Re: st: Factor Analysis for Panel Data

[Cameron McIntosh <cnm100@hotmail.com>]

I'm not surprised that Stata (not STATA) doesn't allow you to just forcibly remove temporal measurement invariance from modeling consideration, because it is illegitimate to do so. The invariance needs to be respected. I would suggest some form of latent growth curve modeling to examine evolution in latent (factor) trajectories over time, and which takes measurement invariance into account. The following may be of interest:
 
Ximénez, C., & Revuelta, J. (2010). Factorial Invariance in a Repeated Measures Design: An Application to the Study of Person-Organization Fit. The Spanish Journal of Psychology, 13(1), 485-493.
http://redalyc.uaemex.mx/redalyc/pdf/172/17213039040.pdf
 
Ferrer, E., Balluerka, N., & Widaman, K.F. (2008). Factorial invariance and the specification of second-order latent growth models. Methodology: European Journal of Research Methods for the Behavioral and Social Sciences, 4(1), 22-36.
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2799302/pdf/nihms114544.pdf
 
Widaman, K.F., Ferrer, E., & Conger, R.D. (2010). Factorial Invariance Within Longitudinal Structural Equation Models: Measuring the Same Construct Across Time. Child Development Perspectives, 4(1), 10–18.
http://onlinelibrary.wiley.com/doi/10.1111/j.1750-8606.2009.00110.x/pdf
 
Drapeau, A., Beaulieu-Prévost, D., Marchand, A., Boyer, R., Préville, M., & Kairouz, S. (2010). A life-course and time perspective on the construct validity of psychological distress in women and men. Measurement invariance of the
K6 across gender. BMC Medical Research Methodology, 10(68).
http://www.biomedcentral.com/content/pdf/1471-2288-10-68.pdf
 
Flora, D.B., Curran, P.J., Hussong, A.M., & Edwards, M.C. (2008). Incorporating Measurement Nonequivalence in a Cross-Study Latent Growth Curve Analysis. Structural Equation Modeling, 15, 676–704.
http://www.unc.edu/~curran/pdfs/Flora,Curran,Hussong&Edwards(2008).pdf
 
Chan, D. (1998). The conceptualization of change over time: An integrative approach incorporating longitudinal means and covariance structures analysis (LMACS) and multiple indicator latent growth modeling (MLGM). Organizational Research Methods, 1, 421–483.
 
Millsap, R.E. (2009). Testing Measurement Invariance Using Item Response Theory in Longitudinal Data: An Introduction. Child Development Perspectives, 4(1), 5-9.
 

Cam
 
> Date: Mon, 27 Aug 2012 12:39:01 +0200
> Subject: st: Factor Analysis for Panel Data
> From: mschimmer@gmail.com
> To: statalist@hsphsun2.harvard.edu
> 
> Dear Statalisters:
> 
> I currently try to reduce the number of performance variables in a
> panel dataset, starting from 17 variables. Theoretically, there should
> be three factors (profitability, efficiency, growth). However, when
> doing a factor analysis per year, I receive between 3 and 5 factors,
> depending on the year.
> 
> If I would do the factor analysis in a fashion that considers the
> panel character of the dataset (which would enforce an equal number of
> factors for all years), I believe I should receive variable weights
> that lead to a stable 3 factor solution over all years.
> 
> The factor command of STATA (11) seems not to support this feature.
> Does anyone have a suggestion on how to resolve the issue?
> 
> Thank you and best,
> Markus
> *
> * 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/