Improving the Reading Comprehension of Middle School Students With Disabilities Through Computer-Assisted Collaborative Strategic Reading

This study investigated the effects of computer-assisted comprehension practice using a researcher-developed computer program, Computer-Assisted Collaborative Strategic Reading (CACSR), with students who had disabilities. Two reading/ language arts teachers and their 34 students with disabilities participated. Students in the intervention group received the CACSR intervention, which consisted of 50-min instructional sessions twice per week over 10 to 12 weeks. The results revealed a statistically significant difference between intervention and comparison groups’ reading comprehension ability as measured by a researcher-developed, proximal measure (i.e., finding main ideas and question generation) and a distal, standardized measure (i.e., Woodcock Reading Mastery Test, Passage Comprehension). Effect sizes for all dependent measures favored the CACSR group. Furthermore, a majority of students expressed positive overall perspectives of the CACSR intervention and believed that their reading had improved. UNDERSTANDING AND LEARNING FROM TEXT IS at the heart of reading. As students progress through the grades, they are increasingly required to draw on their reading comprehension skills to learn from text (Williams, 1998). One of the most vexing problems facing middle and secondary school teachers today is that many students come into their classrooms without the requisite knowledge, skills, and disposition to read and comprehend the materials placed before them (Rippen & Brewer, as cited in Snow, 2002, p. iii). Students who are successful at meaning-making are able to monitor their understanding and to use various strategies to resolve problems and improve their comprehension. However, many students with learning disabilities (LD) have not developed this metacognitive awareness or the ability to skillfully apply comprehension strategies (Baker & Brown, 1984; Flavell, 1977). Over the years, researchers and their collaborators have developed numerous strategies to help struggling readers interact with text in ways that improve their understanding. The effectiveness of these strategies for students with LD has been well documented (Gersten, Fuchs, Williams, & Baker, 2001; Pressley, Johnson, Symons, McGoldrick, & Kurita, 1989; Snow, 2002; Swanson, 1999; Weisberg, 1988; Wong, 1985). From such empirical evidence, we know that comprehension strategy lessons should include explicit instruction in the targeted strategies (Swanson, 1999) and incorporate careful modeling and the provision of extensive feedback to students (Gersten et al., 2001). COLLABORATIVE STRATEGIC READING Collaborative Strategic Reading (CSR) was designed to facilitate reading comprehension for students with reading diffi© 2006 Hammill Institute on Disabilities. All rights reserved. Not for commercial use or unauthorized distribution. at PENNSYLVANIA STATE UNIV on April 11, 2008 http://rse.sagepub.com Downloaded from 236 R E M E D I A L A N D S P E C I A L E D U C A T I O N Volume 27, Number 4, July/August 2006 effects on students’ reading comprehension. Reciprocal teaching was originally designed to improve comprehension for students who can decode but who have difficulty with comprehension. A research synthesis on reciprocal teaching revealed a median effect size favoring reciprocal teaching. Particularly, with researcher-developed comprehension measures, reciprocal teaching was associated with the large median effect size of .88 (Rosenshine & Meister, 1994). COMPUTER-ASSISTED INSTRUCTION Computer-assisted instruction (CAI) can provide teachers with a tool for enhancing teaching and learning in their classrooms. CAI has the potential to offer students with LD self-paced, individualized instruction that includes immediate feedback and multiple opportunities for practice (Hall, Hughes, & Filbert, 2000; Lewis, 2000; MacArthur & Haynes, 1995; Rieth & Semmel, 1991; Woodward et al., 1986). Students generally find CAI to be quite motivating, yet Wissick and Gardner (2000) cautioned that to maximize the benefits of technology, students with disabilities should not be left to their own devices but should receive assistance as needed. Hall et al. (2000) reviewed 17 studies on CAI in reading interventions for students with LD. They noted that 3 of these studies focused on strategy instruction and included improving reading comprehension as a goal (Bahr, Kinzer, & Rieth, 1991; Keene & Davey, 1987; Woodward et al., 1986). Woodward et al. (1986) found significant differences favoring CAI. On the other hand, neither Bahr et al. (1991) nor Keene and Davey (1987) found significant differences between those students who used a software program and those who did not. In their review of meta-analyses in special education, Forness et al. (1997) described CAI as an intervention that “shows promise” in effectively helping students, rather than as an intervention that we know “works” (p. 6). In general, studies using CAI as a provider of teaching practices (e.g., providing the main ideas or definitions) have demonstrated significant improvements in reading comprehension (Horton, Lovitt, Givens, & Nelson, 1989; MacArthur & Haynes, 1995), whereas studies using CAI as a simple tool in the classroom (e.g., providing text on the screen) did not yield significant improvements in reading comprehension for students with reading difficulties (Elkind, Cohen, & Murray, 1993; Farmer, Klein, & Bryson, 1992). These findings suggest that effective reading CAI programs should provide effective, specific comprehension instruction. COMPUTER-ASSISTED COLLABORATIVE STRATEGIC READING Given the importance of helping students with LD become more efficient at comprehending, the potential of reading comprehension strategy instruction for achieving this goal, culties (Klingner & Vaughn, 1999; Klingner, Vaughn, & Schumm, 1998). CSR is an adaptation of reciprocal teaching (Palincsar & Brown, 1984) and includes many features associated with effective instruction (e.g., collaborative group work, interactive dialogue, clearly specified procedures). In CSR, students learn prereading, during-reading, and postreading strategies. Before reading, they preview by brainstorming what they already know about a topic and predicting what they think they will learn. During reading, they monitor their comprehension and apply fix-up strategies to help them figure out unknown words when comprehension breaks down. This strategy is called click and clunk. During reading, they also get the gist by identifying the most important ideas about a topic in a section of text. After reading, students wrap up by generating questions and reviewing the key ideas they have learned. Initially, the teacher presents the strategies to the whole class using modeling and think-alouds. After the students have developed proficiency using the strategies, they are then divided into collaborative groups, in which each student performs a defined role to implement the strategies collaboratively while learning from expository text. CSR was first implemented with 26 Latino middle school students with LD who were also English language learners (Klingner & Vaughn, 1996). The results were promising— even students who had very poor decoding skills made improvements in reading comprehension. In the next investigation of CSR, Klingner et al. (1998) provided strategy instruction in fourth-grade inclusive classrooms. Students who used CSR made statistically significantly greater gains than students in a comparison condition on the GatesMacGinitie Reading Test (MacGinitie & MacGinitie, 1989) and demonstrated equal proficiency in their knowledge of the social studies content. Separate data for students with LD were not available in this study. In a more recent study, Klingner, Vaughn, Arguelles, Hughes, and Ahwee (2004) examined teachers’ year-long implementation of CSR in fourth-grade inclusive settings. Students in CSR classrooms improved significantly in reading comprehension when compared with students in comparison classrooms. For students with LD, there was no statistically significant difference between the two groups on the comprehension measure, although the result was promising for students with LD in the CSR condition (standardized mean difference = .40). Although a limited number of studies were conducted with students with LD to examine the effects of CSR, there appears to be considerable research support for several features of CSR. First, Gersten et al. (2001) identified two instructional components associated with improved comprehension for students with LD, and CSR incorporates these two components: (a) the use of small, interactive groups (e.g., applying CSR strategies in small, collaborative groups) and (b) the teaching of specific formats for students’ generation of questions about what they read (e.g., wrap-up in CSR). Second, the use of reciprocal teaching, which significantly influenced the development of CSR, has resulted in positive © 2006 Hammill Institute on Disabilities. All rights reserved. Not for commercial use or unauthorized distribution. at PENNSYLVANIA STATE UNIV on April 11, 2008 http://rse.sagepub.com Downloaded from 237 R E M E D I A L A N D S P E C I A L E D U C A T I O N Volume 27, Number 4, July/August 2006 and the challenges faced by classroom teachers when trying to teach comprehension strategies, models of strategy instruction that teachers find feasible and students find engaging remain a necessity. One promising CAI model for addressing these issues is Computer-Assisted Collaborative Strategic Reading (CACSR; Kim, 2002). CACSR was designed to build on what is known about the critical features of effective comprehension strategy and CAI instruction. Thus, CACSR provides students with an interactive learning environment intended to maintain their interest while teaching them how to apply comprehension strategies as they read expository text passages. CACSR provides an individualized learning pace, choices in learning paths and reading passages, and reading level options. The CACSR intervention also includes feedback and correction procedures through CACSR’s builtin function of recording student performance. Teachers and students can monitor and evaluate student progress using the recorded data, and students receive immediate, corrective feedback based on their performance. In a previous investigation, Kim (2002) taught high school students with LD how to use CACSR and compared the reading comprehension gains of these students with those of students in a language arts resource class. The results revealed that both groups improved significantly in reading comprehension during the 12 weeks of the study, with no statistically significant differences between groups. Effect sizes, however, favored the CACSR group—that is, the pretest to posttest mean difference effect size for the CACSR group was .81, whereas it was .33 for the comparison group. Students generally expressed favorable opinions about CACSR but did offer suggestions for improvement. The purpose of the current study was to investigate the effects of CACSR on the comprehension of middle school students with LD and to examine the perceptions of participating students and teachers regarding the efficacy of CACSR. This study builds on previous work with CACSR in several ways. First, students worked in pairs to increase the interaction between students and to facilitate discussions, rather than each student interacting individually with the CACSR program, as in the previous study. Research findings have suggested that working with peers promotes interactive dialogue about text, thus encouraging students with LD to think while they read (Gersten et al., 2001). Also, Swanson and Hoskyn (1998) found that the use of small, interactive groups was associated with improved academic outcomes for students with LD, regardless of domain. Thus, we incorporated the partnering of students as a means of promoting the comprehension of students with LD. Second, the trained teacher and trained research assistant provided supplemental, explicit instruction in the comprehension strategies based on the students’ data (obtained from the CACSR program) at the beginning of each lesson. Explicitness in teaching comprehension strategies is one important factor that influences comprehension outcomes for students with reading difficulties (Rand Reading Study Group, 2002). Many students with LD struggle to understand how to use comprehension strategies when these strategies are presented in an implicit fashion, and they often fail to apply the learned strategies to a new task. Thus, teachers must respond to students’ needs for explicit instruction in comprehension strategies (Gersten & Carnine, 1986; Gersten et al., 2001). Third, a proximal measure of the intervention (i.e., the CSR measure) and a standardized comprehension measure (i.e., the Woodcock Reading Mastery Test–Revised Passage Comprehension subtest) were used. In the previous study, the standardized comprehension measure was used as the sole dependent measure. Given that reading comprehension has been documented as one of the most difficult components of reading to measure, we believed it was appropriate to use both a proximal measure of the comprehension strategies taught to students in the study (i.e., identifying the main idea and generating questions about what had been read) and a standardized measure to better understand students’ comprehension outcomes. Finally, a larger sample (n = 34) than that used in the previous study (n = 23) was included to increase statistical power.