[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
[Skip to Content Landing]
November 11, 1983

Multivariate Analysis

Author Affiliations

University of Alabama Medical Center Birmingham

JAMA. 1983;250(18):2471-2472. doi:10.1001/jama.1983.03340180033019

To the Editor.—  The article by Lew et al (1983;249:641) describes some important principles about the multivariate analysis; however, some practical limitations and other equally important considerations were omitted. There were also misleading statements and inaccuracies in the text, particularly in citing the results from our melanoma articles.The authors advocate the use of cross-tabulations or survival curves for all possible combinations of variables. A well-known limitation of these statistical summaries is the difficulty in extending analysis beyond two or three variables unless a large sample is available. In analysis of survival data, this is further complicated by the possibility of varied censoring pattern and duration of patient follow-up within different subgroups. No adequate statistical inferences can be drawn from "overstretched" cross-tabulations. The authors failed to recognize these limitations when criticizing one of our articles.1Contrary to the authors' contention, in our studies the surgical treatment effect was routinely