I am using the discrepancy analysis in TraMineR. I performed univariate analysis with "dissassoc" to measure the effect of each variable. Then, I did a multi-factors analysis using "dissmfac" to assess the effect of all the variables. I didn't understand how the total Pseudo R2 is calculated in the multi-factor analysis?
1 Answer
$\begingroup$
$\endgroup$
These methods are fully described in the following article:
Matthias Studer, Gilbert Ritschard, Alexis Gabadinho, and Nicolas S. Müller (2011), Discrepancy Analysis of State Sequences, Sociological Methods & Research 40: 471-510, doi:http://doi.org/10.1177/0049124111415372
The behavior of these functions may depend on whether the distance matrix is squared or not. In the above reference you can find an in-depth discussion on whether you should square it or not.
By default dissmfac square the distance matrix before computing the Pseudo-R2, whereas it is not the case for dissassoc. This is to be consistent with previous literature on the methodology behind dissmfac. The dissmfacw function does not square the distance matrix by default. In any case, you can change this behavior by using the argument squared=FALSE in all the above-mentioned functions.
