2014-Aug-12, 03:23:36

Concerning rotation in PC/FA, I have always considered Promax to be the best, see here, although Oblimin should produce similar results.

Also, if you want a test (requested by Dalliard) for knowing which number of factors to retain in EFA, you should better avoid eigenvalue>1 and scree plot. You should better use Parallel analysis. It's a Monte Carlo simulation technique which evaluates what minimum eigenvalues are needed to reject the null hypothesis by subjecting a data set of random numbers to analysis (repeated many times, e.g., 1000 replications) when the data is adjusted for sample size and the number of variables.

See below for why you should prefer this one :

http://pareonline.net/pdf/v12n2.pdf

Determining the Number of Factors to Retain in EFA: an easy-to-use computer program for carrying out Parallel Analysis

In R, it's nFactors package you need.

http://cran.r-project.org/web/packages/n...actors.pdf

http://www.statmethods.net/advstats/factor.html

It should also give you a plot that helps you to determine the n factors to be retained.

Also, if you want a test (requested by Dalliard) for knowing which number of factors to retain in EFA, you should better avoid eigenvalue>1 and scree plot. You should better use Parallel analysis. It's a Monte Carlo simulation technique which evaluates what minimum eigenvalues are needed to reject the null hypothesis by subjecting a data set of random numbers to analysis (repeated many times, e.g., 1000 replications) when the data is adjusted for sample size and the number of variables.

See below for why you should prefer this one :

http://pareonline.net/pdf/v12n2.pdf

Determining the Number of Factors to Retain in EFA: an easy-to-use computer program for carrying out Parallel Analysis

In R, it's nFactors package you need.

http://cran.r-project.org/web/packages/n...actors.pdf

http://www.statmethods.net/advstats/factor.html

It should also give you a plot that helps you to determine the n factors to be retained.