A good example of this frustration came recently when I advised my wife
(a research psychologist) to conduct some exploratory analyses via some
interesting graphical and modeling methods. The study she had set-up
was primarily interesting in exploratory questions. There was no need
(or relevance) to test statistical hypotheses. When the reviews came
back from her article they stated that she "had not done any statistics
and that they could not evaluate the article until some hypothesis tests
were conducted". The divorce hearings are now settling out of court (just
kidding) but it certainly was frustrating because she now had to conduct
endless numbers of hypothesis tests just to get her article reviewed. I
might also add that the journal was the top rated APA journal in her
area.
What resulted from the story above story about my wife is possibly the
answer to Dick Lehman's question. That is, I have since conducted a workshop
and written an article on the appropriate use of exploratory data analysis
for possible publication in a journal directly in her area. What I have
done is set aside some of my research time to prepare "didactic" or
"consumer's guides" on various issues. For example, last year the journal
"Canadian Psychology" (Canada's version of the American Psychologist)
published a symposium I had coordinated on nonparametric stat and other
alternatives to normal-theory stat. There were several papers all of which
were written for practising behavioral scientists.
Therefore, the answer to Dick's question is that we should set some time
aside as scholars to communicate with the practitioners with a practitioners
concerns in mind. We need to stick our necks out and make serious
recommendations. I am certainly not suggesting that we devote all of our
time to this but even a bit of time helps.
My two cents... Cheers, Bruno
Bruno D. Zumbo, Ph.D.
Measurement & Evaluation
University of Ottawa
PO Box 450 STN A
Ottawa, Ontario
Canada K1N 6N5
e-mail address: Zumbo@acadvm1.Uottawa.ca (internet)
Phone (613) 564-4222
FAX (613) 564-9098