aliquote.org

Contemporary or modern psychometrics

December 11, 2011

That is certainly a minor issue of terminology, but what is best: “contemporary” or “modern” psychometrics?

I have often encountered the term modern psychometrics when speaking of Item Response Theory models as opposed to the Classical Test Theory approach where all statistical indicators rely on untransformed (raw) scores. There is even a book with this title, precisely: Rust, J. and Golombok, Modern Psychometrics: Science of Psychological Assessment, 2008, 3rd ed., Routledge.

I don’t have this book but the table of content does suggest that no particular chapter is devoted to IRT, but it rather looks like a discussion of the foundation of measurement and test theory, much like The new psychometrics: science, psychology, and measurement (emphasis is mine), by Paul Kline (Routledge, 1998), with applications in intelligence and personality assessment.

On the contrary, the Handbook Statistics, volume 26 (North Holland, 2006), includes a larger number of chapters describing the foundations and applications of IRT models. I recently brought another book which looks very interesting as it features distinct chapters for factor-analytic models, structural equation modeling, multidimensional scaling (an often forgotten aspect of psychometrics): Maydeu-Olivares, A. and McArdle, J.J., Contemporary Psychometrics, 2005, Psychology Press. Still no chapter on IRT.

In sum, I cannot find any distinctive feature when one is talking about modern, or new, or contemporary psychometrics. Of course, I have a lot of books dealing with IRT models. I personally like:

although the latter is mostly a series of chapters on applications of IRT in various fields of research. I have noticed that Chapman & Hall/CRC now has an “R series” like Springer where Mair and coll. plan to discuss their Extended Rasch Modeling framework (see the related JSS paper or the UseR! 2010 slides) (PDF).

See Also

» Dimensions or categories? » Cronbach's alpha yet again » Computing intraclass correlation with R » Fitting Rasch model with R lme4 » Psychoco 2011