Classwide peer tutoring : three experiments investigating the generalized effects of increased oral reading fluency to silent reading comprehension

Reading deficits are commonly displayed by students referred for school psychology services. The ability to read fluently (i.e., rapidly and accurately) correlates with reading comprehension (Martson, 1989), the goal of reading. Thus, a common instructional goal for students experiencing reading deficits is to increase their reading fluency. Classwide Peer Tutoring (CWPT) has been shown to improve the oral reading rates and level of reading comprehension of students with learning disabilities and academic deficits. However, students typically do not read aloud for comprehension. As well, with more skilled readers, comprehension rates may provide a more educationally valid measure of reading proficiency than oral reading fluency or comprehension accuracy alone (Skinner, Neddenriep, Bradley-Klug, & Ziemann, 2002). Therefore, this three-experiment study was conducted to investigate the effects of CWPT on students' rate and level of both oral and silent reading comprehension rates. This study extended research on the effectiveness of CWPT with students having reading skill deficits in three ways: (1) the study incorporated a rate of reading comprehension measure (Skinner, 1998); (2) the study assessed near generalization of oral reading fluency to oral reading comprehension; and, (3) the study assessed far generalization to silent reading comprehension. The study consisted of three, singlesubject design experiments. Experiment 1 used an alternating treatment design to compare two, 6th-grade, general-education students' rates and level of oral reading comprehension when tutored (experimental passages) to their rates and level of oral reading comprehension when not tutored (control passages). Both students evidenced deficits in reading (i.e., standardized reading achievement scores in at least one area below the 20th percentile), and their instructional reading level was found to be 1 year below grade level (i.e., fifth grade). Significant differences were found on all dependent measures (i.e., comprehension rate, comprehension level, oral reading fluency) across participants favoring tutored passages. Two additional students with reading skill deficits participated in Experiment 2. Although Experiment 2 used the same design, students were only exposed to a portion of the passage during tutoring; thus, near generalization to the remainder of the tutored passage was assessed. Significant differences were not found in the direction of CWPT partially- tutored passages across any of the measures. Correlations between oral reading fluency and both level of reading comprehension and rate of reading comprehension also were calculated. Across the four subjects, correlational data showed that oral reading fluency correlated more strongly with comprehension rates (r = .87) than with comprehension levels (r = .57). Although the correlation ·data suggest that reading comprehension rates may be a more reliable and sensitive measure than merely assessing comprehension accuracy, the two experiments suggest that it may be difficult to obtain measures of comprehension (levels or rates) unless students read entire intact passages. Experiment 3 used a multiple baseline design to determine the generalized effects of increased oral reading fluency to students' level and rate of silent reading comprehension of intact, novel passages (i.e., passages not previously tutored). Seven, 6th-grade students having reading deficits participated in the experiment. Two participants began the intervention phase, participating in CWPT 4 days per week for 30-min sessions. All participants were assessed once weekly. Pairs of students were added to the intervention as improvement in silent reading comprehension was verified. Results demonstrated an increase in average number of words read correctly per minute for six students and a reduction in average number of errors made per minute across all seven participants favoring the intervention condition. The majority of participants, however, evidenced similar trends across baseline and intervention conditions, making conclusions regarding the treatment effect inconclusive. Improvement in the ability to accurately comprehend a passage read silently was demonstrated for one participant while silent reading comprehension rate also reflected improvement in both speed and accuracy of comprehension for the same student. Limitations of this research and implications for future research are discussed.