This cross-sectional study examined eye movements during reading across grades in students with differing levels of reading efficiency. Eye-movement recordings were obtained while students in grades 2, 4, 6, 8, 10, and 12 silently read normed grade-leveled texts with demonstrated comprehension. Recordings from students in each reading rate quartile at each grade level were compared to characterize differences in reading rate, number of fixations, number of regressions, and fixation durations. Comparisons indicated that students in higher reading rate quartiles made fewer fixations and regressions per word, and had shorter fixation durations. These indices of greater efficiency were also characteristic of students in upper as compared to lower grades, with two exceptions: (a) between grades 6 and 8, fixations and regressions increased while reading rates stagnated and fixation durations continued to decline, and (b) beyond grade 6 there was relatively little growth in the reading efficiency of students in the lower two reading rate quartiles. These results suggest that declines in fixation duration across grades may in part reflect broader maturational processes, while higher fixation and regression rates may distinguish students who continue to struggle with word recognition during their high school years.

No abstract

This study examined silent reading rates (SRRs) in relation to students' estimated academic vocabulary grade levels (EVGLs) and comprehension accuracy (Comprehension Items Correct; compIC). Analyses were based on data from 288,934 students in grades 2-12 who completed an adaptive silent reading assessment that yielded measures of the three variables of interest. Silent reading rate was measured while students read five 150-to 300word passages. Each student's initial passage difficulty was aligned with their EVGL. Each passage was followed by five comprehension questions, such that in total, students could answer up to 25 comprehension items correctly. Twolevel Multilevel Models (MLMs) were fitted to evaluate SRR in relation to EVGL, compIC, and their interactions. The final MLM included the random intercept and three random slopes for the two level-1 predictors (schoolmean-centered EVGL as the focal predictor and school-mean-centered com-pIC as the moderator) and their interactions. Results indicated that: (a) the fixed effect of higher EVGL on SRR was positive and significant, (b) the fixed effect of higher compIC on SRR was negative and significant, and (c) there was a significant interaction indicating that the relationship between schoolmean-centered EVGL and SRR grew stronger as school-mean-centered compIC increased. These results suggest that vocabulary knowledge and SRR increase in concert among students with good comprehension, whereas SRRs measured in the absence of good comprehension are less meaningful and may indicate inadequate skills or insufficient motivation to fully comprehend what is being read.