

Generating outofsample factor scores 

Message/Author 

Tom Hanson posted on Tuesday, May 10, 2011  12:51 pm



I have conducted a confirmatory factor analysis using categorical indicators on a large sample. I would like to use the results of this model to approximate factor scores for individuals not in the analysis dataset. It seems possible to use fixed parameters from the estimated model for a new data set comprised of individuals for whom I am trying to get factor scores for. (1) Does this seem like the proper way to do this? (2) If so, should I also fix the thresholds to what they were in the analysis dataset? (3) Does fixing the thresholds plus all the other parameters mean that any differences between the analysis data set and the new data set in levels of the factors will be picked up in differences in the factor score means? Thanks so much for any insight you can provide. 


You can fix all parameters of the CFA solution and then generate factor scores for a new sample. The model should have no free parameters. You make the assumption that the new sample comes from the same population as the sample used in the CFA. 

Tom Hanson posted on Wednesday, May 11, 2011  12:09 pm



Thanks so much Linda... 

Back to top 

