Paper short abstract:

The purpose of this study is to examine the context validity of the reading fluency measurement tool. From the results of a pilot study was conducted on 308 learners, we made 10 sets of 'Japanese texts' and '10 content comprehension questions' with high validity at the elementary completion level.

Paper long abstract:

The purpose of this study is to examine the validity (context validity) of the reading fluency measurement tool under development, based on the social cognitive framework (Weir, 2005). In recent years, extensive reading and speed reading, which are expected to improve reading fluency, are attracting attention in the Japanese language education field. However, a few studies have demonstrated these effects. To begin with, it has been pointed out that the problem is that there is no valid tool that can measure reading fluency (e.g. Watanabe, Xu, Yamashita, Yokoyama, Oidaira, 2015). Therefore, in this study, we first made 10 sets of three-language versions (Japanese, English, and Chinese) of 'Japanese texts' and '13 four-choice multiple-choice content comprehension questions' at the elementary completion level. While preparing this, we sought opinions from experts, and consulted prior research (Quinn, Nation and Millett, 2007) regarding English speed reading tools. Next, a pilot study was conducted on a total of 308 Japanese language learners from various backgrounds (from those who had completed a beginner-level class to those studying at an advanced level, and from over 15 countries). The survey was conducted by explaining its method, handing out books, giving a practice test, and having participants fill out an information sheet and a consent form. We gathered data about reading speeds, content comprehension question scores, evaluations of the text content in terms of familiarity, difficulty, and interestingness. From the results of the item analysis using descriptive statistics of these data and the classical test theory, we were able to select 10 content comprehension questions with high validity (with an item discrimination index and bilinear correlation coefficient greater than .40). In this presentation, by comprehensively examining not only language content parameters, but also the process of developing a reading fluency measurement tool and a pilot survey, we propose a better method of enforcement, under which this tool will be used for moving forward. In the future, we will collect longitudinal data, and, along with continuing to verify its validity, advance the development of an intermediate-level reading fluency measurement tool.