Per Bech: Clinical Psychometrics
Jose de Leon' s comment on Donald F. Klein's response to Per Bech's and Martin Katz's reply to Klein's comment
I am perplexed that Dr. Klein, a psychiatrist, gets himself further into the shifting sand of statistics by making a second comment. This time, without hiding my thoughts in a rambling introduction, I would like to say that I obviously agree with Dr. Klein’s comment, “I don't think that factor analysis can effectively resolve mixtures.” However, his second comment, “I would appreciate it if Bech could refer me to studies where differences in Rasch scores provided effective comparative measures. A comparison to standard techniques, such as ANCOVA, would be valuable,” is quite perplexing. ANOVA and ANCOVA techniques are frequently used by psychiatrists who, according to Vickers in his “Analysis of variance is easily misapplied in the analysis of randomized trials: a critique and discussion of alternative statistical approaches,” usually have no clue of the major limitations of these techniques.
My recommendation concerning this set of critiques is to bring a more worthy adversary for Dr. Bech, who understands IRT. Having people like Doctors Katz, Klein and me swimming in the troubled waters of complex statistics may not be a good idea; Dr. Bech may be one of the few psychiatric researchers who may be qualified for that type of swimming. The downside is that a statistician may have difficulty grasping the issues involved in dealing with the complexity of mental symptom measurement.
Vickers AJ. Analysis of variance is easily misapplied in the analysis of randomized trials: a critique and discussion of alternative statistical approaches. Psychosom Med. 2005; 67:652-5.
November 28, 2019