Critical difference table for word recognition testing derived using computer simulation

Abstract

Purpose: To construct a table for upper and lower limits of the 95% critical range for changes in word recognition scores obtained with monosyllabic word lists (of lengths 10, 25, 50, and 100 words) using newly available methods. Although such a table has been available for nearly 30 years (A. R. Thornton & M. J. M. Raffin, 1978), the earlier table was constructed by calculation and used an approximation to the variance of the difference score between 2 administrations of word lists of identical size. It has been used clinically, reproduced, and recommended for use by clinicians in handbooks and textbooks.

Method: The new table was created using computer simulation of the relevant distributions and a direct estimate of the variance of the difference score between 2 tests, calculated using the simulated results.

Results: The new table differed from the previous table in 23% of entries. Critical ranges were both narrowed (82%) and expanded (18%). No range changed by more than 1 word correct in any direction. The original table was most accurate for list sizes of 25 words each.

Conclusion: Using the new table will provide more accurate estimates of the 95% critical range for successive administrations of word recognition tests.