I am doing my research on Personality (the big five factors) and if personality can predict commitment to an organization. However, commitment is composed of three dimensions.

What is the best method I can use when trying to link the big five factors of personality with commitment to an organization?

  • $\begingroup$ Consider for example canonical correlation analysis $\endgroup$
    – ttnphns
    Apr 20, 2012 at 16:51
  • $\begingroup$ Due to its vague statistical content and implicit reliance on a specific psychological theory, this question might be more suitable for the cognitive science site. $\endgroup$
    – whuber
    Apr 20, 2012 at 17:51
  • $\begingroup$ Just for closure, the question was subsequently asked on cogsci.se here $\endgroup$ Apr 23, 2012 at 7:52

1 Answer 1


Your question is fairly general, but based on what you've said, it sounds like a structural equation model may be the best choice. I say this because you are working with observed indicators of latent factors, some of which have a directional effect onto others.

The example you give sounds like it would fit in well with many of the conceptual examples that exist in the SEM literature, textbooks, online help, etc. You should have an easy time mapping your application onto the examples. Programs such as AMOS (an SPSS add-on), SAS, LISREL, and M-Plus can fit the type of model you'll need.

Be sure to observe the assumptions of SEM, most of which will apply to the endogenous factors of organizational commitment, and to add disturbances to the exogenous factors.

Hope that helps,

  • $\begingroup$ One more question, 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. So do you think SEM is better than the procedures used thus far? Thanks a Bunch, Angie $\endgroup$
    – Angie
    Apr 20, 2012 at 19:53
  • $\begingroup$ There doesn't seem to be an objectively "better way," Angie. The methods you mention have different aims and different scopes. SEM (or canonical correlation, as @ttnphns said) would potentially be more comprehensive than the others in helping you map out a whole network of relationships. But SEM also has some pretty restrictive assumptions and requires greater expertise than do partial correlation or hierarchical regression. $\endgroup$
    – rolando2
    Apr 20, 2012 at 23:04
  • $\begingroup$ Sorry for the delay, Angie. I was away from the interwebs for a while. @rolando2 is right; there's no single correct approach. Regression is just fine, but it isn't as holistic as SEM since it dives the dataset into separate outcomes, each with their own model (in spite of what was stated above, SEM is less restrictive than regression). Canonical correlation is, well, just correlation; you don't get all that much from it. SEM gives tests, fit indices, regression weights, and more. But it also needs to be handles by a professional. $\endgroup$ Apr 25, 2012 at 21:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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