Programs that Work
Cooperative Integrated Reading and Composition
Children Succeeding in School
Students performing at grade level or meeting state curriculum standards
Age of Child
Early Childhood (0-8)
Middle Childhood (9-12)
Type of Setting
Type of Service
Type of Outcome Addressed
Cognitive Development / School Performance
Evidence Level (What does this mean?)
Cooperative Integrated Reading and Composition (CIRC) is a school-based program that targets reading, writing, and language arts in grades 2 through 6. The three principle program elements are direct instruction in reading comprehension, story-related activities, and integrated language arts/writing instruction. Each student is paired with another student and then assigned to a group of students at the same or different reading level. These learning teams work cooperatively on program-related activities. All activities follow a cycle that involves teacher presentation, team practice, peer pre-assessment, additional practice, and testing. Students are encouraged to cooperate and help one another, because studentsí scores on individual assessments are summed to form team scores.
Several years ago, CIRC was adapted to form one component of Reading Wings, a comprehensive reading program in the Roots and Wings whole-school reform model. The Roots and Wings model consists of elementary school age learning programs, reading and language arts instruction, tutoring, family support and integrated services, social studies and science instruction, and mathematics instruction. CIRC has been incorporated into a primer-level reading program called Reading Wings.
Students in grades 2 through 6
Madden et al. (1986a) studied CIRC among 461 third- and fourth-grade students in 21 classes in a suburban Maryland school district. The sample consisted of 11 treatment classes from six schools that were matched on California Achievement Test (CAT) Total Reading scores with ten control classes from four schools. Treatment and control teachers volunteered to participate in the 12-week study, with treatment classes implementing CIRC, and control teachers continuing to use their usual teaching methods and curriculum materials until the following academic year. Teaching methods utilized by the control teachers involved dividing each class into three reading groups by reading level, giving traditional language arts and writing instruction to the whole class, and using workbooks during follow-up time. Outcome measures at 12 weeks included performance on reading comprehension, reading vocabulary, spelling, language expression, and language mechanics subtests on the CAT. In addition, studentsí writing samples were analyzed for ideas, the level of organization, and language mechanics (i.e., punctuation and capitalization, spelling, word choice, syntax). An analysis of baseline differences between the groups revealed that the control group had better scores than the treatment group on the CAT Total Language test and on the pretest writing samples that indicated studentsí knowledge of language mechanics. To account for these differences, studentsí scores on the CAT pretest were used as covariates in all analyses.
In a similar study, Madden et al. (1986b) assessed the effects of CIRC in a sample of 450 third- and fourth-grade students from 22 classes in a suburban Maryland school district. The sample included nine treatment classes from four schools and 13 control classes from five schools. The treatment and control classes were matched on CAT Total Reading and Total Language scores, and the research team also attempted to match the groups on ethnic background and socioeconomic status by selecting classes from schools in the same or similar neighborhoods. An analysis of baseline differences between the two groups revealed no significant differences between them. Both treatment and control teachers volunteered to participate in the study, and CIRC was implemented for 24 weeks. Outcome measures included performance on reading comprehension, reading vocabulary, language expression, and language mechanics subtests from the CAT. Student writing samples were also analyzed for ideas, organization, and mechanics. Additionally, a subsample of students was given informal inventories of their reading knowledge (consisting of lists of words to be defined and oral reading passages from the Durrell Analysis of Reading Difficulty test). The student subsample consisted of six pairs of students (one from the treatment group and one from the control group in each pair) from each class. The students in two pairs were randomly selected from the top third of their class according to reading level, students in two pairs were randomly selected from the middle third of their class, and students in two pairs were randomly selected from the bottom third of their class. This random selection of the students in each matched pair provided for representation from all levels of students in each classroom.
Stevens et al. (1989) studied CIRC in a sample of 529 students from 29 classes in grades 2 through 6 in a suburban Maryland school district. The sample of 15 treatment classes from three schools was matched on CAT Total Reading scores with 14 control classes from three schools. Zero percent to 10 percent of the student population in these schools were racial/ethnic minorities, and 6 to 13 percent were disadvantaged students as determined by the number receiving free or reduced-price lunches. No significant pretest differences were found between the treatment and control groups on CAT Total Reading or Total Language scores or on an index of reading awareness. Outcome measures included performance on reading vocabulary, reading comprehension, language mechanics, and language expression subscales from the CAT, and scores on an index of reading awareness (which measured studentsí awareness of comprehension strategies).
CIRC was also assessed by Stevens and Durkin (1992) in a sample of 1,223 sixth-grade students from six middle schools in an urban Maryland school district. The sample included 20 treatment classes from three schools and 34 control classes from three schools, who were matched on CAT Total Reading pretest scores. Racial/ethnic minority students in the schools ranged from 27 percent to 99 percent of the student populations, and the percentage of disadvantaged students ranged from 38 percent to 77 percent. Outcome measures included reading vocabulary and reading comprehension subscales from the CAT.
Students in another state participated in a study by Bramlett (1994), which examined CIRC among 392 third-grade students in eight school districts in rural southern Ohio. The treatment group consisted of 198 students from nine classes, and the control group consisted of 194 students from nine classes. Teachers volunteered to participate in CIRC, and their teaching experience ranged from 4 to 21 years. The teachers were assigned to treatment and control groups based on geographic representation, and an attempt was made to equally distribute teachers according to years of teaching experience. Outcome measures included CAT subscales of reading vocabulary, reading comprehension, Total Reading, and word-analysis.
Finally, Stevens and Slavin (1995) studied 1,299 students in grades 2 through 6 in a suburban, working-class school district in Maryland. The sample consisted of 31 treatment classes in three schools, and 32 control classes in four schools. The treatment-group and control-group schools were matched on socioeconomic variables, and the student body in both groups of schools had similar ethnic backgrounds. Minority and disadvantaged students in both schools ranged from zero to10 percent and from 6 to13 percent of the student body, respectively. An analysis of pretest scores between treatment and control groups indicated that the groups differed significantly in language and reading abilities. To account for these differences, pretest scores were used as covariates in the analyses. Outcome measures included reading comprehension, reading vocabulary, language mechanics, and language expression subscales from the CAT, and scores on an index of reading awareness.
Key Evaluation Findings
The Madden et al. (1986a) study of 461 third and fourth graders found the following:
- Treatment-group students scored significantly higher than control group students on four of the five CAT subscales, including reading comprehension, reading vocabulary, language expression, and spelling. No significant differences were found between groups on the fifth subscale, language mechanics.
- As measured by grade equivalents, the CIRC classes outscored the control classes by 0.41 grade equivalents for reading comprehension, 0.34 grade equivalents for reading vocabulary, 0.56 grade equivalents for language expression, and 0.71 grade equivalents for spelling.
- The assessment of student writing samples indicated that the treatment group outscored the control group in organization of writing. No significant differences were found between groups for ratings of writing language mechanics or ideas.
- The CIRC group significantly outscored the control group on reading comprehension, language expression, and language mechanics, with effect sizes of 0.35, 0.29, and 0.30, respectively. (An "effect size" is a quantified measure of the effectiveness of a treatment. Conventional values of effect sizes are: small = .20, medium = .50, and large = .80). No significant differences in reading vocabulary were found between the two groups.
- In grade equivalents, the CIRC classes outscored the control classes by 0.66 on reading comprehension, 0.64 on language expression, and 0.66 on language mechanics.
- The CIRC group scored marginally higher on the writing sample analysis of ideas, with an effect size of 0.31. No significant differences in the assessments of writing organization or mechanics were found between the two groups.
- The CIRC subsample of students outscored the control group subsample on all oral reading measures, including word recognition (with an effect size of 0.64), word analysis (effect size 0.47), grade placement (effect size 0.55), time on a common paragraph (effect size 0.62), and number of errors in a common paragraph (effect size 0.44).
- Significant differences favoring the treatment group on three of the four CAT tests, including reading vocabulary (effect size 0.29), reading comprehension (effect size 0.25), and language mechanics (effect size 0.28). There were no significant differences between the treatment and control groups for language expression.
- Treatment effects varied by grade level, i.e., while there were significant program effects for students in the second through fifth grades, there were no significant differences between the two groups in sixth-grade scores.
- The treatment group scored significantly higher than the control group on the index of reading awareness.
- The treatment group outscored the control group on reading comprehension, with an effect size of 0.11.
- There were no significant differences between the two groups in their reading vocabulary.
- Additional analyses of studentsí race and gender did not indicate a differential program impact on racial/ethnic minorities or on students of either gender.
- The treatment group scored significantly higher than the control group in reading comprehension, with a small effect size of 0.11.
- There were marginally significant but small effects favoring the treatment group on tests of Total Reading (effect size 0.07) and word analysis (effect size 0.10).
- No significant differences were found between the two groups on their reading vocabulary scores.
- Scores for the bottom 33 percent of the CIRC group were significantly higher than the scores for the bottom 33 percent of the control group. Treatment-group students in the bottom-third percentile outperformed control group students on all four subtests, including vocabulary (effect size 0.30), reading comprehension (effect size 0.33), CAT Total Reading (effect size 0.35), and word analysis (effect size 0.56).
- Significant differences between the two groups were not found for students in the middle or upper percentile levels.
- After the first year of the CIRC program, treatment group classes outscored control group classes on reading vocabulary (effect size 0.22) and reading comprehension (effect size 0.24). No significant program effects were found for language mechanics or language expression.
- After the second year of the program, CIRC students outscored control group students on tests of their reading vocabulary (effect size 0.20), reading comprehension (effect size 0.26), and language expression (effect size 0.26). No significant differences were found between the two groups for language mechanics.
- Treatment-group students outscored control-group students on the index of reading awareness.
Public and private elementary schools
No information on CIRC funding available at this time.
A unique aspect of CIRC is its focus on cooperative team learning. However, as is also done in traditional reading programs, CIRC teachers use anthologies, basal readers, and/or novels as instructional tools. Students work in pairs on activities that include reading to one another; predicting how stories will end; identifying characters and settings; summarizing stories for each other; writing responses to stories; and practicing spelling, decoding words, and vocabulary. Students also write drafts, revise, and edit each otherís work, and prepare to "publish" their writing. Students then work in teams of four to five to better understand main story ideas and to improve other general comprehension skills. Measurement of individual studentsí contributions to their teams are based on quiz scores and independently written compositions.
CIRC requires two days of training for classroom teachers and school administrators, plus review of materials including story anthologies, basal readers, and novels. Additional training and follow-up sessions are recommended.
Issues to Consider
This program received a "promising" rating. The evaluations of CIRC are quasi-experimental in design, and utilized reasonably convincing comparison groups along with statistical analyses that attempted to account for any pre-existing differences between the experimental and control groups. The evaluations demonstrate that treatment-group students scored significantly higher than control-group students on standardized tests of reading comprehension, reading vocabulary, language expression, language mechanics, and spelling. Analyses of studentsí writing samples indicate that CIRC students exhibited superior performance as compared with control students on organization of their writing and presentation of written ideas.
The Bramlett (1994) study suggests that CIRC may have a greater impact on students with poorer reading skills than on students with average or strong reading skills. Study results showed that students in the bottom third of the group at baseline experienced significant program impacts when compared with the control group, whereas students in the middle and upper thirds of the treatment group did not demonstrate any differences when compared with control-group students.
One possible issue of concern with several of the evaluations (i.e., Madden et al., 1986a; Madden et al., 1986b; Stevens et al., 1989; Bramlett, 1994; and Stevens & Slavin, 1989) is that classroom teachers volunteered to participate in the experiment. It is possible that these teachers are intrinsically different than teachers who would opt not to participate in the studies, e.g., they may be larger risk-takers, or be particularly concerned about their studentsí reading difficulties. Beyond participation in a study on CIRC, such innate differences may lead to teaching practices and classroom styles that differ from other kinds of teachers. The results of these evaluations therefore do not necessarily generalize to classrooms with teachers who would choose not to participate in such an experiment.
It should be noted that the program developers were program evaluators on five of the six studies cited in this program description.
Nancy Madden, President
Success for All Foundation
200 W. Towsontown Blvd.
Baltimore, MD 21204
Phone: 410 616 2330
Fax: 410 324 4440
None at this time.
Bramlett, Ronald K., "Implementing Cooperative Learning: A Field Study Evaluating Issues for School-Based Consultants,"
Journal of School Psychology,
Vol. 32, No. 1, 1994, pp. 67-84.
Madden, Nancy A., Robert J. Stevens, and Robert E. Slavin, A Comprehensive Cooperative Learning Approach to Elementary Reading and Writing: Effects on Student Achievement (Report No. 2), Center for Research on Elementary & Middle Schools, Baltimore, Md.: The Johns Hopkins University, 1986a.
Madden, Nancy A., Robert J. Stevens, and Robert E. Slavin, Reading Instruction in the Mainstream: A Cooperative Learning Approach (Report No. 5), Center for Research on Elementary & Middle Schools, Baltimore, Md.: The Johns Hopkins University, 1986b.
Stevens, Robert J., and Robert E. Slavin, "Effects of a Cooperative Approach in Reading and Writing on Academically Handicapped and Nonhandicapped Students," The Elementary School Journal, Vol. 95, No. 3, 1995, pp. 241-262.
Stevens, Robert J., and Scott Durkin, Using Student Team Reading and Student Team Writing in Middle Schools: Two evaluations (Report No. 36), Center for Research on Effective Schooling for Disadvantaged Students, Baltimore, Md.: The Johns Hopkins University, 1992.
Stevens, Robert J., Robert E. Slavin, and Anne Marie Farnish, A Cooperative Learning Approach to Elementary Reading and Writing Instruction: Long-Term Effects (Report No. 42), Center for Research on Elementary & Middle Schools, Baltimore, Md.: The Johns Hopkins University, 1989.