Title: SELECTION FOR TREATMENT AS A SOURCE OF BIAS IN BEFORE-AND-AFTER STUDIES
Abstract: The importance of the bias-by-selection phenomenon is twofold. Firstly, it tends to make results of specific projects look better than they are. Secondly, the cumulative effect of several biased before-and-after comparisons is to bring about an erroneous consensus about the degree of effectiveness for some treatment. This overestimate finds its way into manuals and handbooks, and may lead to the implementation of projects of doubtful validity. In this paper, a procedure for the elimination of the bias-by-selection is explained. It suffers from two principal limitations: (1) when the number of systems treated is small, the estimate of the bias is very unreliable; and (2) in most cases, the process by which systems are selected for treatment is not as clearly defined as is assumed in the analysis. Therefore, it is seldom clear to what extent the estimate of the bias applies in a specific case. Nevertheless, since the phenomenon is real, and appears to be important, the eliminatino of the bias should be undertaken by the best method available. (Author/TRRL)
Publication Year: 1980
Publication Date: 1980-08-09
Language: en
Type: article
Access and Citation
Cited By Count: 21
AI Researcher Chatbot
Get quick answers to your questions about the article from our AI researcher chatbot