Wolfgang Radner1, Stephan Radner2, Gabriela Diendorfer3. 1. Austrian Academy of Ophthalmology, Mollgasse 11, A-1180, Vienna, Austria. wolfgang.radner@inode.at. 2. Austrian Academy of Ophthalmology, Mollgasse 11, A-1180, Vienna, Austria. stephan.radner@gmx.net. 3. Department of Phoniatrics, Medical University of Vienna, Währinger Gürtel 18-20, 1090, Vienna, Austria. gabriela.diendorfer@meduniwien.ac.at.
Abstract
PURPOSE: To investigate the reliability, validity, and statistical comparability of long paragraphs that were developed to be equivalent in construction and difficulty. METHODS: Seven long paragraphs were developed that were equal in syntax, morphology, and number and position of words (111), with the same number of syllables (179) and number of characters (660). For validity analyses, the paragraphs were compared with the mean reading speed of a set of seven sentence optotypes of the RADNER Reading Charts (mean of 7 × 14 = 98 words read). Reliability analyses were performed by calculating the Cronbach's alpha value and the corrected total item correlation. Sixty participants (aged 20-77 years) read the paragraphs and the sentences (distance 40 cm; font: Times New Roman 12 pt). Test items were presented randomly; reading length was measured with a stopwatch. RESULTS: Reliability analysis yielded a Cronbach's alpha value of 0.988. When the long paragraphs were compared in pairwise fashion, significant differences were found in 13 of the 21 pairs (p < 0.05). In two sequences of three paragraphs each and in eight pairs of paragraphs, the paragraphs did not differ significantly, and these paragraph combinations are therefore suitable for comparative research studies. The mean reading speed was 173.34 ± 24.01 words per minute (wpm) for the long paragraphs and 198.26 ± 28.60 wpm for the sentence optotypes. The maximum difference in reading speed was 5.55 % for the long paragraphs and 2.95 % for the short sentence optotypes. The correlation between long paragraphs and sentence optotypes was high (r = 0.9243). CONCLUSIONS: Despite good reliability and equivalence in construction and degree of difficulty, a statistically significant difference in reading speed can occur between long paragraphs. Since statistical significance should be dependent only on the persons tested, either standardizing long paragraphs for statistical equality of reading speed measurements or increasing the number of presented paragraphs is recommended for comparative investigations.
PURPOSE: To investigate the reliability, validity, and statistical comparability of long paragraphs that were developed to be equivalent in construction and difficulty. METHODS: Seven long paragraphs were developed that were equal in syntax, morphology, and number and position of words (111), with the same number of syllables (179) and number of characters (660). For validity analyses, the paragraphs were compared with the mean reading speed of a set of seven sentence optotypes of the RADNER Reading Charts (mean of 7 × 14 = 98 words read). Reliability analyses were performed by calculating the Cronbach's alpha value and the corrected total item correlation. Sixty participants (aged 20-77 years) read the paragraphs and the sentences (distance 40 cm; font: Times New Roman 12 pt). Test items were presented randomly; reading length was measured with a stopwatch. RESULTS: Reliability analysis yielded a Cronbach's alpha value of 0.988. When the long paragraphs were compared in pairwise fashion, significant differences were found in 13 of the 21 pairs (p < 0.05). In two sequences of three paragraphs each and in eight pairs of paragraphs, the paragraphs did not differ significantly, and these paragraph combinations are therefore suitable for comparative research studies. The mean reading speed was 173.34 ± 24.01 words per minute (wpm) for the long paragraphs and 198.26 ± 28.60 wpm for the sentence optotypes. The maximum difference in reading speed was 5.55 % for the long paragraphs and 2.95 % for the short sentence optotypes. The correlation between long paragraphs and sentence optotypes was high (r = 0.9243). CONCLUSIONS: Despite good reliability and equivalence in construction and degree of difficulty, a statistically significant difference in reading speed can occur between long paragraphs. Since statistical significance should be dependent only on the persons tested, either standardizing long paragraphs for statistical equality of reading speed measurements or increasing the number of presented paragraphs is recommended for comparative investigations.
Entities:
Keywords:
Reading performance; Reading speed; Reading test
Authors: Konrad R Koch; Philipp S Muether; Manuel M Hermann; Robert Hoerster; Bernd Kirchhof; Sascha Fauser Journal: Graefes Arch Clin Exp Ophthalmol Date: 2011-09-08 Impact factor: 3.117
Authors: Antonia M Joussen; Sandra Joeres; Nader Fawzy; Florian M A Heussen; Helene Llacer; Jan C van Meurs; Bernd Kirchhof Journal: Ophthalmology Date: 2007-03 Impact factor: 12.079
Authors: Robert P Finger; Peter Charbel Issa; Rolf Fimmers; Frank G Holz; Gary S Rubin; Hendrik P N Scholl Journal: Invest Ophthalmol Vis Sci Date: 2008-11-07 Impact factor: 4.799