Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Harman's single-factor test in Stata


From   John Antonakis <John.Antonakis@unil.ch>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Harman's single-factor test in Stata
Date   Fri, 13 Apr 2012 18:28:38 +0200

Hi:

Take a look at my "Causal claims" paper. Fixed-effects can only eliminate unmodeled variance due to constant effects; however, if your modeled independent variables are not exogenous then you may have a problem; and, the Harman "test" (which is far from being a test), will not help you.

Best,
J.

__________________________________________

Prof. John Antonakis
Faculty of Business and Economics
Department of Organizational Behavior
University of Lausanne
Internef #618
CH-1015 Lausanne-Dorigny
Switzerland
Tel ++41 (0)21 692-3438
Fax ++41 (0)21 692-3305
http://www.hec.unil.ch/people/jantonakis

Associate Editor
The Leadership Quarterly
__________________________________________


On 13.04.2012 18:09, Bernini, Michele wrote:
Hi John,

thanks for your answer.  I am using self-reported variable from survey data both on the LHS and on the RHS of my model and I intend to check for omitted variable bias (assuming they are time invariant) with fixed effect (I have got two cross sections for each individual). I planned to use Harman's to check for bias more closely related to the structure of the questionaire:

-for example as both my dependent and independent are ordered categorical, I want to see if questionnaire design induce to choose high categories.


do you think Harman's is redundant in this context?

Thanks again,

Michele





On 13 Apr 2012, at 16:52, John Antonakis wrote:

Hi Michele:

I would strongly advise you to not conduct such a test--thus, your Stata
question is not really of issue.

Common-method variance bias is a kind of omitted variable problem that
creates endogeneity. The problem you have is that you don't know how
this omitted variable affects the other variables in the model. The
Harman approach has been shown to not recover true model parameters:

Richardson, H. A., Simmering, M. J.,&  Sturman, M. C. (2009). A Tale of
Three Perspectives: Examining Post Hoc Statistical Techniques for
Detection and Correction of Common Method Variance. Organizational
Research Methods, 12(4), 762-800.

Antonakis, J., Bendahan, S., Jacquart, P.,&  Lalive, R. (2010). On
making causal claims: A review and recommendations. The Leadership
Quarterly, 21(6), 1086-1120. (this paper also shows how endogeneity
creates bias and that only instruments can help you).

Only a design that eliminates common-method variance, or instrumental
variables, can save the day.  The latter are hard to come by but if you
have them then you can purge the model from the endogeneity bias.  See also:

Podsakoff, P. M., MacKenzie, S. B.,&  Podsakoff, N. P. (2012). Sources
of Method Bias in Social Science Research and Recommendations on How to
Control It. Annual Review of Psychology, 63(1), 539-569.

HTH,
JOhn.

__________________________________________

Prof. John Antonakis
Faculty of Business and Economics
Department of Organizational Behavior
University of Lausanne
Internef #618
CH-1015 Lausanne-Dorigny
Switzerland
Tel ++41 (0)21 692-3438
Fax ++41 (0)21 692-3305
http://www.hec.unil.ch/people/jantonakis

Associate Editor
The Leadership Quarterly
__________________________________________


On 13.04.2012 17:35, Bernini, Michele wrote:
Dear Statalister,

I am trying to perform an Harman's single-factor test with stata to check for Common Method Variance. However I am not sure which method should I use when I do Factor Analysis with Stata. For example Principal factor confirms the presence of CMV while Principal-Component factor rejects it.

Thanks for your help!



Michele Bernini
Phd Candidate
School of International Studies (SIS)
University of Trento
Via Verdi, 8/10
I-38122 Trento
Italy

Tel. +39 3491831687





*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
Michele Bernini
Phd Candidate
School of International Studies (SIS)
University of Trento
Via Verdi, 8/10
I-38122 Trento
Italy

Tel. +39 3491831687





*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index