An examination of third graders’ reading behavior, strategy use, and accuracy on an expository text: An eye movement examination

An examination of third graders’ reading behavior, strategy use, and accuracy on an expository text: An eye movement examination

First Author: Katherine Binder -- Mount Holyoke College
Additional authors/chairs: 
Scott P. Ardoin; Kathryn Tremblay; Amani Talwar; Elizabeth Tighe
Keywords: Comprehension, Eye-tracking, Individual Differences, Item characterictics, Reader-text interactions
Abstract / Summary: 

PURPOSE: Typically, reading comprehension tests merely yield an accuracy score, which limits knowledge of how individual characteristics and item-level characteristics influence performance. The purpose of this study was to investigate how individual characteristics of 3rd grade children are related to reading behavior and reading comprehension, as measured by accuracy on multiple choice (MC) questions. METHOD: We examined third graders’ eye movement behavior (n= 78) as they read an expository text and then answered corresponding MC questions. We also collected data on decoding, word identification, reading fluency, reading comprehension, and working memory abilities. We were interested in which individual characteristics were related to performance, what sort of differences we might observe between the reading and answering behaviors of students who answered each question correctly and students who did not, and whether question type (literal vs. inferential) contributed to performance. RESULTS: Correct responders often did not re-read the text, but when they did, they were able to find the relevant information quickly and efficiently. Incorrect responders, however, were not efficient at finding the relevant information in the text, and the incorrect responders were more likely to spend greater time reading the questions and answers rather than consulting the text. In addition, depending on question type (literal vs inferential) different component skills differentiated correct and incorrect responders. CONCLUSIONS: Student characteristics and item-level characteristics are related to test-taking behavior, these differences can lend important information related to previous research findings that RC tests produce varying results within and between students.