Reliability Theory and Research Decision in Selection of a Criterion Score

Abstract

The problem of selecting a criterion measure when several trials are available was considered in the light of pertinent principles of reliability analysis. Based upon reliability theory, it was shown that if no trend in trials was present the correct criterion measure was the mean of all available trials. When a trend was present, denoting systematic trial-to-trial error variance, several alternatives were available. The recommended procedure when a trend was present was to search for a measurement schedule free of systematic measurement error variance. In the event a search for a reliable measurement schedule was impossible, a logical and theoretically defensible procedure for selection of high or low score as a criterion measure was described along with necessary cautions.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 41 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.