Saris, Willem E.2017-04-052017-04-052014-06http://hdl.handle.net/10230/30791At several occasions we have been asked why our research with respect to quality and invariance of survey questions has been based on analyses of product-moment (Pearson) correlations and not on polychoric and polyserial correlations or latent trait models for categorical variables. In this research note a description will be given of the possible options, their advantages and disadvantages based on Monte Carlo experiments, real experiments and a meta-analysis. After that we will indicate the empirical arguments that brought us to the choice we have made.application/pdfengThis is an Open Access article distributed under the terms of the Creative Commons Attribution License Creative Commons Attribution 4.0 International, which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properlyattributed.info:eu-repo/semantics/workingPaperinfo:eu-repo/semantics/openAccess