This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
To the Editor:—
The article on "Treatment of Chronic Schizophrenia" in the Sept. 28 issue of The Journal, page 333, contains a serious misstatement which should be pointed out to readers of The Journal, The authors state that they calculated a chi-square of 13.85 by using a fourfold table, though they do not make clear what the two entries were other than 61 and 28. They say that this chisquare gives a value for p of 0.0003, and go on to say that a "p value of less than 0.05 is adequate to establish a hypothesis." This is entirely erroneous and, if believed, can result in serious errors. No p, however small, can ever establish that a hypothesis is correct, nor can it even measure the probability that it is correct; p merely is the probability that if a given hypothesis is correct, then chi-square will be found at least
Zeisler EB. TREATMENT OF CHRONIC SCHIZOPHRENIA. JAMA. 1957;165(13):1739. doi:10.1001/jama.1957.02980310091023