Commitment Profiles and Latent Transition Analysis 1 Are Commitment Profiles Stable and Predictable? a Latent Transition Analysis

Recent efforts have been made to identify and compare employees with profiles reflecting different combinations of affective (AC), normative (NC), and continuance (CC) organizational commitment. To date, the optimal profiles in terms of employee behavior and well-being have been found to be those in which AC, NC and CC are all strong, or those where AC, or AC and NC, dominate. The poorest outcomes are found for profiles where AC, NC and CC are all weak, or CC dominates. The primary goal of the current study was to use Latent Profile Analysis (LPA) and Latent Transition Analysis (LTA) to identify profile groups and examine changes in profile membership over an 8month period in an organization undergoing a strategic change. We also tested hypotheses concerning the relation between perceived trustworthiness of management and employees’ commitment profile within and across time. We found that commitment profiles have substantial temporal stability and that trustworthiness positively predicts memberships in more desirable commitment profiles. There was also some, albeit weak, evidence that changes in perceived trustworthiness were accompanied by corresponding shifts in the commitment profile. COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 3 INTRODUCTION The three-component model of commitment (Meyer & Allen, 1991; Meyer & Herscovitch, 2001), defines commitment as a force that binds an individual to a target or course of action. However, this force can be characterized by three distinct mindsets – desire (affective commitment), obligation (normative commitment), and perceived cost (continuance commitment) – that can have different implications for behavior. Although Meyer and colleagues argued that the mindsets combine to influence behavior, most research has focused on their independent or additive effects. It is only recently that studies have examined the behavioral consequences of “commitment profiles” (e.g., Gellatly, Meyer, & Luchak, 2006; Wasti, 2005). These studies generated new insights into the nature and implications of commitment and served as the impetus for recent developments in commitment theory (Meyer, Becker, & Van Dick, 2006; Meyer & Maltin, 2010; Meyer & Parfyonova, 2010). The shift in attention to commitment profiles reflects a broader trend in organizational research toward greater use of a person-centered approach (see Wang & Hanges, 2011). In contrast to the more common variable-centered approach that aims to explain relations among variables, the person-centered approach involves the identification of homogeneous subgroups of individuals within a population. The person-centered approach treats individuals in a more holistic fashion and allows for the possibility that a set of attributes (e.g., commitment mindsets) might be experienced differently, and have different implications, in combination than they do individually. Consequently, the personcentered approach affords a different perspective on a phenomenon of interest and complements the variable-centered approach (Marsh, Lüdke, Trautwein, & Morin, 2009; Meyer, Stanley, & Vandenberg, 2013). To date, commitment profile studies have been cross-sectional and do not adequately address the important issue of profile stability. If the commitment profiles found across samples differ radically, or the profile structure within a sample is reactive to situational cues, it will be difficult to make meaningful recommendations. Therefore, our primary objective was to determine whether there is temporal stability in commitment profiles within a sample of employees. This study was conducted in an organization undergoing a large-scale change in strategy and culture, providing a strong test of within-sample stability. COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 4 Also, most of the attention in existing profile studies has been directed at their implications for behavior (Somers, 2010; Wasti, 2005) and well-being (Meyer, Stanley, & Parfyonova, 2012; Somers, 2009), with little concern for how these profiles are formed or change over time. In order to take advantage of what has been learned about the consequences of commitment profiles, we need to know more about what managers can do to foster desirable profiles and maintain them under conditions of change. Therefore, our second objective was to examine the role of one potential contributor to profile formation and change – the perceived trustworthiness of management. We focused on trustworthiness because it, and the trust it engenders, become particularly salient under conditions of change (Thomlinson & Mayer, 2009), and both have been linked to commitment in previous research (Colquitt, Scott, & Lepine, 2007; Dirks & Ferrin, 2002). Thus, there was good reason to believe that perceptions of management trustworthiness would be relevant to formation and change of commitment profiles. Commitment Profile: Theory and Research Meyer and Herscovitch (2001) offered a set of propositions concerning how various combinations (profiles) of the commitment mindsets – affective (AC), normative (NC), and continuance (CC) – would relate to behaviors (e.g., turnover, performance, organizational citizenship). They proposed that the optimal profiles from an outcomes perspective would be characterized by strong AC and relatively weak CC and NC (i.e., the less autonomously-motivated mindsets). The least desirable outcomes were expected for uncommitted employees (all components low) or those whose profile was dominated by strong CC. These propositions have been tested in several studies with mixed support. Although profiles characterized by strong AC were indeed found to be associated with desirable behaviors, the AC-dominant profile was not necessarily optimal. Indeed, several studies reported that intention to remain, OCB and well-being were greatest among employees with AC/NCdominant or fully-committed (high AC, NC, and CC) profiles (Gellatly et al., 2006; Wasti, 2005; Meyer, L. Stanley et al., 2012; Somers, 2009), suggesting a possible synergy of the three components (see Johnson, Groff, & Taing, 2009). Gellatly et al. (2006) interpreted their findings as evidence that the way any component of commitment is experienced will depend on the context created by the other components. For example, COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 5 combined with strong AC, NC may be experienced as a moral imperative, whereas with weak AC and strong CC it might be experienced as indebted obligation. Similarly, Meyer, L. Stanley et al. (2012) suggested that, on its own, strong CC might reflect entrapment due to lack of alternatives or the economic costs of leaving. Alternatively, when combined with strong AC and NC, CC could reflect awareness of the costs associated with the loss of desirable work and/or work conditions. Thus, the implications of CC and NC will depend on their relative strength within the full commitment profile. Considered together, the results of existing profile studies suggest that the optimal commitment profiles from an outcomes perspective are the fully-committed, AC/NC-dominant, and AC-dominant profiles. The poorest outcomes tend to be associated with the uncommitted and CCdominant profiles. Based on these findings, one might conclude that organizations should invest effort and resources to foster the optimal profiles. However, as noted previously, recommendations such as this rest on the assumption that there is a relatively standard set of distinguishable profiles within the workforce, that there are strategies organizations can use to foster these desirable profiles, and that, once established, commitment profiles remain relatively stable over time. These assumptions remain largely untested. In the discussion to follow we focus first on the issue of stability, and then on profile development. Stability of Commitment Profiles The stability of commitment profiles can be addressed in several ways. First, there is the question of whether a common set of profiles emerges across samples (i.e., cross-sample stability). This question is best answered by comparing profiles across studies. Although research is still limited, Meyer, L. Stanley et al. (2012) noted that several profiles emerge quite regularly. Indeed, all of the studies they reviewed identified fully-committed, AC/NC-dominant, CC-dominant, and uncommitted profiles. Most studies identified an AC-dominant profile and two studies found a CC/NC-dominant profile. Several studies also identified profiles in which scores on all three mindsets fell in the moderate range. The only profile described by Meyer and Herscovitch (2001) that has not been found, other than through median split approaches, is the NC-dominant profile. Thus, some profiles replicate quite consistently. Although other profiles emerge occasionally, this feasible set of profiles is relatively small and easily manageable. COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 6 A second question has to do with the temporal stability of commitment profiles within a sample. That is, will the same profiles be detected for a given sample on separate occasions? Recommendations that organizations select for or promote some profiles over others (e.g., Meyer, L. Stanley et al., 2012), or use different management strategies for different types of employees (e.g., Morin, Morizot, Boudrias, & Madore, 2011), assume that profiles persist over time. However, their temporal stability has yet to be investigated. The evidence for cross-sample consistency suggests, but does not provide direct evidence for, within-sample temporal stability. Consequently, addressing this issue was one of our key objectives. Finally, there is the question of temporal stability of individual employees’ commitment profiles. The questions of within-person and within-sample temporal stability are highly related. Temporal stability at the individual level virtually assures within-sample stability. However, even if individual employees’ profiles change over time, within-sample stability remains a possibility if the change involves balanced movement (i.e. switching) between existing profiles. For example, if some employees shift from a CC-dominant profile to an AC/CC-dominant profile over time, while other employees shift in the opposite direction, the profile structure of the sample should remain the same over time. Even when the switching is not fully balanced, the profile structure might still remain the same across time, although their relative sizes may differ. Only large and uniform shifts in individual employee profiles are likely to lead to within-sample instability. Thus, if a dramatic event caused a large proportion of employees with an AC/CC-dominant profile to shift to a CC-dominant profile, with no one moving in the opposite direction, the former profile might be detected before the event but not after. In addressing temporal stability, it is important to consider factors that might contribute to stability and change in individuals’ profiles. In theory, there are several reasons to expect the commitment mindsets to remain relatively stable over time. Mowday, Porter, and Steers (1982) argued that, by its very nature, (affective) commitment is a stable attitude emerging in part from a dispositional propensity to commit. Weiner (1982) proposed that NC develops largely as a function of socialization forces presumably designed and intended to create stability. Becker (1960) suggested that (continuance) commitment develops when individuals make “side bets” (e.g., investing time to develop organization-specific skills) that COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 7 make it costly to change one’s course of action, potentially for a considerable period of time. Although empirical evidence for dispositional influences on commitment is sparse (Meyer et al., 2002), a few recent studies have reported correlations between personality and AC (e.g., Erdheim, Wang, & Zickar, 2006; Panaccio & Vandenberghe, 2012), as well as relations between commitment mindsets and cultural values (Clugston, Howell, & Dorfman, 2000; Fischer & Mansell, 2009; Meyer, D. Stanley et al., 2012; Wasti, 2003). Finally, Morin, Morizot et al. (2011) found evidence of a general factor underlying AC to seven distinct work-relevant foci, suggesting the existence of a general tendency to commit. Thus, to the extent that these internal factors are free to operate (i.e., without strong counterforces in the environment), individual employees’ profiles should be expected to remain stable. There are also strong theoretical and empirical bases for expecting instability in commitment profiles over time. Indeed, in the initial formulation of the three component model, Meyer and Allen (1991, 1997) focused almost exclusively on situational factors as determinants of commitment. In contrast to research on dispositions, there has been an extensive body of research linking commitment to work conditions (or perceptions of these conditions). Meta-analyses provide strong evidence linking AC (and to a lesser extent NC) to perceived organizational support (Rhoades & Eisenberger, 2002), organizational justice (Colquitt, Conlon, Wesson, Porter, & Ng, 2001), empowerment (Seibert, Wang, & Courtright, 2011), trust (Colquitt, Scott, & Lepine, 2007), high involvement work practices (Jiang, Lepak, Hu, & Baer, 2012), and transformational leadership (Jackson, Meyer, & Wang, 2013). CC has been linked to lack of employment alternatives and non-transferability of skills and education (Meyer et al., 2002). Each of these situational factors is subject to change and, based on their relations with commitment, could contribute to changes in one or more of the commitment mindsets. Few studies have examined relations between situational factors and commitment over time, and the findings have been mixed (see Morrow, 2011). Some of the earliest longitudinal studies involving established employees provided little evidence for time-lagged relations between work conditions and commitment (Bateman & Strasser, 1984; Curry, Wakefield, Price, & Mueller, 1986). However, Meyer and colleagues (Meyer & Allen, 1988; Meyer, Bobocel, & Allen, 1991) found significant time-lagged relations between work experiences (e.g., job challenge) and commitment among new employees. These findings suggest that situational factors may play a role in shaping COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 8 commitment, but that they are most likely to do so under novel or changing conditions. Once formed, commitment might remain quite stable. In sum, there are reasons to expect both stability and changes in commitment over time. There is no strong evidence to suggest that organizational changes such as the one experienced by employees in our study will be sufficiently strong or uniform to produce temporal instability in the profile structure of an entire sample of employees. Therefore, we predicted that we would find several of the more common profiles in our sample, and that these would remain stable over time. Hypothesis 1: Our sample will be heterogeneous with regard to commitment profile and should include the following: fully committed, AC/NC-dominant, AC-dominant, CCdominant, uncommitted. Other possible profiles include CC/NC-dominant and all-mid profiles. Hypothesis 2: The same profiles will exist prior to and following the change. Perceived Management Trustworthiness and Commitment As noted earlier, there has been little research to identify factors involved in the formation of, or change in, commitment profiles (see Gellatly, Hunter, Currie & Irving [2009] for an exception). Therefore, our second objective was to investigate the roles of perceived management trustworthiness, and change in perceived trustworthiness, respectively, in the formation and changes in commitment profiles. Trust is commonly conceptualized as a willingness to make oneself vulnerable to the decisions or actions of others, whereas trustworthiness is a quality of the trustee (Mayer, Davis, & Schoorman, 1995; Colquitt et al., 2007). According to Mayer et al. (1995), judgments of trustworthiness reflect an assessment of three characteristics: benevolence (concern for the trustor’s well-being), ability (situation-relevant competence), and integrity (adherence to acceptable moral and ethical principles). Trust and trustworthiness are inextricably intertwined. Indeed, many measures of trust make direct reference to two or more of the facets of trustworthiness (Salamon & Robinson, 2008). We focused on trustworthiness rather than trust per se because the findings are likely to be more directly actionable. A basic theoretical underpinning of organizational commitment is social exchange (Meyer & Allen, 1991). At the heart of high quality exchanges is the belief that the other party will fulfill its COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 9 obligations (Blau, 1964; Cropanzano & Mitchell, 2005). Trust is therefore important at any stage of a relationship, but becomes central under conditions of uncertainty such as a large-scale organizational change (Mayer et al., 1995). In these contexts, employees are likely to be guided by their perceptions of managements’ trustworthiness. Indeed, there is considerable empirical evidence linking commitment, particularly AC, to trust (Dirks & Ferrin, 2002) and trustworthiness (Colquitt et al., 2007). In their meta-analysis, Dirks and Ferrin found that AC correlated positively with trust in top management and immediate supervisor, but that the former relation was stronger. They argued that the difference might be because top management plays a greater role in developing strategy and policy. Thus, when it comes to their willingness to commit to the organization, employees may pay particular attention to whether they trust top management to steer the organization in the proper direction. In contrast to its positive relation with AC, trust in management has generally been found to have a negative (Albrecht & Travaglione, 2003; Laschinger et al., 2000) or non-significant (Hopkins & Weathington, 2006; Ozag, 2006) relation with CC. Although these studies did not address the issue of causality, to the extent that a negative relation exists, we expect it may be because employees with strong CC find it difficult to leave the situation despite concerns about management’s trustworthiness. It is unlikely that lack of trust contributes directly to the perceived cost of leaving. To our knowledge, only two studies have examined the relation between trust and NC. Ozag (2006) found a positive correlation with a combined measure of trust in supervisor and the organization. Colquitt et al. (2012) found positive correlations with both affectand cognition-based measures of trust in supervisor. However, when the cognition-based measure of trust (conceptually similar to our trustworthiness measure) was included in a structural equation modeling analysis, the relation with NC disappeared. Thus, the findings pertaining to both CC and NC are somewhat inconsistent. It is important to note, however, that research is limited and has not considered CC or NC as they might be experienced within a commitment profile. Based on the foregoing theory and research, we developed hypotheses pertaining to profile formation. First, we expected that employees who perceived management to be trustworthy would be more likely to have a profile characterized by strong AC. Employees who see management as untrustworthy may have little reason to commit to the organization, and might therefore be COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 10 uncommitted (waiting for an opportunity to leave) or CC-dominant (seeing no alternative but to stay). Finally, employees who perceive management as trustworthy might also develop a felt obligation to remain (NC-dominant) as a means of reciprocation, or a sense of indebtedness due to expectation from other individuals (e.g., CC/NC dominant), although both NCand CC/NC-profiles were rare in past research. A trusting environment is likely to be perceived positively, in which case NC might combine with AC to form an AC/NC-dominant or fully-committed profile. Hypothesis 3: Employees’ perceptions of management trustworthiness will relate positively to their likelihood of having a fully-committed, AC/NC-dominant, or AC-dominant profile, and negatively to the likelihood of having an uncommitted or CC-dominant profile. Although some studies have examined the relations between trust and commitment under conditions of change (Albrecht & Travaglione, 2003; Laschinger et al., 2000), we are unaware of any studies that investigated how changes in perceptions of management trustworthiness related to changes in commitment mindsets over time. Consequently, our hypotheses were guided by the broader literature on commitment and change (see Meyer, 2009). Moreover, to be consistent with Hypothesis 2 regarding the temporal stability at the sample level, we focused our attention on the role that changes in perceptions of management trustworthiness might have on individual employees’ transitions between profiles over time. Morrow’s (2011) review of longitudinal studies revealed that commitment can increase or decrease as a consequence of organizational change. Although she did not address the role of trust per se, it is interesting to note that the strongest and most consistent evidence for a decrease in AC was obtained in the case of downsizing (e.g., Armstrong-Stassen, 1998). Changes such as this can lead employees to engage in a process of sense-making (Tomlinson & Mayer, 2009) with implications for the nature and strength of their commitment (Meyer, Allen, & Topolnytsky, 1998). For example, employees may see the change as unjust (Caldwell, Liu, Fedor, & Herold, 2009) or a violation of its psychological contract (Korsgaard, Sapienza, & Schweiger, 2002), thereby reducing perceptions of trustworthiness. However, large-scale organizational changes can also provide an opportunity for management to build trust by using fair procedures and/or communicating the need to “rewrite” the psychological contract in a mutually satisfactory manner (Meyer, 2009). Therefore, how employees COMMITMENT PROFILES AND LATENT TRANSITION ANALYSIS 11 react to management’s actions may depend on how they interpret the situation, and this could vary from employee to employee. This may have implications for temporal movement of profiles at the