Confusion about Negative Studies

Publisher Website
Google Scholar
Abstract
As Angell notes,1 a study that is well designed and conducted makes a positive contribution to biomedical knowledge; thus, it seems wrong to call it a "negative study." Others have requested that similar terms not be used,2 I believe with good cause. What one is really concerned with is the study results, for which the term "negative study" is too broad. Generally speaking, one can obtain one of five possible results in testing a single hypothesis: (1) a "positive" association between factors of interest (in favor of the test hypothesis) that is "statistically significant," (2) a "positive" association that is not statistically significant, (3) no association between the factors of interest (the findings favor the null hypothesis), (4) a "negative" or inverse association that is statistically significant, or (5) a "negative" association that is not significant. Presumably, the term "negative study" could refer to an investigation with any one of the last four results. In fact, the editorial1 used the term to refer to both results 2 and 3 listed above.

This publication has 2 references indexed in Scilit: