Design and validation of matching tests to form equal groups for tracking experiments

Abstract

The main obstacle to using matched groups in human performance experiments is the general unavailability of valid matching tests. To remove this deficiency, a tracking performance test was developed for use in forming matched groups for continuous tracking experiments. The need for such a matching test and the advantages of matched groups over randomized groups (in between-subjects experimental designs) are discussed, as well as the authors' preference for between-subjects over within-subjects designs. The matching test's reliability was measured (099) and predictive and matching validities were determined in two validation experiments. Validation coefficients of 0·79-0·93 were obtained, which are considerably higher than necessary to prefer matched groups to randomized groups. The strategy used in developing this matching test is explained and may be useful to those who wish to develop their own matching tests for specific research purposes.