10
$\begingroup$

Which statistical method should I use when Predicting organization commitment (composed of 3 latent variables) using the five factors of personality?

Its a cross-sectional design with survey data. The Personality factors are Five independent factors (derived out of factor analysis) and each factor is measured on a likert scale, same goes for the three components of commitment and they are also measured on a likert type scale.

The studies I have read upto now uses partial correlations and hierarchical regression to link these contructs but I somehow think there has to be a better way.

Note: this question was originally asked here on Stats.SE, but the moderator there suggested posting it on cogsci.se

$\endgroup$
5
  • $\begingroup$ It might be hard to answer your question without knowing how the study is designed and which type of data you are collecting. $\endgroup$
    – H.Muster
    Apr 20, 2012 at 20:24
  • $\begingroup$ Its a cross-sectional design with survey data. The Personality like I said is derived out of Five independent factors (derived out of factor analysis) and each factor is measured on a likert scale, same goes for the three components of commitment and they are also measured on a likert type scale. The studies I have read upto now uses partial correlations and hierarchical regression to link these contructs but I somehow think there has to be a better way $\endgroup$
    – Angie
    Apr 20, 2012 at 20:40
  • $\begingroup$ Adding in as much helpful info as you can into the body of the question will help us come up with the best solution. I'm afraid I don't know as much on the stats side myself though. $\endgroup$
    – Ben Brocka
    Apr 20, 2012 at 21:51
  • $\begingroup$ Angie, your question is welcome, but more directed to mods and active users, this raises the somewhat unresolved issue of when a question should be migrated to stats.se? This links in with existing meta here $\endgroup$ Apr 21, 2012 at 2:01
  • $\begingroup$ The question is fine on here (though it'd be fine on stats) my problem was the original question wasn't elaborated on enough $\endgroup$
    – Ben Brocka
    Apr 21, 2012 at 15:11

1 Answer 1

8
$\begingroup$

I'm guessing you are relating the Big 5 factors of personality to Meyer and Allen's three factor model of commitment (normative, continuance, affective).

You certainly could adopt the simple approach of creating observed composite scores for each variable and reporting the correlation matrix and a set of multiple regressions.

However, the language of your question also suggests that you are interested in modelling latent variables rather than observed variables. This may suggest a structural equation modelling approach may be suitable.

Benefits of structural equation modelling approach

  • You can model your measurement model (i.e., do the observed variables load appropriately on your 8 latent factors; is there an overall higher order commitment factor that represents the three commitment factors)
  • You could model relationships between latent factors rather than observed variables (i.e., you can adjust for reliability of measurement); this often aligns analyses more with theoretical interests in psychology.
  • You could explore more sophisticated model comparison approaches. One approach I really like involves comparing models with various constraints in either the correlations or regression coefficients between the predictor set (i.e., personality) and the outcome set (i.e., commitment). E.g., one model could constrain all 15 regression coefficients between personality and commitment variables to be the same; another model could say that the relationships are constrained to be equal across commitment variables but different across personality variables.
$\endgroup$
2
  • 1
    $\begingroup$ No worries. I had a similar problem when relating the big 5 to 7 factors of well-being. In the end I adopted a similar approach to that described above. I found the model comparison SEM approach led to some unique insights: see: Grant, S., Langan-Fox, J., & Anglim, J. (2009). Big Five Traits as predictors of subjective and psychological well-being. Psychological Reports, 105, 201-231. $\endgroup$ Apr 21, 2012 at 13:32
  • $\begingroup$ Jeromy, I am trying to get access to this article through athens, but its not available in Science Direct,Springer,Proquest etc only on Amsci. So I signed up but still cant view it. I tried searching your Blog aswell but cant access the journal. How should I get hold of it?Thanks $\endgroup$
    – Angie
    Apr 21, 2012 at 14:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.