A COMPUTER-PROGRAM FOR PERFORMING METAANALYSES WHEN THE OUTCOME VARIABLE IS DICHOTOMOUS - RELEVANCE TO NEUROPSYCHOLOGY AND BIOMEDICAL-RESEARCH

The research clinical neuropsychologist is increasingly faced with the critical challenge of deciding whether the corpus of data in a given area of inquiry does or does not produce consistent or replicable findings. For our focus of interest, the most general case involves any clinical problem that can be cast into a 2 x 2 or fourfold table. One specific class of relevant queries would address those studies focusing on the efficacy of two forms of treatment (A or B) on the outcome (success, failure) of a specific type of brain injury. A second class would involve the effects of two treatments on the outcome (positive, negative) of specific types of neuropsychological disorders. This report describes a computer program that performs appropriate meta-analyses by providing statistics of choice that vary according to both the number of studies comprising the meta-analytic research endeavor and to the sample sizes within each study of focus.