Multiple True‐False Items: A Study of Interitem Correlations, Scoring Alternatives, and Reliability Estimation

Abstract

Intercorrelations among multiple true‐false items were examined to determine to what extent each true‐false option can be treated as independent. Results from 157 health science students and 170 medical students showed that correlations between true‐false options associated with the same stem were from 2.6 to 7.0 times larger than those from different stems. This suggests that results from previous research indicating that each true‐false option could be treated as an independent item cannot be generalized to other tests and examinee populations without supporting evidence. Four scoring methods were explored which varied chance success levels and scoring for partial knowledge. The results showed that scoring methods incorporating partial knowledge were more reliable and possessed greater concurrent and predictive validity than those minimizing chance success. Methods for computing reliability estimates were compared and suggestions were offered regarding practical use

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 — 40 citations exist upstream.

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