Sternberg, J. (1988). Family caregiving: Impact of patient funcuoning and underlying causes of dependency. The Gerontoloeist. 45 f1). 25-29. Smith, C. E., Mayer, L. S., Perkins, S. B., Gerald, K., Pingleton, S. K. (1994). Caregiver learning needs and reactions to managing home mechanical ventilation. Heart & Lung: Journal of Critical Care. 23 f2). 157-63. Smyth, K. & Williams, P. (1991). Patterns of coping in black working women. Behavioral Medidne. 17 fl). 40-45. Smyth, K. & Yarandi, H. N. (1996). Factoranalysisofthe ways of coping questionnaire for African American women. Nursing Research. 45 f H. 25-29. Soldo, B. J., & Myllyluoma, J. (1983). Caregivers who live with dependent elderly. The Gerontologist. 23 r6). 605-611. Song, K. T. (1995). Measures and dimensions of filial piety in Korea. The Gerontologist. 35 (2). 240-247. Song, K.T. (1990). A new look at filial piety: Ideas and practices of family- centered parent care in Korea. The Gerontologist. 30 r5). 610-617. Sparks, M. B., Farran, C. J., Donner, E., & Keane-Hagerty, E. (1998). Wives, husbands, and daughters of dementia patients: predictors of caregivers' mental and physical health. Scholarly Inquiry for Nursing Practice. 12 f3). 221-38. Statistical Bulletin, (1997). A protrait of Asians and Pacific Islanders in the United States. Statistical Bulletin, (1996). Increasing diversity of the U.S. population. 279 Steele, R, G. & Fitch, M.1. (L996), Coping strategies of family caregivers of home hospice patients witli cancer. Oncology Nursing Forum. 23 fg). 955-960. Stephens, M. A., Norris, V. K., Kinney, J. M., Ritchie, S. W., & Grotz, R. C. (1989). Stressful situations in caregiving: relations between caregiver coping and well- being. Psychology & Aging. 3 (2). 208-9. Stephens, M.A., & Zarit, S. H. (1989). Family caregiving to dependent older adults: stress, appraisal, and coping. Psychology & Aging. 4 (4). 387-8. Stewart, A. L., Ware, J. E., & Brook, R. H. (1981). Advances in the measurement of functional status: construction of aggregate indexes. Medical Care.19 {5}^ 473-488. Stommel, M. Wang, S., Given, C. W., & Given, B. (1992). Focus on psychometrics confirmatory factor analysis as a method to assess measurement equivalence. Research in Nursing «fe Health. 15 (5). 399-405. Strong, C. (1984). Stress and caring for elderly relatives: Interpretations and coping strategies in American Indian and while sample. The Gerontologist. 24 (3). 251- 255. Suchman,E. A., Phillips, B.S.,&Streib,G.F. (1958). An analysis of the validity of health questionnaires. Social Forces. 36.223-232. Sullivan, D. F. (1966). Conceptual problems in developing an index of health. National Center for Health Statistics. Series 2. No 17. U.S. Department of Health, Education, & Welfare. Sashington. Sung, K. T. (1995). Measures and dimensions of filial piety in Korea. The Gerontologist. 35 (21240-247. Sung, K.T. (1992). Motivationsforparentscare:Thecaseoffnialchildrenin Korea. International Journal of Human Development. 34 f2\ 109-124. Szmukler, G. I., Burgess, P., Herrman, H., Benson, A., Colusa, S., & Bloch, S. (1996). Caring for relatives with serious mental illness: the development of the Experience of Caregiving Inventory. Social Psychiatry & Psychiatric Epidemiology. 31 (3-4). 137-48. Tabak, N., Ehrenfeld, M., & Alpert, R. (1997). Feelings of anger among caregivers of patients with Alzheimer's disease. International Journal of Nursing Practice fCarlton). 3(2184-8. 280 Taylor, R. (1985). The extend family as a source of support to elderly blacks. The Gerontoloeist. 25 (5). 488-495. Teel, C S., Duncan, P., & Lai, S. M. (2001). Caregiving experiences after stroke. Nursing Research. 50 fl). 53-60. Tiden, V. (1986). New perspectives on socical support. Nurse oractitiner. 11(8), 60,62. Tobin, D. L., Kohroyd, K. A., Reynolds, R. V., & Wigal, J. K. (1989). The hierarchical factor structure of the coping strategies inventory. Cognitive Therapv & Research. 13 r4). 343-361. Toseland, R. W., Blanchard, C. G., & McCallion, P. (1995). A problem solving intervention for caregivers of cancer patients. Social Science & Medicine. 40 (4). 517-28. Toseland, R. W., Labrecque, M. S., Goebel, S. T., & Whimey, M. H. (1993). An evaluation of a group program for spouses of firail elderly veterans. Gerontologist. 32 (3). 382-90. Turner, R. H. (1970). Familv Interaction. New York: John Wiley & Sone. Thompson, S. C., Medvene, L. J., & Freedman, D. (1995). Caregiving in the close relationships of cardiac patients: Exchange, power, and atributional perspectives on caregiver resentment. Personal Relau'onships. 2 (2). 125-142. Thompson, E. Jr., & Doll, W. (1982). The burden of families coping with the mentally ill: An invisible crisis. Familv Relations. 31(3). 379-388. Lr.S. Bureau of the Census (1991). Census Bureau releases 1990 Census counts on specific racial groups. Press release CB 91-215. Washington, DC. U.S. Bureau of the Census (1996). [On line]. Available: httD://^www.census.gove/poDulation/proiections/staie/vi96toQ0 U.S. Census Bureau. (2000). Demographic Profile: Census 2000. [Online]. Available: http://www.census.gov/Press-ReleaseAvww/2001/demoprofile.html VaIle,R. (1994). Culture-fair behavioral symptom differenaal assessment and intervention in dementing illness. Alzheimer Disease & Associated Disorders. 8 Suppl 3. 21-45. 281 Vandenneer, J. L. (1997). Personal and financial costs to family systems providing informal home care for rural elders with Alzheimer's disease. Unpubh'shed doctoral dissertation, University of Alabama at Birmingham. Vedhara, K., Shanks, N., Anderson, S., & Lightman, S. (2(XX)). The role of stressors and psychosocial variables in the stress process: A study of chronic caregiver stress. Psychosomatic medicine. 62 (3). 374-385. Vitaliano, P. P., Young, H. M., & Russo, J. (1991). Burden: A review of measures used among caregivers of individuals with dementia. The Gerontoloeist. 31 (U. 67-75. Vitaliano, P. P., Breen, A. R., Russo, J., Albert, M., Vitiello, M. V., & Prinz, P. N. (1984). The clinical utility of the dementia rating scale for assessing Alzheimer patients. Journal of Chronic Disease. 37 (9-10). 743-753. Wahl, A., Hanestad, B JR., Wiklund, L, & Moum, T. (1999). Coping and quality of life inpatients with psoriasis. Quality of Life Research. 8 (5). 427-433. Wallhagen, M. L (1988). Perceived control and adaptation in elderly caregivers. Unpublished doctoral dissertation. University of Washington. Waley, A. (1938). The Analects of Confucius. New York: Vintage Books. Walsh, M., Smith, R., Morales, A., & Sechrest, L. (1999). Ecocultural Research: A mental health research's guide to the study of race, ethnicity and culture (In press). Wang, Y., Shyu, Y., Yao. J., & Wu. S. (1998). A correlational study of "finding a balance point" in family caregiving process for frail elders [Chinese]. Nursing Research fChina). 6(g). 489-500. Ward,R. (1985). Mormal networks and well-being in later life: A research agenda. The Gerontologist. 25.55-81. Ware, I. E. (1976)> Scales for measuring general health perceptions. Health Services Research. 11(4). 396-415. Ware, I.E., &Karmos. (1976). Development and validation of scales to measure perceived health and patient role propensiQr. Final report. Springfield, VA: National Technical Information Service (Publication No. 288-331). Ware, J. E., Brook, R. H., Davies-Avery, A., Williams, K. N., Stewart, A. L. (1980). Conceptualization and measurement of health for adults in the health insurance 282 study, volumel, Model of Health and Methodology. Sponsored in part by Grant DHEW- 016B-7901-P202L Ware, J. E., Keller, S. Bentler, P. M., Sullivan, M., Brazier, J., & Gander, B. (1994). Comparisons of health status measurement models and the validitv of SF-36 In Great Britain. Sweden and the U.S.A. Ware, J. E, Kosinski, M. & Keller, S. D. A 12-item Short-Form Health Survev: Construction of scales and preliminarv. [On line] Available: http://www.sf- 36.com/general/sfl2.html Ware, J. E. Jr., & Sherboume, C. D. (1992). The MOS 36-item short-form health survey (SF-36). Conceptual framework and item selection. Medical Care. 30 (6). 473- 483. Ware, J. E., Snow, K. K., Kosinski, M., & Gandek, B. (L993). SF-36 Health Survev: Manual and interpretation guide. Boston, MA: The health institute. New Englan Medical Center. Wener, L., & Campbell, D. T. (1970). Translation working through interpreters and the problem of descentering. In Narrow & R. Cohen (Eds.), Handbook of methods in cultural anthropology fop. 398-420). Garden City, New York: Natural History Press. Whitlatch, A. M. (1993). The relationships of perceived stress, coping strategies, and specificallv religiositv on subiective well-being of familv caregivers for individuals affected bv Alzheimer's disease. Unpublished doctoral dissertation, Ohio State University. Wicks, M. L. N. (1992). Family health in chronic illness. Unpublished doctoral dissertation, Wayne State University. Wijeratne, C. (1997). Review: pathways to morbidity in carers of dementia sufferers. Ditemational Psvchogeriatrics. 9 Tl). 69-79. Williamson, G. M., & Schulz, R. (1993). Coping with specific stressors in Alzheimer's disease categiving. Gerontologist. 33 (6). 747-55. Wineman, N. M., Durand, E. J., & Steiner, R. P. (1994). A comparative analysis of coping behaviors in individuals with multiple sclerosis or a spinal cord injury. Research in Nursing & Health. 17 (3). 184-194. Winslow, M. N. B. (1993). Commentary on Suffering: a relatively unexplored phenomenon among family caregivers of non-institutionalized patients with cancer Nursing Scan in Oncology. 2 f 1). 3-4. 283 Wilson, H. S. (1989). Family caregiving for a relative with Alzheimer's dementia: coping with negative choices. Nursing Research. 38 f2\ 94-8. Wood, J. (1987). Coping with the absence of Perceived Control: Ethnic and Cultural Issues in Family Caregiving for Patients with Alzheimer's Disease. Unpublished doctoral dissertation, Virginia Commonwealth University at Richmond. Worcester, M. I. (1990). Family coping: caring for the elderly in home care. Home Health Care Services Quarterly. 1If 1-2). 121-85. Worcester, M. I., & Quayhagen, M. P. (1985). Correlation of caregiving satisfaction: Prerequisites to older home care. Research in Nursing & Health. 6 (2). 61- 67. World Health Orgam'zation. (1958). Constitution of the World Health Orgam'zation, Annex I, in the First Ten Years of the World Health Organization. Geneva, World Health Organization. Wright, L. K. (1994). AD spousal caregivers: longitudinal changes in health, depression, and coping. Journal of Gerontological Nursing. 20 f10). 33-45.48. Wright, L. K., Clipp, E. C., & George, L. K. (1993). Health consequences of caregiver stress. Medicine. Exereise. Nutrition, and Health. 2.181-195. Wu, S. C., Li, C. Y., & Chang, A. L. (1997). The influence of intergenerational exchange on nursing home admission in Taiwan. Journal of Cross-Cultural Gerontology. 12^ 163-174. Wu, Agnes Chew-Chung. (1993). A cross-cultural, cross-national study of the family caregiver burden of a severely mentally ill family member. Dissertation Abstracts International. 54 (3-B). Wu, S. C., Li, C. Y., & Chang, A. L. (1997). The influence of intergenerational exchange on nursing home admission in Taiwan. Journal of Cross-Cultural Gerontology. 12^ 163-174. Wykle, M., & Segal, M. (1991). A comparison of Blactc and White family caregivers experience with dementia. Journal of National Black Nurses Association. 5 0129-41. 284 Yates, M. E., Tennstedt, S., & Chang, B. (1999). Contributors to and mediators of psychological well-being for informal caregivers. Journals of Gerontology Series B- Psvchological Sciences & Social Sciences. 54Bfl). 12-22. Yang, K. S., Yeh, K. H., & Huang, L. (1988). A social attitudinal analysis of Chinese filial piety: Concept and assessment. Bulletin of the Institute of Ethnology. Academia Sinica. 65.117-227. Yeh, S. (1995). Caregiver burden and the changes in caregiying following nursing home placement in Taiwan. Unpublished doctoral dissertation. University of Utah. York, N. (1995). Coping with caregiving: supporting the informal caregiver. Caring. 14 (4). 44-7. Young, J. L. B. (1995). Caring for a seriously mentally ill adult family member cooing strengths and strategies of black family caregivers. Unpublished doctoral dissertation. University of Virginia. Young, J. J., & Gu, N. (1995). Demographic and socio-econonuc characteristics of elderly Asian and Pacific Island Americans. Seattle, WA: National Asian Pacific Center on Aging. Young, R. F. & FCahana, F. (1989). Specifying caregiver outcomes: Gender and relationship aspects of caregiving strain. Gerontologist. 29 (5). 660-666. Yu, E., Liu, W. T., Wang, Z. Y., Levy, P. S., Katzman, R. Zhang, M. Y., Qu, G. Y., & Chen, FP. (1993). Caregivers of the cognitively impaired and the disabled in Shanghai, China. Di S. H., Zarit, & L. L, Pearlin et al. (^.) et al. fl993). Caregiving systems: Informal and formal helpers. Social structure and aging, (pp. 5-30). ffillsdale, NI: Lawrence Erlbaum Assodates, Inc. Zabielski, M. (1984). Giving and receiving in the neomatemal period: A case of distributive inequity. Maternal-Child Nursing Journal. 13.19-46. Zanetti, O., Magni, E., Sandri, C., & Frisom", G. B. (1996). Detenm'nants of burden in an Italian sample of Alzheimer's patient caregiyers. Journal of Cross-Cultural Gerontology, llfl). 17-27. Zarit, S.U. 91989). Issues and direction in family interventions research. fiiE. Light & B. Lebowits ^ds.), Alzheimer's Disease Treatment and Family Stress: Directions for Research, pp.458-486. 285 Zarit, S. H, & Eggebeen, D. J. (L995). Parent-child relationships in adulthood and old age. In M. H. Borwkin Handbook of parenting. Vol 1: Children and Parenting, pp.119-140. Hillsdale, NJ: Lawrence Erlbaum Assodates Inc. Zarit, S. H., Reever, K. E., & Bach-Peterson, J. (1980). Relatives of the impaired elderly: correlates of feelings of burden. Gerontologist. 20 f6\ 649-655. Zarit,S.,Todd,P., &2^t,J. (1986). Subjective burden ofhusbands and wives as caregivers: A longitudinal study. The Gerontologist. 26 (3). 260-266. Zunzunegui, M. V., Beland, E, Llacer, A., Keller, I. (1999). Family, religion, and depressive symptoms in caregivers of disabled elderly. Journal of Epidemiology & Community Health. 53f6). 364-9.