The Russell Sage Social Relations Test

Using a technique that controlled exposure of items, the investigator examined the effect on mean test score, item difficulty index, and reliability and validity coefficients of the reordering of items within a power test containing ten letter-series-completion items. The results suggest that effects on test statistics from item rearrangement are, generally, minimal. The implication of these findings for test designs involving an item sampling procedure is that performance on an item is minimally influenced by the context in which it occurs.