Abstract
Purpose: :
Standardized texts are necessary to assess reading performance. A set of multiple equivalent texts is needed for repeated measurements and equated texts across languages are required for multi-language studies. Paragraphs are preferable to single sentences for more precise speed measurement and judgement of fluency and mistakes. In a previous European Research-Project (AMD-READ) we developed such texts in 4 different languages (International Reading Speed Texts, IReST). The aim of the here presented study was to develop texts in more languages to provide this tool for a wide range of users and countries.
Methods: :
Ten texts were designed for 17 languages each by a linguistic expert and adapted for length, content, difficulty and linguistic complexity (in Arabic, Chinese, English, Finnish, French, German, Hebrew, Italian, Japanese, Dutch, Polish, Portuguese, Swedish, Slovenian, Spanish, Russian, and Turkish). Then the texts were used to measure reading speeds of 436 normally sighted native speakers (18-35 years, 25 per language, 36 in Japanese), presented at a distance of 40 cm and size 1 M, i.e. 10 point Times New Roman. Reading time (aloud) was measured by stop watch. Reading speeds were calculated in 4 different units: texts/min, words/min, syllables/min and characters/min.
Results: :
Reading speed in texts/min showed the lowest coefficient of variation (CV)(9.6%) - followed by words/min (11.9%), syllables/min (21.6%) and characters/min (27.2%). When Hebrew, Arabic and the non-alphabetic languages are excluded, char/min showed the lowest CV (6.1%). The choice of unit depends on the language structure and the specific question being investigated. Variability of reading speeds between the 10 texts in each language is mainly caused by differences among subjects (75-93%), whereas the texts account only for an average of 11.5%.
Conclusions: :
The IReST (2nd ed) can now be provided in 17 languages (www.amd-read.net). It closes a gap in the diagnostics of reading performance and offers the advantages of standardized assessment of reading speed, comparability of results before and after interventions and is a useful tool for multi-language studies. IReST is suitable for a wide field of applications: low vision patients, neurological reading disorders, developmental dyslexia and studies with normal subjects.
Keywords: reading • low vision • clinical (human) or epidemiologic studies: systems/equipment/techniques