The problems of R in clinical settings and in research: suggestions for the future

It has been suggested that we control for R in the interpretation of individual Rorschach records. A related, but distinctly different issue concerns controlling for or equating R in research. I suggest we should not control for R in interpreting individual records. There are also some research paradigms in which controlling for R does not seem advisable. In other types of research, I argued that some method of controlling for or equating R appears appropriate. The distributions of many Rorschach factors are markedly non-normal, and in some cases the shapes of the distribution for a particular factor may be different among different populations. In these instances, when attempting to control statistically for R, improper model specification may lead to erroneous findings. Finally, I argue for a theory-based, empirical approach to all future Rorschach research.