Purpose
To evaluate the test-retest reliability of the MNREAD charts in children with normal or impaired vision and to compare MNREAD maximum reading speeds to that of a longer, grade appropriate passage.
Methods
72 children with vision impairment (VI) and 35 children with normal vision grades 1-12 had a comprehensive ocular examination, completed MNREAD chart-based testing, paragraph reading using the Jerry Johns Basic Reading Inventory (BRI) and the Slosson Intelligence Test. The MNREAD and BRI were each administered at 2 study visits, 1 to 3 weeks apart. Children wore their habitual correction for all testing and were given the choice of regular or large print versions of the BRI. Reading acuity, critical print size and maximum reading rate (MRR) were determined for the MNREAD. Reading rate at the child’s independent reading level was determined for the BRI. Statistical analyses included Intra-Class Correlations (test-retest reliability), Pearson correlations (between test associations), Bland-Altman plots and paired t-tests.
Results
Children with and without VI were mostly male (64% vs. 57%), white (74% vs. 86%), and had average intelligence (mean 105 ± 16 vs. 109 ± 12). There were no statistical differences in demographics. MNREAD test-retest reliability was greatest for reading acuity (0.99 overall, 0.96 VI, 0.94 control) followed by Critical Print Size (0.91 overall, 0.73 VI, 0.52 control). Reliability was modest for MRR (0.39 overall, 0.89 VI, 0.37 control). Assessment of the Bland-Altman plots showed that there was greater variability for MRR when reading speeds were faster. Correlations between MNREAD MRR and BRI reading rate were stronger for VI than control (0.39 overall, 0.81 VI, 0.38 control). Children, (grouped into grades 1-3, 4-6, 7-9 & 10-12) with VI had significantly slower MNREAD MRR than controls in the same grade groups (p<0.01 for all comparisons). Reading rate on the BRI was not different between VI and control grade groups. Reading rates were faster on the MNREAD test than the BRI for both VI and control (p< 0.01), but were more similar between the two tests for children with VI than for children with normal vision.
Conclusions
MNREAD chart based testing is a reliable test for children with and without VI, however it yields faster reading speeds compared to paragraph testing. Differences are greatest for children with normal vision.