New standardised texts in four European languages for assessing reading performance

Abstract

Abstract: : Purpose: To develop standardised texts for longer paragraph reading, because the available texts use single sentences only. Methods: The texts were designed by linguistic experts in English, Finnish, French and German. The texts were matched with respect to length (median 830 characters), word frequency ( 0.0001% or higher) and syntactic complexity (Syntactic Prediction Locality Theory of Gibson) on the level of 6th grade reading material. 100 normally–sighted native speaking volunteers (25 per language; age18 to 35 years) were tested with 10 texts each presented in randomised order. Oral reading times were measured with a stopwatch and errors were marked. After excluding inhomogeneous data sets and a Huynh–Feldt correction, an ANOVA of the reading times was performed. Results: The homogeneity of the reading material is shown by a quite low intra–individual variability of the reading times. Due to different reading habits, there is a considerable inter–subject variation. However, all the reading speed values lie in the range of fluent reading. Mean values (standard deviation) in characters per minute are: English 1234 (49.0), Finnish 1263 (58.8), French 1214 (50.7), German 1126 (39.2); due to different word lengths this corresponds to a reading speed in words per minute: English 228 (16.7), Finnish 161 (8.4), French 197 (11.1), German 179 (8.2). In each language we found a set of texts that showed no significant difference of reading time values in the ANOVA. Conclusions: We developed and tested a set of homogenous, comparable, and standardised texts in four European languages (English, Finnish, French, German) as a tool for follow–up tests and international studies in the field of reading and low–vision research.

Keywords: reading • low vision • clinical research methodology