Correlations Between Latent Variables The Programs CORDIM And FMERGE
Door: Verhelst, N.D., Veldhuijzen, N.H. | 01-10-2002 In the analysis of test data, the most popular IRT models are the ones that assume a unidimensional latent variable.Sometimes, however, the correlations between two or more latent dimensions are needed. The present report
discusses a possible estimation procedure. One of the programs that does estimate correlations is the program MULTI by Frans Kamphuis. The procedure is a two-step procedure. In the first step the item parameters are estimated using CML-estimation for two or more scales. In the second step the item parameters are fixed, and the vector of means and the covariance matrix is estimated, using a Bayesian approach.
Kunnen we je helpen?
Stel je vraag via onze kanalen of kijk in de veelgestelde vragen.
Voor scholen: Vergeet niet om het brinnummer bij de hand te hebben en/of in de mail te vermelden, zodat we jouw vraag sneller kunnen behandelen!