Precision, reliability and application of the Wilkins Rate of Reading Test

Background: The Wilkins Rate of Reading Test (WRRT) enables rapid measurement of reading speed using text passages that have no semantic content and demand minimal word recognition skills. It is suited to applications where the primary interest is in the influence of visual and ocular motor factors on reading rate. Methods: We obtained estimates of precision and reliability of WRRT from four data samples (A-D) collected independently by the authors: (A) n = 118 adults; (B) n = 90 adults; (C) n = 787 children; (D) n = 134 children. Each participant was asked to read aloud as quickly and accurately as possible, for 1 min, and results were recorded as number of words read correctly per minute (wcpm). Results: Estimates of precision are given by the within-subjects standard deviation s_w, and reliability by the intraclass correlation coefficient for single measurements r₁. For each sample these estimates were (A) s_w = 11.5 wcpm, r₁ = 0.85; (B) s_w = 3.8 wcpm, r₁ = 0.98; (C) s_w = 6.7 wcpm, r₁ = 0.93; (D) s_w = 6.2 wcpm, r₁ = 0.94. Conclusion: The reliability of WRRT reflects large variation in reading rate between individuals compared to within-individual variability, indicating that it is a good test for discriminating differences in reading speed between individuals. The precision of the test varies from 3.8 wcpm to 11.5 wcpm among samples, and the pooled value of 7.2 wcpm, provides a basis for setting a population-wide criterion for minimum detectable change of reading rate in individuals over time. Nevertheless, a preferable way of monitoring change in an individual would be to use a criterion determined from estimates of that individual's baseline variation in WRRT scores.