How useful are work samples in validational studies?

Some job tasks do not lend themselves to formal on-the-job assessment becau se they do not occur with sufficient regularity to permit the standardized measurement required in validational research. The preparation of incident reports by police and security officers is such a job task. The production of accurate, literate incident reports is important because these reports a re often required in legal proceedings. Their standardized evaluation on th e job is not practical because incidents occur at unpredictable intervals w ith highly variable content. Given the limitations of on-the-job performanc e criteria, we developed a standardized work sample by preparing sets of no n-verbal drawings depicting incidents, each of which required a written des criptive report by security officers. A total of 187 security officers comp leted a cognitive and personality test battery and criterion incident repor ts based on the standardized materials. Reports supported the usefulness of the standardized work sample, as well as the validity of the test battery - 100% of participants in the upper quartile of the test score distribution produced satisfactory reports, while only 17% of those in the lowest quart ile produced satisfactory incident reports. A number of advantages of struc tured work samples as criterion measures are noted, including their greater standardization, the elimination of range restriction problems by administ ering them to all job candidates, the opportunity to obtain expert evaluati ons of work samples at remote sites, and their face and content validity re sulting in acceptability to job candidates and to decision makers.