Alternative techniques were judged on appropriateness, methodological strength, flexibility, sensitivity, confidence, transferability, costs, and understandability. Simpler techniques, such as ordinary least squares analysis, appeared to have lower costs and were easier to understand. However, the application of such simple techniques is questionable on technical grounds. The researcher using simpler techniques should be warned that estimates obtained may fail to have desirable properties (i.e., lack of efficiency). It is particularly likely that estimated standard errors will be biased and that tests of significance may be inappropriate. Lists of statistical techniques assessed and 22 references are supplied. (Author summary modified)