March 1983

Study of Gluten Effect in Schizophrenia

Author Affiliations

Psychiatry Service Veterans Administration Medical Center 1030 Jefferson Ave Memphis, TN 38104
Department of Psychology Bronx Psychiatric Center 1500 Waters Pl Bronx, NY 10461

Arch Gen Psychiatry. 1983;40(3):345. doi:10.1001/archpsyc.1983.01790030115016

To the Editor.—  The study of the effects of gluten on schizophrenics that was reported by Storms et al (Archives 1982;39:323-327) was scientifically weak and did not provide an adequate basis for rejection of the hypothesis being tested. Its conclusions were predicated on the fallacy that failure to reject the null hypothesis is equivalent to its confirmation, thus violating Fisher's inference model in which the two logical alternatives are "reject" and "inconclusive."1 Statisticians have long stressed this point, warning that a methodologically weak study—for example, one with few subjects, little homogeneity of variance, and measures of low reliability—increases the risk of a type II error.The adequacy of the experiment used to test the null hypothesis that "gluten has no effect on schizophrenics" can be directly measured by calculating its power value. Using Cohen's methods,2 we find the power value of this study (two groups of 13 subjects

First Page Preview View Large
First page PDF preview
First page PDF preview