Measurement of reading speed with standardized texts: a comparison of single sentences and paragraphs

Elke Karin Altpeter, Tobias Marx, Nhung Xuan Nguyen, Aline Naumann, Susanne Trauzettel-Klosinski

We examined the influence of text length (single sentence versus a paragraph of several sentences) on the repeatability of reading speed measurements in normal-sighted subjects. We compared reading speeds for the German versions of the Radner charts (single sentences of 14 words each) and the International Reading Speed Texts (IReST) charts (paragraphs, on average 132 words) in 30 normal-sighted elderly subjects aged 51-81 years (mean 64.5 years ± 7.2 SD). Three texts each of both lengths were read aloud in random order. The influence of text length (single sentence or paragraph) and text sample (each single text) on reading speed was calculated by a regression model and Bland-Altman analysis. Mean reading speed (words per minute) showed no significant difference for single sentences (170 wpm ± 33 SD) and paragraphs (167 wpm ±31 SD). Differences in reading speeds within one type of reading material were higher between single sentences than between paragraphs. Correlation coefficients between speeds were higher for paragraphs (r = 0.96-0.98) than for single sentences (r = 0.69-0.78). Variations between reading speeds for three texts of each length were markedly lower for paragraphs than for single sentences: (median, interquartile range [IQR]): 6.7, IQR 13.9; 3.0, IQR 8.3; -2.0, IQR 9.7 versus -8.8, IQR 29.6; 15.6, IQR 29.4; 22.7, IQR 19.4, respectively. Since reading speeds assessed with paragraphs show lower variance among texts than those for single sentences, they are better suited for repeated measurements, especially for long-term monitoring of the course of reading performance and for assessing effects of interventions in subjects with reading disorders.