Organizational Commitment Profiles: A Cross-National Comparison

International audienceA growing academic literature dedicated to organizational commitment has adopted a person-centered approach in order to identify homogenous groups of employees with similar mindset profiles (i.e., relative levels of affective [AC], normative [NC], and continuance [CC] commitment; Meyer et al., 2013). Although multiple studies have identified a relatively common set of commitment profiles (cf. Meyer et al., 2012), and profiles have been found to remain stable with samples over time (Kam et al., 2013), there has been little systematic attempt to statistically compare profile similarity across diverse samples. For instance, relatively little attention has been paid to cross-cultural variations in the nature of observed commitment profiles. Our aim is to enrich the emergent knowledge in this area by proposing a cross national comparison, and illustrating a methodological approach specifically designed to investigate likely differences in the nature of latent profiles. For this purpose, we use a cross-cultural data set simultaneously collected in the United States (n = 492) and France (n = 476) and based on the same set of AC, NC, and CC measures (Meyer et al., 1993). As a preliminary verification of the measurement invariance of these measures across these two samples differing linguistically, we first conducted a series of confirmatory factor analyses across both groups of participants. The results from these analyses confirmed the complete equivalence of the measures (strict measurement invariance of the models, factor loadings, item intercepts, and item uniquenesses). The factor scores from this strictly invariant model were then saved, and submitted to a series of latent profiles analyses first conducted separately in both countries. In both countries, the results converged on highly similar 5-profile solution, although a few differences between these solutions could be noted. To systematically test whether these differences reflected true differences in the underlying population model, or simple sampling variations, we then conducted a systematic test of the measurement invariance of the identified profiles across countries. These tests supported the invariance of the within-profiles means and variance of the indicators across countries, but showed that the relative sizes of the profiles groups systematically differed across countries. The results from this invariant model are reported in Figure 4 and suggest the existence of 5 distinct profiles, associated with different sample sizes in both countries: (1) Uncommitted (US: 12.1%; France: 26.2%); (2) Weakly committed (US: 23.4%; France: 27.5%); (3) Morally devoted (US: 31.4%; France: 6.3%); (4) Morally allied (US: 26.5%; France: 14.2%); (5) Weakly emotionally committed (US: 6.6%; France: 25.9%)