Correlation

Abstract

A great many questions in applied research and evaluation in-volve the degree to which two variables "go together." For example, a researcher might ask whether, and to what extent, students who earn high scores on an aptitude test also tend to earn good grades in school. An employer might ask whether, and to what degree, job applicants who earn high scores on a screening test tend to receive high ratings from their supervisors on the job. A psychologist might ask whether, and to what extent, examinees who earn the highest scores on an anxiety scale tend to earn the lowest scores on an examination. All of these questions are concerned with the degree of covariation of two variables or measures. That is, they are concerned with the tendency of the measures to vary together, in the sense that one increases as the other increases (in which case we say that the variables "covary positively"), or in the sense that one variable increases as the other decreases (in which case we say that the variables "covary negatively"). If they covary, one can predict one variable from the other, for example, college performance from test results.