Statistics

Abstract

I wonder if it would be worthwhile to call the attention of readers of Pediatrics to an error which is made from time to time in the statistical analysis of clinical data, and which has escaped the eyes of editors more than once. The error arises when an investigator has a large mass of facts, a punch card machine but no hypothesis. To illustrate, consider a clinical survey in which 100 different relationships are tested by the chi-square method.