Organizational Psychology Article Review
Organizational Psychology Review 2016, Vol. 6(3) 273–302 Viewing the interpersonal ª The Author(s) 2015 Reprints and permission: sagepub.co.uk/journalsPermissions.nav mistreatment literature DOI: 10.1177/2041386615607095 through a temporal lens opr.sagepub.com
Michael S. Cole Texas Christian University, USA
Abbie J. Shipp Texas Christian University, USA
Shannon G. Taylor University of Central Florida, USA
Abstract Given increasing awareness of time’s critical role, we assess the current position of time in the workplace mistreatment literature. Focusing on four mistreatment constructs (viz., abusive supervision, workplace bullying, workplace incivility, and social undermining) found in the orga- nizational psychology literature, our search revealed 266 studies that have empirically examined the consequences of these forms of interpersonal mistreatment. We examine and critique these studies, finding that with a few exceptions, most have failed to design and test theoretical rela- tionships in a manner consistent with construct definitions. As interpersonal mistreatment research has neglected the role of time, we conclude that the substantial number of existing studies offer limited insight into the true nature of mistreatment’s consequences over time. We go on to elaborate on the types of theoretical insights that might emerge when a temporal lens (objective time and/or subjective time) is adopted by mistreatment researchers.
Keywords Abusive supervision, bullying, incivility, interpersonal mistreatment, longitudinal, objective time, subjective time, temporal lens, time, undermining
Paper received 5 January 2015; revised version accepted 19 August 2015.
Corresponding author: Michael S. Cole, Department of Management, Entrepreneurship, and Leadership, Neeley School of Business, Texas Christian University, Fort Worth, TX 76129, USA. Email: [email protected]
Downloaded from opr.sagepub.com at TEXAS CHRISTIAN UNIV on July 15, 2016 274 Organizational Psychology Review 6(3)
Over the past two decades, research on the of temporality (e.g., treating a dynamic phe- topic of interpersonal mistreatment at work has nomenon as static), the most likely conse- increased exponentially. In doing so, researchers quences are weak hypotheses and ambiguous—if interested in studying workplace mistreatment not erroneous—results (Ployhart & Vandenberg, have proposed several concepts designed to tap 2010). Clearly, questionable inferences have the different aspects of this harmful work behavior, potential to denigrate the accumulated knowledge including abusive supervision (Tepper, 2000), in this important area of research. By the same bullying (Einarsen, 1999), incivility (Anders- token, the explicit consideration of time-related son & Pearson, 1999), and social undermining issues will result in better theory building and a (Duffy, Ganster, & Pagon, 2002). Abusive richer understanding of the phenomena of supervision, for example, refers to hostile interest (Mitchell & James, 2001). According verbal and nonverbal supervisory behaviors to George and Jones (2000), this is because (Tepper, 2000) and is estimated to cost U.S. constructs ‘‘exist in and through time; time is businesses (in terms of absenteeism, health intimately bound up with the content of human care costs, and lost productivity) more than experience [and thus] ... cannot be separated $20 billion annually (Tepper, 2007). Tracking from it’’ (p. 666). the adverse consequences of abusive super- With hundreds of published articles on vision, workplace incivility focuses on a per- interpersonal mistreatment at work, our main petrator’s behavior with an ambiguous intent purpose is to take stock of this literature’s to harm another (Andersson & Pearson, 1999) knowledge base by evaluating existing research and is believed to touch 98% of U.S. employees, studies through a temporal lens (Ancona, with an associated annual cost in the millions Goodman, Lawrence, & Tushman, 2001). We (Porath & Pearson, 2013). To be sure, the therefore provide a temporal review of the frequency with which individuals are being mistreatment literature. We begin with a brief mistreated at work, whether by supervisors or review of the conceptual definitions and theircoworkers,suggestsitisaserioussocial empirical findings associated with the four problem warranting continued scholarly atten- exemplar mistreatment constructs (viz., abu- tion (Hershcovis & Barling, 2010; Tepper, sive supervision, bullying, incivility, and social 2007). undermining).1 Next, we offer a short primer Given the importance of this research, we on time to delineate the important aspects of sought to appraise the current state of affairs a temporal lens that shape our review and by determining the extent to which the inter- critique of the mistreatment literature. Here, personal mistreatment literature has generated we address the question of ‘‘what is time,’’ theoretically substantive findings. Toward observing that time can be viewed in terms of this end, we consider that, despite meaningful objective (clock) time or subjective (psy- conceptual differences among the mistreat- chological) time. As we detail later in our ment constructs (Tepper & Henle, 2011), they review, this particular issue has salient impli- all share a key feature—namely, the mistreat- cations for how mistreatment researchers con- ment constructs are inherently dynamic and ceive, design, and conduct their studies. We thus change over time. As such, an empirical then summarize empirical studies that focus studyonabusivesupervisionorcoworker on interpersonal mistreatment’s downstream bullying that treats its focal construct as a consequences, with special consideration toward static input would not be testing theoretical the temporal issues raised in this review. As relationships in a manner prescribed by the a result of our findings, we articulate how construct’s definition. This is potentially pro- the addition of objective and subjective time blematic because when a study sidesteps issues can change the way mistreatment is studied,
Downloaded from opr.sagepub.com at TEXAS CHRISTIAN UNIV on July 15, 2016 Cole et al. 275 offering interesting new research questions. acts’’ (Hershcovis, 2011, p. 501). The sys- Finally, we discuss how incorporating time tematic and gradually evolving nature of bul- into one’s study requires careful attention lying is a notable hallmark of this mistreatment to research design, and in doing so, allows phenomenon. For example, according to Einar- for tests of new and interesting research sen (1999), one or two negative experiences questions. cannot be construed as bullying but, rather, repeated exposure on a weekly basis (for several months) is indicative of experienced bullying Overview of interpersonal (p. 16). A recent meta-analysis by Nielsen and mistreatment constructs Einarsen (2012) provides supportive evidence for the detrimental effects of workplace bullying. Abusive supervision It found that employees exposed to bullying are Tepper (2000) first introduced the concept more likely to experience adverse psychological of abusive supervision, which he defined as and job-related consequences, including mental ‘‘subordinates’ perceptions of the extent to and physical health problems, increased inten- which supervisors engage in the sustained tions to quit, reduced job satisfaction and com- display [emphasis added] of hostile verbal mitment, and more frequent absenteeism. and nonverbal behaviors, excluding physical contact’’ (p. 178). According to Tepper’s con- Workplace incivility ceptualization, then, a supervisor’s abusive acts must be recurrently experienced by subordinates. Concerns about civility can be traced as far This particular aspect of abusive supervision is back as the founding of the United States of noteworthy because it implies that supervisory America (Washington, 1888/1971). Despite its mistreatment must occur over time. Reviews of extensive history in social norms, the notion the literature have found that abusive supervision of incivility at work was first introduced by has been consistently linked to undesirable levels Andersson and Pearson (1999). They char- of subordinate job satisfaction, commitment, acterized workplace incivility as ‘‘low-intensity and psychological distress (Martinko, Harvey, deviant behavior with ambiguous intent to harm Brees, & Mackey, 2013; Tepper, 2007). They the target, in violation of workplace norms for also find that subordinates of abusive super- mutual respect’’ (p. 457). In contrast to the visors are more likely to respond with deviant other mistreatment constructs which explicitly retaliatory acts, fewer citizenship behaviors, incorporate time into their definitions, work- and less effort directed at job tasks (Martinko place incivility’s definition does not account for et al., 2013; Tepper, 2007). its temporal nature. And yet workplace incivi- lity does indeed consider the role of time and its implications. This stems in large part from Workplace bullying Andersson and Pearson’s theorizing, in which It proved difficult to determine who first they describe uncivil behavior as unfolding in a introduced the concept of bullying to the series of ‘‘tit-for-tat’’ exchanges that ‘‘spiral’’ organizational psychology literature, although or escalate over time. Andersson and Pearson some of the earliest research began appearing go on to explicitly assert that workplace inci- in the early 1990s (e.g., Einarsen, Raknes, & vility has the potential to substantively impact Matthiesen, 1994). When placed in the work targets’ attitudes towards work because the context, bullying is defined as ‘‘instances where effects associated with incivility’s low-intensity an employee is repeatedly and over a period of behaviors are likely to accumulate over time. In time [emphasis added] exposed to negative such a scenario, initial instances of incivility
Downloaded from opr.sagepub.com at TEXAS CHRISTIAN UNIV on July 15, 2016 276 Organizational Psychology Review 6(3) have little to no discernable effect on a target, connection, Duffy and her colleagues (2002; but eventually a subsequent incivility episode Duffy, Ganster, Shaw, Johnson, & Pagon, could create significant harm by constituting 2006) have found that social undermining is the proverbial straw that ‘‘breaks the camel’s associated with lower levels of job satisfaction back’’ (Andersson & Pearson, 1999, p. 462). A and commitment, and greater intentions to quit. number of studies have demonstrated that the Studies have also found that targets of social experience of incivility is associated with undermining experience increased levels of psy- diminished levels of job satisfaction and com- chological distress (e.g., Nahum-Shani, Hen- mitment (e.g., Cortina, Magley, Williams, & derson, Lim, & Vinokur, 2014), and are likely to Langhout, 2001; Taylor, Bedeian, & Kluemper, engage in retaliation (e.g., Duffy et al., 2002) as 2012), increased physiological and psycholo- well as poor job performance (e.g., Ng & Feld- gical forms of stress and withdrawal (e.g., Bunk man, 2012). & Magley, 2013; Salomon & Jagusztyn, 2008), and more frequent retaliatory acts (e.g., Penney & Spector, 2005; Taylor & Kluemper, 2012). A Issues related to time and use few studies have examined performance-related of a temporal lens in outcomes, with incivility experiences correlat- ing with fewer acts of citizenship (e.g., Taylor mistreatment research et al., 2012), decreased task performance Illustrated by the brief previous overview, it is (Porath & Erez, 2007), and increased absen- quite clear (to us) that temporal issues are an teeism (Sliter, Sliter, & Jex, 2011). integral aspect of all four mistreatment con- cepts. Such an observation mirrors the organi- zational psychology literature as a whole. It is Social undermining widely acknowledged, for example, that time is The social undermining construct was brought a critical component of most any topic dis- into the work domain by Duffy et al. (2002), cussed in organizational psychology and related though they acknowledge that the term ‘‘social management disciplines (e.g., Ancona et al., undermining’’ was first introduced by Vinokur 2001; Bluedorn, 2002; George & Jones, 2000; and van Ryn (1993). According to Duffy et al. McGrath & Kelly, 1986; Shipp & Fried, 2014; (2002), social undermining is best characterized Sonnentag, 2012). This is because we—as as ‘‘behavior intended to hinder, over time, organizational scientists—cannot fully under- [emphasis added] the ability to establish and stand why individuals behave as they do with- maintain positive interpersonal relationships, out considering the temporal context in which work-related success, and favorable reputa- they are embedded (Lewin, 1943; Murray, tion’’ (p. 332). Once again, the passage of time 1938). Seeing value in the adoption of this plays a pivotal role in delineating this phe- perspective, Ancona et al. (2001) were among nomenon’s episodic qualities. For example, the first to call for the use of a ‘‘temporal lens’’ undermining behaviors are assumed to ‘‘weaken when studying human behavior in the work its target gradually or by degrees’’ (Duffy, domain. According to these scholars, time ‘‘has Scott, Shaw, Tepper, & Aquino, 2012, p. 643) always been at the foundation of organization but only to the extent one’s efforts are under- theory’’ but it has only recently moved ‘‘from the mined on a recurring basis (Duffy et al., 2002). background to the foreground’’ (p. 512). Moreover, the consequences associated with We therefore draw from the work of Ancona successive undermining episodes are believed et al. (2001), among others (e.g., Fried & to be subtle but insidious because their effects Slowik, 2004; Roe, 2008; Shipp & Fried, 2014), on targets can add up over time. In this to frame our discussion of time-relevant issues
Downloaded from opr.sagepub.com at TEXAS CHRISTIAN UNIV on July 15, 2016 Cole et al. 277
Table 1. Temporal review of interpersonal mistreatment research.
k A-temporal Objective time Subjective time
Abusive supervision 124 120 (97%) 3 (2%) 1 (1%) Bullying 79 78 (99%) 1 (1%) 0 (0%) Incivility 50 42 (84%) 5 (10%) 3 (6%) Undermining 13 13 (100%) 0 (0%) 0 (0%) Total 266 253 (95%) 9 (3%) 4 (2%)
Note. k ¼ number of studies. as they pertain to mistreatment research. In sense of the present situation. Perhaps more doing so, we begin by introducing the two interestingly, the target’s present-day response predominant conceptualizations of time. The is shaped not only by remembered past experi- first refers to objective time, most commonly ences (as contrasted to current levels), but also represented by the passage of ‘‘clock time.’’ by his or her anticipated interactions with the Triggered from the physical universe (e.g., the alleged perpetrator. daily rising and setting of the sun), objective time implies that workplace events flow in an irreversible sequence from past to present to Current state of the existing literature future. Objective time is measured as linear, quantifiable units (e.g., each passing minute is Using these temporal conceptualizations as a the same as any other minute in time), which backdrop, we examined existing studies and leads to the passage of time as an absolute determined the extent to which prior research across all individuals and situations (Ancona on abusive supervision, bullying, incivility, and et al., 2001; Shipp & Fried, 2014). When undermining has accounted for time when incorporated into a study on interpersonal mis- testing theoretical predictions.2 For instance, a treatment, the ‘‘passage of time’’ is typically not study exploring the relationship between abusive the study’s primary focus but rather provides the supervision and subordinate job performance context needed to capture the inherent dyna- might use an experience sampling method to mism in a mistreatment construct or its effects. repeatedly assess both variables for an extended The second conceptualization refers to time period of time (e.g., over a period of days or as a subjective experience based on psy- weeks). Such a study would be considered as chological and sociological constructions (e.g., having adopted a temporal lens because it uses Rousseau & Fried, 2001; Shipp & Jansen, 2011). the objective passage of clock time as the That is, individuals are assumed to interpret medium through which dynamic relationships workplace events based on normative experi- are explored. In a similar vein, a study also ence (Fried & Slowik, 2004). Individuals’ ret- could be classified as temporal to the extent that rospections and anticipations provide a context it considers a target’s current mistreatment for their current experience, meaning that sub- perceptions as well as recollected past mis- jective time is not necessarily linear or quanti- treatment experiences or anticipations regard- fiable, but instead interpretive and malleable ing future incidents (i.e., a subjective time lens). (Shipp & Fried, 2014). When placed into the Our search of the four mistreatment constructs context of the mistreatment literature, a sub- yielded 545 samples (heretofore referred to as jective view of time assumes that a target of ‘‘studies’’) from 465 articles. Of these, 266 mistreatment will mentally ‘‘time travel’’ to studies met our inclusion criteria.3 Table 1 recall his or her past experiences when making summarizes our classification results.
Downloaded from opr.sagepub.com at TEXAS CHRISTIAN UNIV on July 15, 2016 278 Organizational Psychology Review 6(3)
Our review indicates that the mistreatment somewhat addressed subjective time. In three of literature has yet to adopt a temporal lens, with these studies (i.e., Cortina & Magley, 2009, 253 of the 266 studies (95%) neglecting both Studies 1, 2, and 3), the employee participants objective and subjective elements of time. We were asked to recall a particular mistreatment also noted that 190 of the 253 studies tested incident (i.e., retrospective mistreatment) to their hypotheses using a cross-sectional research assist them in narrowinginonpastreactionsto design whereas 63 studies employed a time-lagged a specific situation. Only Greenbaum, Hill, research design, in which the mistreatment Mawritz, and Quade (2014, Study 3) examined construct and outcome variable(s) were assessed how the recollection of past mistreatment at different times. Although a time-lagged design influenced current reactions. Finally, we may help to alleviate common-method variance observed that these four studies focused only concerns, it is only capable of capturing between- on retrospected mistreatment and neglected the unit differences (Ployhart & Vandenberg, 2010). potential impact of anticipatory mistreatment In other words, a time-lagged design is similar to a when asking employee participants to make cross-sectional design in that both place the mis- sense of their current situation. treatment variables and their outcomes in static In summary, a majority of the identified form (Ployhart & Vandenberg, 2010; Ployhart & studies on abusive supervision, bullying, work- Ward, 2011). place incivility, and social undermining employed Another interesting finding was that in over astaticresearchdesign.Consequently,itappears half (54%) of these studies (136 out of 253), the that these streams of research have largely authors were seemingly aware of the limitations failed to design and test conceptual schemes associated with their static research design in a manner consistent with the mistreatment given the repeated calls for future research to use constructs’ core definitions and assumptions. a truly longitudinal design. By ‘‘truly long- This finding was somewhat surprising to us itudinal,’’ we mean that the focal mistreatment given that these well-known and widely used construct and study outcomes are both measured mistreatment constructs are innately relational repeatedly for three or more time periods (Chan, and dynamic (e.g., Hershcovis & Reich, 2013). 1998; Ployhart & Ward, 2011). Further, 22% of The lack of empirical evidence for the dynamics the time-lagged studies (14 out of 63) incorrectly of mistreatment over time is not just surpris- stated that they had used a ‘‘longitudinal’’ ing, it also carries with it severe implications. design. With very few exceptions, then, when Within the broader organizational psychology researchers have examined the consequences of literature, scholars have strongly criticized interpersonal mistreatment at work, they primarily static research because most (if not all) orga- have done so in a manner that can be classified as nizational theories and the constructs used to a-temporal. test them are fundamentally dynamic (e.g., Of the remaining 13 studies that did account Ployhart & Vandenberg, 2010). Consequences for temporal factors, nine studies (3% of 266) associated with putting dynamic phenomena examined dynamic longitudinal relationships and their associations in static form include (Beattie & Griffin, 2014a, 2014b; Meier & ambiguous tests of theory, biased parameter Spector, 2013; Taylor, Bedeian, Cole, & Zhang, estimates, and quite possibly erroneous infer- 2014; Thau & Mitchell, 2010, Study 3; Tuckey ences made by the researchers (George & & Neall, 2014; A. R. Wheeler, Halbesleben, & Jones, 2000; Maxwell & Cole, 2007; Pitariu Whitman, 2013; Whitman, Halbesleben, & & Ployhart, 2010). For example, it has been Holmes, 2014; Zhou, Yan, Che, & Meier, 2014) shown that findings from a static study can through an objective time lens. We also iden- reverse in sign or disappear altogether when tified four studies (2% of 266) that we felt tested in a dynamic manner that is consistent with
Downloaded from opr.sagepub.com at TEXAS CHRISTIAN UNIV on July 15, 2016 Cole et al. 279 underlying theory (see, e.g., Vancouver, Tama- that nearly one quarter of these time-lagged nini, & Yoder, 2010; Vancouver, Thompson, & studies wrongly stated that they had used a Williams, 2001). Given the real possibility of longitudinal design. The top half of Table 2 inaccurate conclusions when mistreatment shows a few descriptive statistics for the vari- variables and their outcomes are relegated to a ous time lags used by mistreatment researchers. static form, the all-too-common practice of An inspection of this information suggests recommending a longitudinal research design considerable differences within and across the as ‘‘the next researcher’s responsibility’’ should mistreatment constructs (e.g., time lags ranged come to an end. Mistreatment scholars must from 10 days to 14 months), and we noted that begin to account for the role of time (e.g., the identified studies rarely justified their time George & Jones, 2000; Mitchell & James, lag choices. As will be discussed in a later 2001) and incorporate longitudinal research section, different time lags can yield sub- designs (e.g., Chan, 2014; McArdle, 2009) stantively different effect sizes even when the into their work. The need for longitudinal same mistreatment–outcome relationships are mistreatment research has been consistently under study. acknowledged (e.g., Hershcovis, 2011; Mar- We also observed that a-temporal studies, as tinko et al., 2013; Tepper, 2007), but it appears part of their research design, oftentimes asked that very few scholars have heeded these calls. individual respondents to summarize their This observation begs the question of why exposure to the negative acts over some specific more mistreatment researchers have not wel- period of time (e.g., ‘‘think back over the past comed a temporal lens with open arms. year, how often have you experienced ...’’) and correlate these summary judgments with Methodological challenges focal outcome variables. As shown in the bot- tom half of Table 2, there is considerable var- and misunderstandings iation in the use of recall instructions within and We attribute the current state of affairs (at least across the mistreatment constructs. For exam- partly) to methodological challenges and the ple, researchers interested in studying incivility fact that the term ‘‘longitudinal’’ is frequently typically (79% of the studies) asked participants misunderstood. According to Ployhart and to report on experienced incivility over a spe- Ward (2011), for example, the term long- cific period of time; the most frequently used itudinal ‘‘gets tossed about so much that it is recall timeframe was 1 year, although six confusing to know what is a longitudinal study’’ studies requested participants to summarize (p. 14). Following Chan (1998) and Ployhart their incivility experiences over a period of and his colleagues (Ployhart & Vandenberg, 5 years (M ¼ 540 days; SD ¼ 613 days). About 2010; Ployhart & Ward, 2011), we submit that a one quarter of bullying and social undermining study is considered longitudinal when it (a) studies explicitly reported use of a temporal emphasizes construct change and (b) contains a scale (i.e., timeframe under consideration), with minimum of three repeated observations. This bullying scholars occasionally asking partici- means that a study measuring the mistreatment pants to think back as far as 5 years into the past variable at Time 1 and the criterion variable at and social undermining scholars using recall Time 2 does not constitute a longitudinal study. instructions ranging from 1 week to 1 month. In We identified a number of time-lagged studies contrast, there was simply not enough infor- that separated the timing of such measurements. mation reported in most of the abusive super- Although the use of time lags is better than a vision studies (97%) to determine if recall cross-sectional design (i.e., it helps deal with instructions were used and what the temporal common method concerns), our analysis found scale for asking about past experiences might
Downloaded from opr.sagepub.com at TEXAS CHRISTIAN UNIV on July 15, 2016 Table 2. Research design issues of a-temporal studies by interpersonal mistreatment construct. Downloaded from Study characteristic Abusive supervision Bullying Incivility Social undermining
Time lag: opr.sagepub.com No lag used 68% (81 of 120 studies) 90% (70 of 78 studies) 76% (32 of 42 studies) 54% (7 of 13 studies) Yes, lag used 32% (39 of 120 studies) 10% (8 of 78 studies) 24% (10 of 42 studies) 46% (6 of 13 studies) M 3 months 342 days 88 days 246 days SD 3 months 158 days 67 days 165 days atTEXASCHRISTIAN UNIVonJuly15, 2016 280 Min 10 days 6 months 14 days 1 month Max 1 year 2 years 6 months 14 months
Temporal scale of recall instructions: Not reported 97% (17 of 20 studies) 77% (60 of 78 studies) 21% (9 of 42 studies) 77% (10 of 13 studies) Yes, recall instructions were reported 3% (3 of 120 studies) 23% (18 of 78 studies) 79% (33 of 42 studies) 23% (3 of 13 studies) M 250 days 244 days 540 days 22 days SD 191 days 218 days 613 days 13 days Min 1 month 1 month 14 days 7 days Max 1 year 5 years 5 years 1 month Cole et al. 281 have been. We also observed (though not pre- Finally, we identified very few studies that sented in our tables) considerable variation in attempted to understand the mistreatment phe- the design choices (e.g., temporal scale, overall nomena through a subjective account of time. study span) made by mistreatment researchers Although this finding is consistent with Shipp when developing longitudinal studies (details and Cole’s (2015) review of the micro literature available from the authors). To illustrate, con- as a whole, we took note of a few important sider two recent objective time studies on issues that pertain to the mistreatment literature workplace incivility (i.e., Meier & Spector, specifically. For instance, while coding the 2013; Zhou et al., 2014). We find it interesting mistreatment studies, we noticed that a study’s that both of these repeated-measures studies instructions quite often requested participants focused on the consequences associated with to ‘‘think back’’ or recollect past instances of incivility change and even used the same inci- mistreatment. Yet, these studies were not vility instrument, and yet they differ in the slice focusing on retrospected mistreatment as a the- of time under consideration: a day versus a oretically meaningful construct. Instead, when month. Such differences are important because, researchers asked participants to recollect past although the two research teams were interested mistreatment experiences for a specific period of in understanding the same phenomenon, the time, they did so as a means to capture sufficient experience of (and responses to) such behavior instances of transient mistreatment (i.e., a form will certainly have a different meaning when of summary evaluation). It seems that the use of considered daily for 10 days versus monthly for this type of retrospective recall blurs the lines 5 months (interested readers should consult between current and retrospected mistreatment. Zaheer, Albert, & Zaheer, 1999, for a detailed For this reason, we did not code these studies as discussion on time scales). adopting a subjective lens. We maintain that these relatively subtle design differences can have important theoretical and practical implications, including confused read- Accounting for time when ers and the potential for different conclusions exploring interpersonal based on study results. When deciding what temporal scale is most appropriate, we encourage workplace mistreatment researchers to do their homework. One’s theore- From our review of the literature, it is safe to tical justification for a specific temporal scale say that a vast majority of theory testing in the could, for example, be supplemented with mistreatment literature (i.e., abusive super- empirical evidence from prior longitudinal stud- vision, bullying, incivility, and social under- ies that employed the same mistreatment mea- mining) still uses a cross-sectional design, sure. Moreover, when discussing how a study’s wherein a researcher’s inferences are based on results connect with the existing literature, associations between two or more static vari- researchers may wish to incorporate a discussion ables. In this section, we discuss how incor- on how their design decisions could have porating one or more temporal lenses promises impacted the generalizability of findings. For to open new doors for empirical research on example, Taylor et al. (2014) clarified how their workplace mistreatment. We begin by elabor- dynamic mediation model—although receiving ating on the types of theoretical insights that empirical support when based on repeated mea- might emerge when an objective temporal lens sures data assessed weekly—may not be con- is adopted. We then explore areas in which a ceptually fitting for a study focusing on how subjective temporal lens is likely to have the targets react to incivility incidents as they happen largest impact on mistreatment research. In moment by moment. doing so, we integrate several perspectives on
Downloaded from opr.sagepub.com at TEXAS CHRISTIAN UNIV on July 15, 2016 Table 3. How to apply a temporal lens to mistreatment research.
Checklist item Reasons for importance Key considerations for implementation References
1. So, I am interested in conducting a longitudinal field study on one of the mistreatment constructs covered in this review and I plan to adopt an objective temporal lens. Now what? Identify the study’s A longitudinal research design does not, by Value added: Are you most interested in Chan (1998); Ployhart & core purpose default, make for a novel theoretical descriptive longitudinal research (i.e., Vandenberg (2010) contribution. What is it that you want to fundamental dynamics of the mistreatment accomplish? construct) or explanatory longitudinal Downloaded from How will your study on mistreatment research (i.e., investigating dynamic contribute to the literature and/or change the relationships between a mistreatment conversation? construct and outcomes)? 2a. I want to conduct a descriptive longitudinal study to explore when and how the mistreatment construct changes opr.sagepub.com over objective time. What’s next?
Conceptualize (and Hypothesizing the form of change (e.g., linear Onset, duration, offset: When does Monge (1990); Pitariu &
atTEXASCHRISTIAN UNIVonJuly15, 2016 graph) the predicted vs. nonlinear) is essential for making mistreatment begin, how long does it last, and Ployhart (2010); 282 form of change predictions that are falsifiable. when does it end? Ployhart & Vandenberg Providing a figure to illustrate the Trajectory: What is the shape and trend of (2010); Roe (2008) hypothesized form of change will help by mistreatment over time? Will there be no graphically depicting the predicted change change, or will the change increase (upwardly trend. positive) or decrease (downwardly negative)? Developing a theory for mistreatment change Could mistreatment change exponentially? will also inform research design issues; for Should it plateau with time, or exhibit multiple example, if interest is in the growth of abusive upturns or downturns over time? supervision perceptions, a researcher may Cycles and rhythms: Could mistreatment wish to sample only recently hired employees reoccur over time in a predictable way? Is as opposed to employees with varying levels there a reason to expect mistreatment to of tenure with a supervisor. spiral in an increasing or decreasing manner over time? Decide on a data The number of repeated measurements and Incident reporting: Is the exposure to Fisher & To (2012); Ployhart collection schedule the spacing between measurements (i.e., time mistreatment triggered by a specific event or & Vandenberg (2010); interval) will impact the study’s ability to does it routinely reoccur? Zaheer et al. (1999) detect and model meaningful forms of change. How often and for how long: Include a sufficient number of measurements to appropriately (continued) Table 3. (continued)
Checklist item Reasons for importance Key considerations for implementation References
model the hypothesized form of change (e.g., nonlinear change requires more measurements). Practicality: Given pros and cons of the various sampling approaches, which alternative is most likely to produce valid inferences Downloaded from without overburdening participants? 2b. Although the descriptive approach sounds interesting, I’m more interested in developing an explanatory longitudinal study to explore dynamic relationships between mistreatment experiences and targets’ responses over objective time. What’s next? opr.sagepub.com
Examine the functional Predict whether there should be change over Disclaimer: Just because data are collected Grimm et al. (2012) form of each time in each focal variable. repeatedly over time does not guarantee that
atTEXASCHRISTIAN UNIVonJuly15, 2016 dynamic construct variables will exhibit variability over time.
283 Types of change: Change in each variable can be constant or proportional (i.e., relative to its previous level). Conceptualize the Focus on theorizing why a change in Time: When will the relationship exist (e.g., Grimm et al. (2012); dynamic mistreatment will lead to a change in another within a day, from yesterday to today, week to Monge (1990); Pitariu & relationship(s) substantive variable. week, and so on)? Ployhart (2010) Duration: How long should the dynamic relationship exist? Shape: Will the dynamic relationship steadily increase (or decrease) over time? Or perhaps it will accelerate, decelerate, or even plateau? Hypotheses: Do the formal hypotheses precisely describe the dynamic relationship? Level of analysis Longitudinal data are hierarchical, with Single-level model: A within-person approach Singer & Willett (2003) repeated measures nested within participants. acknowledges that individuals may perceive Therefore, the level of change of interest they are mistreated more on some occasions needs to be clearly specified. and less on others, and that their attitudes and behaviors may change accordingly. (continued) Table 3. (continued)
Checklist item Reasons for importance Key considerations for implementation References