On artificial results due to using factor analysis for dichotomous variables

This paper serves to remind the reader, that factor analysis in the case of dichotomous variables will often lead to artificial factors. In other words, the factors correspond primarily to certain levels of item difficulty. A numerical example will be given in order to illustrate this. It is argued that, for instance, factoring tetrachoric correlations instead of conventionally used Pearson correlations would lead to more content valid results.