Purchase this article with an account.
Koichi Oda, Madoka Ohnishi, Terumi Otsukuni, Aoi Takahashi, Michiko Sugiyama, Seiji Yamagami; Effects of Length of Reading Materials on Key Parameters of Reading Speed Function. Journal of Vision 2016;16(12):1421. doi: https://doi.org/10.1167/16.12.1421.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose: Reading evaluation has been gradually accepted as a clinically important vision test in low vision rehabilitation. There are several versions and reading materials varies from unrelated random words to a paragraph of text (approx. 130 words). Purpose of this paper is to investigate into the effect of length of reading materials on two key parameters, location of reading function(LRF) which is closely related to critical print size(CPS) and maximum reading speed(MRS). Methods: In one experiment, reading materials were both 4-character clauses and 30-character sentences in Japanese. In the other experiment, they were random combinations of a Kanji and a Kana, 4-character clauses, and 10-character short sentences. In an experimental session, one type of reading material was tested. Different reading content was presented in each trial and participants were asked to read aloud as quickly and precisely as possible. A session started with the largest character size of 61.93 min of arc and proceeded to 0.1 log smaller trials until no single character was read. Reading time and errors were recorded and reading speed function was plotted for each reading material condition. Two parameters were estimated using Weibull fitting. Thirty three Japanese for one experiment and 28 for the other participated and all had corrected/uncorrected normal visual acuity. Results and Discussion: Combining results of two experiments, there was a moderately good agreement in LRF estimates among different reading material conditions with 0.5 through 0.7 correlations. As for MRS, there was a very systematic increase along with the length of reading material. We conducted a mixed effect model analysis on MRS data with a model, reading time (msec) = latency + unit reading time (msec/char) x character length. The model fit well (R2=0.99). We conclude that these findings help us compare and convert results of different reading evaluations.
Meeting abstract presented at VSS 2016
This PDF is available to Subscribers Only