The present study concerns an examination of the predictive validity of several potential predictors of the work proficiency of computer programmers. The study was conducted by the Research Department of a large, multi-line insurance company. When the research design for this study was developed, it was anticipated that data would be collected on 150--200 computer programmers hired within one year. Unfortunately, it became impossible for this large a number of programmers to be hired. Therefore, the author was left with a research design which had been developed for a much greater number of subjects than it was possible to have. The data analyses were completed as planned even though the smaller-than-hoped-for number of subjects made some of the analyses inappropriate. The data were available so they were used in exploratory analyses, the intent of which was to gain insights which might be valuable in future research. This paper presents all the analyses which were performed in the hope that the method and techniques used will be informative and useful to some readers.
/lp/association-for-computing-machinery/an-examination-of-the-predictive-validity-of-several-potential-WDJgAJrkmr