A Profile of Profiles: A Meta-Analysis of the Nomological Net of
Commitment Profiles
Adam H. Kabins
Hay Group/Korn Ferry, Dallas, Texas
Xiaohong Xu and Mindy E. Bergman
Texas A&M University
Christopher M. Berry
Indiana University
Victor L. Willson
Texas A&M University
Although the majority of empirical commitment research has adopted a variable-centered approach,
the person-centered or profiles approach is gaining traction. One challenge in the commitment
profiles literature is that names are attached to profiles based on the within-study comparison among
profiles and their relative levels and shapes. Thus, it is possible that different studies name the
same profiles differently or different profiles similarly because of the context of the other profiles
in the study. A meta-analytic approach, combined with multilevel latent profile analysis (LPA) that
accounts for both within- and between-sample variability, is used in this study to examine the
antecedents and outcomes of commitment profiles. This helps solve the naming problem by
examining multiple data sets (K = 40) with a large sample (N = 16,052), obtained by contacting
commitment researchers who voluntarily supplied primary data to bring further consensus about the
phenomenology of profiles. LPA results revealed 5 profiles (Low, Moderate, AC-dominant, AC/
NC-dominant, and High). Meta-analytic results revealed that high levels of bases of commitment
were associated with value-based profiles whereas low levels were associated with weak commit-
ment profiles. Additionally, value-based profiles were associated with older, married, and less
educated participants than the weak commitment profiles. Regarding outcomes of commitment,
profiles were found to significantly relate to focal behaviors (e.g., performance, tenure, and
turnover) and discretionary behaviors (e.g., organizational citizenship behaviors). Value-based
profiles were found to have higher levels of both focal and discretionary behaviors for all analyses.
Implications for the commitment and profile literature are discussed.
Keywords: organizational commitment, commitment profiles, profile analysis, latent profile analysis,
meta-analysis
Organizational commitment has traditionally been investigated
with three mindsets: affective (AC), normative (NC), and contin-
uance (CC) commitments; respectively, they reflect bonds based in
desire, obligation, and need (Meyer & Allen, 1997; Meyer &
Herscovitch, 2001). Most research to date has investigated the
three mindsets of commitment in terms of the effects that each one
individually has on outcomes (i.e., variable-centered approach),
with the rare article examining interactions among the mindsets
and their effects on outcomes (e.g., Gellatly, Meyer, & Luchak,
2006; Meyer, Stanley, & Parfyonova, 2012; Wasti, 2005). Re-
cently, commitment research has begun to utilize person-centered
approaches, whereby commitment profiles are defined by grouping
individuals into homogenous subsets based on relative endorse-
ment of AC, NC, and CC (Meyer & Parfyonova, 2010) using
methods such as cluster analysis (Wasti, 2005), median splits
(Gellatly et al., 2006), or latent profile analysis (LPA; Meyer et al.,
2012).
Although there is growing consensus regarding the commitment
profiles that occur (Meyer et al., 2012), there are a number of factors
that limit the previous research on profiles. First, although many
profile studies have relatively large sample sizes, profile analyses
segment the sample into homogenous subgroups, oftentimes leading
to comparisons between small groups. While this segmentation is the
purpose of profile creation, segmented subpopulations in relatively
small studies can significantly limit a researcher’s ability to make
inferences from the results. Second, virtually all profile studies have
focused on outcomes of commitment profiles, largely ignoring pre-
dictors of commitment. As a result, a number of theoretical questions
about how commitment profiles develop are left unanswered. Third,
the few studies that have taken a person-centered approach to com-
mitment have used a variety of profile-generating techniques, some of
This article was published Online First March 7, 2016.
Adam H. Kabins, Hay Group/Korn Ferry, Dallas, Texas; Xiaohong Xu
and Mindy E. Bergman, Department of Psychology, Texas A&M Univer-
sity; Christopher M. Berry, Kelley School of Business, Indiana University;
Victor L. Willson, Department of Educational Psychology, Texas A&M
University.
Correspondence concerning this article should be addressed to Adam H.
Kabins, Hay Group/Korn Ferry, 5001 Spring Valley Rd #800, Dallas, TX
75244. E-mail: Ahk325@gmail.com
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Applied Psychology © 2016 American Psychological Association
2016, Vol. 101, No. 6, 881–904 0021-9010/16/$12.00 http://dx.doi.org/10.1037/apl0000091
881