Statistical approaches in alcohol research: a comparative survey of two major alcohol journals with four major psychiatric journals.

This study surveyed and compared the statistical methods used in two major alcohol journals with those used in four major psychiatric journals. The alcohol specialty journals were Alcoholism: Clinical and Experimental Research and Journal of Studies on Alcohol. The psychiatry journals studied were the American Journal of Psychiatry, British Journal of Psychiatry, Archives of General Psychiatry, and Acta Psychiatrica Scandinavica. The aim of this study was to examine the extent to which alcohol researchers use statistical methods and their level of statistical sophistication. A second aim focused on the extent to which alcohol researchers are attentive to design and sample size issues. Comparisons between papers published in the four psychiatry journals and the two alcohol specialty journals published in 1990 revealed that the percentage of articles without any numerical results was substantially different among the journals: 21% of the psychiatry journals and 7% in the alcohol journals. There was a significant difference in favor of the alcohol journals with respect to frequency of intermediate statistical techniques. The number of papers published in the two alcohol journals using 20 or fewer subjects was similar to the four psychiatric journals. However, the alcohol journals contained more articles in which > 400 subjects were studied.