Assumptions for well-known statistical techniques: Disturbing explanations for why they are seldom checked

A valid interpretation of most statistical techniques requires that the criteria for one or more assumptions are met. In published articles, however, little information tends to be reported on whether the data satisfy the assumptions underlying the statistical techniques used. This could be due to self-selection: Only manuscripts with data fulfilling the assumptions are submitted. Another, more disquieting, explanation would be that violations of assumptions are hardly checked for in the first place. In this article a study is presented on whether and how 30 researchers checked fictitious data for violations of assumptions in their own working environment. They were asked to analyze the data as they would their own data, for which often used and well-known techniques like the t-procedure, ANOVA and regression were required. It was found that they hardly ever checked for violations of assumptions. Interviews afterwards revealed that mainly lack of knowledge and nonchalance, rather than more rational reasons like being aware of the robustness of a technique or unfamiliarity with an alternative, seem to account for this behavior. These data suggest that merely encouraging people to check for violations of assumptions will not lead them to do so, and that the use of statistics is opportunistic.

Abstract licence: CC BY

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.