Standard Setting and Cut Score Validation
Standard Setting and Cut Score Validation
The cut score (passing standard) separates competent from non-competent candidates. Standard-setting methods are test-centered (e.g., Angoff, modified Angoff, Yes/No) or examinee-centered (e.g., contrasting groups, borderline group). The Angoff method asks SMEs to estimate the probability that a minimally competent candidate would answer each item correctly; the sum yields the cut score. Bookmark and item-mapping are also test-centered. Contrasting groups uses actual performance of known groups. Borderline group method uses performance of candidates identified as borderline. A common trap is treating the cut score as fixed forever; it must be validated periodically and after major exam changes. Validity evidence includes internal (e.g., consistency of judgments) and external (e.g., relation to job performance). Standard-setting panels must be trained and representative. Compromise methods (e.g., Hofstee) balance multiple criteria. Worked example: In Angoff, if 10 SMEs each rate an item’s probability of correct response for a minimally competent candidate, the average probability is used. If the sum of average probabilities across 100 items is 65, the cut score is 65. Rule of thumb: always document the standard-setting process, including SME selection, training, and judgments, to defend the cut score legally and professionally.