Cooperative learning in the classroom: Contingencies, group interactions, and students with special...

19
Journal of Behavioral Education, Iiol. 2, No. L 1992, pp. 53-71 Cooperative Learning in the Classroom: Contingencies, Group Interactions, and Students with Special Needs Merith A. Cosden, Ph.D., 1,2 and Thomas G. Haring, Ph.D. 1 Accepted: July 6, 1991. Action Editor: Nirbhay N. Singh Cooperative learning activities play a significant role in the schools. As such, these activities should be of interest to applied behavior analysts concerned with educational interventions. A number of factors appear to contribute to the impact of cooperative group activities, including the level of individual accountability, group contingencies, and the types of interactive behaviors in which students engage during group activities. This paper reviews current research on cooperative learning, focusing on the relationships between student behaviors, contingencies of reinforcement and group outcomes. The social and academic skills with which students enter cooperative learning activities, and the impact of these skills on students' behavior within groups and on the social and academic outcomes of groups is considered. Problems associated with the failure of cooperative groups are also addressed. The research potentials for applied behavior analysts are discussed. KEY WORDS: cooperative learning; applied behavior analysis; disabled students; cooperative groups; review. Interest in cooperative learning activities within the educational main- stream has grown substantially over the past ten years. This is due, in part, to the promise of cooperative instruction for affecting a number of student outcomes. Cooperative learning activities have been touted for their posi- tive impact on academic performance (Johnson, Maruyama, Johnson, Nelson, & Skon, 1981), social behavior (Gelb & Jacobson, 1988) and the 1Associate Professor, Department of Education, University of California, Santa Barbara. 2Correspondence should be addressed to either author, Department of Education, University of California, Santa Barbara, CA 93106. 53 1053-0819/92/0300-0053506.50/0 1992 Human Sciences Press, Inc.

Transcript of Cooperative learning in the classroom: Contingencies, group interactions, and students with special...

Journal of Behavioral Education, Iiol. 2, No. L 1992, pp. 53-71

Cooperative Learning in the Classroom: Contingencies, Group Interactions, and Students with Special Needs

Merith A. Cosden, Ph.D., 1,2 and Thomas G. Haring, Ph.D. 1

Accepted: July 6, 1991. Action Editor: Nirbhay N. Singh

Cooperative learning activities play a significant role in the schools. As such, these activities should be of interest to applied behavior analysts concerned with educational interventions. A number of factors appear to contribute to the impact of cooperative group activities, including the level of individual accountability, group contingencies, and the types of interactive behaviors in which students engage during group activities. This paper reviews current research on cooperative learning, focusing on the relationships between student behaviors, contingencies of reinforcement and group outcomes. The social and academic skills with which students enter cooperative learning activities, and the impact of these skills on students' behavior within groups and on the social and academic outcomes of groups is considered. Problems associated with the failure of cooperative groups are also addressed. The research potentials for applied behavior analysts are discussed.

KEY WORDS: cooperative learning; applied behavior analysis; disabled students; cooperative groups; review.

Interest in cooperative learning activities within the educational main- stream has grown substantially over the past ten years. This is due, in part, to the promise of cooperative instruction for affecting a number of student outcomes. Cooperative learning activities have been touted for their posi- tive impact on academic performance (Johnson, Maruyama, Johnson, Nelson, & Skon, 1981), social behavior (Gelb & Jacobson, 1988) and the

1Associate Professor, Department of Education, University of California, Santa Barbara. 2Correspondence should be addressed to either author, Department of Education, University of California, Santa Barbara, CA 93106.

53

1053-0819/92/0300-0053506.50/0 �9 1992 Human Sciences Press, Inc.

54 Cosden and Haring

integration of students with diverse backgrounds and skills (Johnson & Johnson, 1981, Johnson, Johnson, Warring, & Maruyama, 1986; Sharan, 1980). Group instruction which utilizes cooperative goals and group con- tingencies has been encouraged for mainstreaming students with disabilities as well as for managing diverse student needs within the classroom (Conway & Gow, 1988).

This paper reviews current research on factors that contribute to the effectiveness of cooperative learning, and presents an analysis of the po- tential contributions that behavior analysis can provide in the development of this important instructional technology. This review focuses on the iden- tification of effective group behaviors and the relationship of these behav- iors to group incentives and outcomes. Variations in students' responses to group contingencies, and the impact of student behavior on the success or failure of the group, will also be addressed.

Cooperative learning procedures have been defined in several ways. Johnson and Johnson (1982), for example, differentiate cooperative learn- ing experiences from competitive or individualistic experiences on the basis of how goals are attained. In an individualistically structured learning situ- ation students' goal attainments are unrelated and independent, while in a competitive learning situation students' goal attainments are negatively correlated (i.e., one student can achieve his or her goal only when the other student(s) with whom that child is competing does not reach their goal). In a cooperatively structured learning situation, however, students' goal at- tainments are positively correlated; goals are achieved, or not, by the "group" and rewards, when earned, are given to all group members. Thus, any group activity in which students' rewards are linked could be considered a cooperative learning experience.

Slavin (1983) delineates two components of cooperative learning methods: cooperative task structure and cooperative incentive structure. Cooperative task structures are situations in which two or more individuals are allowed, encouraged or required to work together on the same task. Two types of task structures used in cooperative learning are: (a) task spe- cialization, in which each group member is responsible for a unique com- ponent of the activity, and (b) group study methods, in which all group members work together. A cooperative incentive structure, on the other hand, indicates that all group members share the same consequences as a function of succeeding or not succeeding on a task. Incentive structure is further categorized on the basis of whether rewards are provided to indi- viduals within the group on the basis of their own performance, or to ev- eryone in the group on the basis of their group performance. Group performance can be rewarded either by averaging each team member's score, or by using a group product (e.g., one worksheet answered by the

Cooperative Learning 55

group). Thus, Slavin distinguishes six cooperative learning methods: (1) group study with individual reward, (2) group study with group reward for individual learning, (3) group study with group reward for group product, (4) task specialization with individual reward, (5) task specialization with group reward for individual learning, and (6) task specialization with group reward for group product.

Cooperative learning should be of particular interest to behavior ana- lysts working in education for at least two reasons. First, the instructional and reinforcement procedures used in cooperative learning are, on the whole, compatible with behavior analytic constructs. These constructs in- clude: (a) the use of group contingencies such that the reinforcement that the individual receives is affected by the performance of the group as a whole. This component is central to cooperative learning because the group incentive supports the efforts of students within the group to teach each other the critical skills needed to complete an assignment; (b) the use of individual contingencies which may be employed to ensure that the per- formance of each student is assessed and rewarded. This can be pro- grammed through individual testing and grading group activities; and (c) the interactive nature of the cooperative group means that much of the learning and instruction takes place within a social dialogue of group inter- action, rather than the more passive responding of traditional didactic instruction.

Cooperative learning is typically associated with contingency struc- tures in which student rewards are dependent upon successful group out- comes. Beyond this commonality, substantive differences in cooperative learning activities are apparent. Differences in the nature of group incen- tives, the roles and responsibilities of group members, and student inter- action patterns within groups are among the many variations noted in the literature. Based on prior research, it is not clear what components of cooperative learning models are most critical for positive students out- c o m e s .

A second reason that behavior analysts should further participate in this research is to increase the impact of behavior analysis in general edu- cation. The potential impact of current behavioral teaching technology is constrained by limitations in the contexts within which much research is conducted. Although examples of group interventions can be found (e.g., Pfiffner & O'Leary, 1987, Pigott, Fantuzzo, & Clement, 1986), much of the behavior analytic research relevant to general education classrooms is conducted within a one-on-one instructional format (e.g., Gettinger, 1985; Lancioni, Hoogeveen, Smeets, Boelens, & Leonard, 1989; Roberts, Nelson, & Olson, 1987). Of the research published in the Journal of Applied Behavior Analysis over the last 10 years that investigated academic

56 Cosden and Haring

responding, 15 of 22 studies (68%) were demonstrations that involved one- to-one interventions, or pullouts of dyads of students. While these demon- strations promote the development of a technology to remediate individual learning problems, the demonstration of these techniques in situations that are removed from the overall context of group instruction in the classroom limits the application of these techniques.

EFFECTIVENESS OF COOPERATIVE INSTRUCTION

A number of factors appear to influence the effectiveness of coop- erative instruction (see Cotton & Cook, 1982; Johnson, et al., 1981; Sharan, 1980; Slavin, 1983 for reviews). These factors include the behavior in which students engage during group activities, the impact of task and incentive structure on student behavior, differences in student behavior as a result of learning disabilities or group composition, and whether or not the team succeeds in meeting its goals. These factors are further explored below.

Effective Behavior within Groups

The relationship of student interactions during group instruction to academic and social outcomes has been studied from a variety of perspec- tives (Wilkinson, 1988-89). Smith, Johnson, and Johnson (1981), for ex- ample, examined the impact of controversy (i.e., presenting two opposing sides to an issue) and concurrence (reaching a consensus without arguing) on performance within cooperative groups. The investigators reported that when groups were encouraged to engage in controversy they scored higher on a written test of reading comprehension, spent more time viewing an optional film, checked out more materials from the library, presented more ideas and opinions, and rated themselves on a likert scale as more liked by peers, than when groups were encouraged to seek concurrence. While intriguing, follow up studies in this area have not appeared in the literature.

Typically, however, studies in the cooperative learning literature pro- vide little information regarding student behavior during group activities. Many of the studies which focus on group interactions have not been in- cluded in reviews of the cooperative learning literature because these stud- ies do not include control groups against which to contrast the effects of cooperative learning activities. Rather, these studies focus on differences in student interaction patterns within groups, identifying student behaviors

Cooperative Learning 57

which correlate with positive and negative achievement outcomes as a func- tion of group activities.

Webb and her colleagues (e.g., Webb, 1982; 1989; Webb, Ender, & Lewis, 1986) have studied the relationship between student interactions during small group instruction and scores on achievement tests in a variety of areas in mathematics. Through observation of student behavior during group activities, the investigators have identified initiation-response se- quences which correlate with students' test scores. In particular, they have found that receiving explanations in response to one's questions, and pro- viding information in response to questions are positively associated with scores on math achievement tests, while receiving no response or a non- explanatory response to questions appears to be negatively related to test performance.

Impact of Task and Incentive Structure on Group Behavior

Individual accountability associated with group rewards appears to be a significant predictor of group success (Slavin, 1983, 1991). Slavin postu- lates that cooperative group incentives combined with individual account- ability improve students' academic performance because group members are motivated, "to make the group successful by encouraging each other to e x c e l . . . Group members pay a great deal of attention to one another's efforts and socially reinforce efforts that help the group to achieve its goal" (Slavin, 1983, p. 442). Slavin contends that group study, without contingent rewards, is unlikely to have a positive impact on student achievement. He assumes that rewards for cooperative achievement of goals are necessary to promote appropriate interactive behaviors.

There are several weaknesses in his analysis, however. First, Slavin's review of research includes only studies in which cooperative learning con- ditions are contrasted with other types of instruction (i.e., non-group in- structional controls). Thus, studies that assess differences in student interactions within cooperative groups, and which are more likely to identify variations in group behavior as function of personal characteristics, were not considered in his analysis. Although Slavin (1983, 1991) does not in- clude most of Webb's research in his reviews for the reasons noted above, he speculates that it is this type of interactive behavior which is encouraged by individual accountability and group incentives. It is interesting to note, however, that the students in Webb's studies are provided with both indi- vidual and group rewards (i.e., grades) after group participation. Further, while all of the students in her studies experience the same contingencies, some students engage in more productive interactive behaviors than do

58 Cosden and Haring

others. It is critical to better understand the variables that control the dif- ferential impact of cooperative learning procedures on the behavior of in- dividuals within such groups.

Another question raised by Slavin's analysis involves the links between contingencies and interactive behavior. An assumption is made that ma- nipulating reinforcement conditions is sufficient to assure student engage- ment in productive interactive study behaviors. This does not allow for the possibility that simply increasing reinforcement to perform some tasks may not be sufficient when students lack the requisite social interaction skills to engage productively in the group learning.

At issue too, is whether cooperative learning structured in the manner described by Slavin is the only method by which these interactive behaviors can be elicited. For example, in Slavin, Madden, and Leavey (1984), Team Assisted Individualization (TAI), a cooperative learning procedure in which students were asked to study together and monitor each other's perform- ance, and Individualized Instruction (II), a procedure in which students were asked to study alone, were compared to a "traditional" eclectic ap- proach to instruction. There were no significant differences in the academic gains made by students across groups as measured by the average number of correct items on all math achievement tests taken by team members. However, students in both TAI and II groups reported greater satisfaction with the instruction they received than did students in the control groups. This was measured by student responses to questions scored as likert scales; these questions addressed whether or not students liked their math class (e.g. "This math class is the best part of my school day") as well as their self concept in math (e.g. "I'm proud of my math work in this class"). Based on anecdotal evidence, Slavin et al. posited that these satisfaction outcomes were a function of student interactions under both conditions; that is, stu- dents were informally observed to engage in higher levels of task related social interactions during both of the TAI and II treatments. Slavin et al. speculated that, "in schools, duration and quality of contact between stu- dents may be more important in predicting social relationships than the reward structure under which contact occurs" (p. 441). To date, the litera- ture in this area has not sufficiently defined and measured the impact of variation in contingencies on social responding, achievement, or satisfaction with instruction.

It is our contention that cooperative learning activities are effective in large part due to their impact on student behavior during group activi- ties. While cooperative goal structures may increase the likelihood that stu- dents will engage in effective group interactions, other factors, including the social and academic skills of students, need to be considered in pre- dicting the occurrence of these behaviors across instructional conditions.

Cooperative Learning 59

Individual Differences in Student Behavior During Cooperative Group Activities

As students' academic and social histories vary, so do the behavioral repertoires which they bring to cooperative learning activities. Students dif- fer in terms of their level of participation, their social assertiveness, and their individual efforts and success within group activities. Kerr and Bruun (1983), for example, describe a "free rider" effect that can occur within groups; one student may do most of the work and the others assume a "free ride". A higher ability student within a group may also reduce his contribution to the group if he perceives that others are taking advantage of his contributions (Kerr, 1983). Inequities in individual efforts are more likely to occur in disjunctive tasks (e.g., problem solving), where the group outcome depends on the most able member, or conjunctive tasks (e.g., team reading activities) in which group success depends on the skills of the least able member. These problems are less likely to occur when the task is struc- tured t0 require maximal participation from each team member. Assuring appropriate individualization of goals within groups, however, is both labor intensive and more difficult to maintain.

The roles that students play within groups appear to be influenced by a number of factors, including the student's academic and social skills, peer perceptions of that student's skills, and the group composition. Dembo and McAuliffe (1987), for example, found that older students and those who were perceived as having higher ability by their peers were more likely to dominate group discussion and influence group decision making than were other students. The authors suggest that if higher ability students dominate group activities, students perceived by their peers as less able may be deprived of the beneficial effects of both giving and receiving help. It should be noted that in this study perceptions of relative "ability" were induced by the experimenters and were not assessed by direct quantitative performance measures.

Students with Disabilities

Concerns about the effectiveness of cooperative learning for students who have identified disabilities have emerged from the literature. Although cooperative learning activities are often discussed in relation to mainstreaming children with handicapping conditions, few studies have systematically examined the effects of cooperative instruction for special populations. In two recent reviews (Lloyd, Crawley, Kohler, & Strain, 1988; Tateyama-Sniezek, 1990), a combined total of only 15 studies were

60 Cosden and Haring

cited in which cooperative learning activities included children with either mild or severe disabilities. These reviews present equivocal findings with regard to the impact of cooperative group activities for children with disabilities.

In terms of social effects, cooperative learning does appear to facili- tate social acceptance of students with disabilities as measured by students' self reports of wanting to sit near and work with children with disabilities (Lloyd et al., 1988). These findings, however, may be confounded by the fact that most studies are designed so that groups are successful in reaching their goals (e.g., Harris & Covington, 1987). The confound between group success and group instruction that is found in most studies has significant implications for students with disabilities participating in mixed ability co- operative groups. While sharing success with teammates can have positive social effects, students with disabilities may be more vulnerable than other students for receiving the "blame" for group failures (see Ames, 1981 below).

With regard to academic outcomes, the literature is particularly am- biguous. In studies in which results were analyzed separately for disabled and non-disabled students, higher achievement test scores as a function of cooperative group participation often have not been found (e.g., Cosden, Pearl, & Bryan, 1985; Madden & Slavin, 1983; Johnson & Johnson, 1982, 1985; Johnson, Johnson, DeWeerdt, Lyons, & Zaidman, 1983). There are a number of reasons that students with certain types of disabilities might not perform well within cooperative groups. For exam- ple, students with learning disabilities, for example, while identified on the basis of poor academic performance, also exhibit behavioral problems which can impact their group experiences. Students with learning disabili- ties often demonstrate poor pragmatic skills (Dudley-Marling, 1985). They also display a number of problems specifically related to poor peer group communications, including difficulties in initiating and maintaining con- versations (Bryan, Donahue, Pearl, & Sturm, 1981), taking a leadership role during group activities, and being persuasive (Bryan, Donahue, & Pearl, 1981).

The relationship between group study behavior and poor academic performance for students with learning disabilities has been suggested in several studies. Cosden et al. (1985), for example, found that students with learning disabilities who performed poorly on a reading comprehension test on the basis of the fluency of their verbal recall of the story, their verbal responses to eight short answer questions, and their number of correct re- sponses to nine multiple choice questions, also engaged in fewer question and answer behaviors during their study period than did students in dyads with higher scores on these tests. In a subsequent study, Cosden and

Cooperative Learning 61

Goldman (1990) analyzed the social interactions of dyads of students with learning disabilities during a microcomputer-based writing activity. Overall, student interactions were negatively correlated with story production as measured by the number of word types and tokens produced. Students ap- proached the task with higher levels of assertive, demanding behavior than with requests for peer information or collaboration. Even when sequences of requesting and receiving information did occur, they did not have the predicted positive effect on story fluency. This was interpreted as a function of the low level explanatory responses typically given to questions; re- sponses were frequently given in the form of direct answers to questions, and were also sometimes incorrect. This is consistent with Webb's (1982) analysis of instances in which students' responses within groups did not improve academic performance on math achievement tests. Webb observed this to occur when students were younger (e.g., Peterson, Janicki, & Swing, 1981) or when students had lower abilities (Webb & Kenderski, 1984) and could not give clear explanations in response to requests for help:

Group composition

Another factor to consider in developing cooperative activities is the composition of the group. Slavin (1987), in a review of group practices, concludes that heterogeneous grouping patterns should be maintained at most times, with homogeneous grouping in specific skill areas (e.g., reading or math). Webb (1982) found that middle-ability students performed better on math achievement tests after working in relatively homogeneous groups rather than in mixed-ability groups. This was correlated with the increased use of explanatory responses by middle-ability students in homogeneous ability groups. In the Cosden et al. (1985) study nondisabled boys did not engage in as many question and answer behaviors or perform as well on a reading comprehension test when they worked with partners who had learning disabilities as when they worked with partners who did not have learning disabilities. However, other studies (e.g. Madden & Slavin, 1983; Johnson & Johnson, 1982; Johnson et al., 1983) have found that students with disabilities, while not directly benefiting from group instruction, did not disrupt the functioning of other students in the group.

There are several ways in which students who are deficit in the aca- demic or social skills might detract from group activities. A student might provide incorrect information which is then utilized by the group. A student who is off task might influence others to be off task. A student who is less involved in the task may cause other students to feel that they are being taken advantage of, thereby reducing their task involvement. A student with

62 Cosden and Haring

limited skills might require additional help from peers, which may not in- crease the productivity of the group. Students who need help with their social behavior m who have problems attending to others, asking questions or sharing ideas, may use group resources in ways which don't contribute to the academic performance of the group.

Clearly, more needs to be known about the impact of student social responding on the successful functioning of cooperative groups. Concerns are raised for students with disabilities, particularly if those disabilities im- pede their ability to successfully engage in behavior necessary to assure the success of cooperative group activities.

Teams that Fail

As indicated above, most studies published on cooperative learning assume successful cooperative team outcomes (Salomon & Globerson, 1989). Although groups can be structured to maximize their potential for success, it is reasonable to expect that cooperative groups will not always meet their goals, particularly if those goals require individual accountability. While individual accountability allows one to determine the impact of group activities on each student, this accountability also makes the group more vulnerable to failure. Inequalities in effort and participation by team members can have a negative impact on students' perceptions of each other even when group success is not impaired by the differential contributions of each student (Salomon & Globerson, 1989). As groups increase their level of individual accountability, however, the contributions of individuals within the group and the extent to which they meet their personal goals is likely to affect the success or failure of the group attaining its contin- gency. This has serious implications for students for whom problems in at- tention, social behavior, or cognitive skills make their personal success in group efforts questionable.

Ames (1981) examined the impact of success and failure on student's self and peer perceptions as a function of working under cooperative or competitive conditions. In a series of studies, the impact of individual per- formance levels, group success and reward structure (cooperative vs com- petitive) were studied in relation to self and peer perceptions. Self and peer perceptions were measured by student responses to four attribution questions relating to ability ("How smart do you think you were in solving the puzzles?"), effort ("How hard did you try to solve the puzzles?"), task difficulty ("How hard do you think the puzzles were to solve?"), and luck ("How lucky do you think you were in solving the puzzles?"). Each question was followed by nine circles of increasing size with end labels such as "not

Cooperative Learning 63

smart-very smart." Subjects were asked to respond to the same questions with regard to their own attributes and those of their partner. Students were randomly assigned to competitive, cooperative success, and coopera- tive failure conditions. Individual student performance levels were manipu- lated by providing students with puzzles to solve which varied in their solubility. Groups were composed so that each had a high and low per- forming student working together. Students worked independently, but si- multaneously, on their puzzles, and were aware of each other's performance during the task.

As predicted, Ames found that competitive goal structures enhanced the satisfaction of winning for high performing students, as measured by selection of one of five faces (smiling to frowning) which indicated how satisfied they were with their own performance, but also resulted in low ability ratings and low satisfaction ratings for the "losing" low performing students. Competitive goal structures accentuated differences in students' ratings of their abilities to perform the task, as well as their satisfaction with the task. High performers indicated that they were more deserving of reward than were low performers, and low performers concurred with these ratings.

Successful cooperative experiences led to high ratings of ability and satisfaction for both the high and low performers in these groups; in suc- cessful cooperative groups, students' ability ratings and ratings of satisfac- tion were similar despite their individual performance differences. High performers rated themselves and their low performing partners as similarly deserving of reward, although the low performing students indicated that their partners were more deserving. Unsuccessful cooperative group per- formance, however, resulted in different ability ratings by the high and low performing students within these groups; ability ratings and perceptions of satisfaction were significantly lower for the low performing student when their cooperative group failed than when it succeeded, although the stu- dents' performance was held equivalent across these conditions.

In a study designed to replicate and expand this work, Harris & Covington (1987) crossed goal structure (competitive or cooperative) with outcome (success or failure) and individual performance level (high or low). The authors found that success or failure on the task had a more significant impact on student ratings of ability, satisfaction, and deserving of reward than did the goal structure under which students experienced these out- comes. This was the case particularly for less able students. That is, when students with fewer skills were part of a successful group they were rated as having more ability and as being more deserving of reward than when their group was unsuccessful, even though the individual students' scores were the same in each condition. Low performers who were part of groups

64 Cosden and Haring

that failed were often blamed for the group failure. Other negative out- comes associated with the failure of cooperative groups to meet their goals include negative teacher perceptions, even toward students who make a positive contribution to the group (Ames & McKelvie, 1982).

IMPLICATIONS FOR RESEARCH

At present there is wide variation in the manner in which cooperative learning is provided in the schools, with little clarity concerning the effects of these variations (Damon & Phelps, 1989). The potency of cooperative groups for eliciting student interactions which enhance learning is often assumed, and has only rarely been assessed through direct observation. While cooperative group structures appear to support certain types of pro- ductive interactive behaviors, other factors, including the academic compe- tencies and social skills of individuals within the group, need to be considered in predicting group effects.

It is apparent that group success has a significant impact on student ratings of their skills and the skills of their peers. Peers rate less able stu- dents as satisfying to work with when their group is successful but not when they are unsuccessful. Further, groups may be successful, but may not pro- vide equally successful experiences for all team members. When students are given individual goals within a group it is assumed that each student's goal represents an equivalent level of challenge. Unless this difficult bal- ance is maintained, and individual goals vary in terms of their relative dif- ficulty, some students may be bored while others are overwhelmed.

The impact of group failure must also be considered, particularly with regard to the less able students in a group who may be blamed for the failure. Structuring groups to assure success may not always be plau- sible. Monitoring students within groups and establishing realistic individ- ual and group goals can reduce the likelihood of group failure. Methods for addressing group failure which also recognize individuals who success- fully met their goals or contributed positively to the group should be considered.

Being placed in a cooperative group may alone contribute to shaping the behavior of many students. In a study of group entry skills of popular and unpopular students, Gelb and Jacobson (1988), for example, found that unpopular students engaged in fewer disruptive behaviors during a coop- erative activity than during a competitive activity. Group contingencies, peer feedback, and the opportunity to model students with appropriate skills appear to provide some students with sufficient opportunities to learn new, and more effective, interactive behaviors.

Cooperative Learning 65

For students with disabilities, however, making opportunities available to interact with other students and to observe appropriate social behavior is not sufficient to alter behavior patterns (Gaylord-Ross & Haring, 1987). Some students need to be trained to engage in effective group behaviors in order to effectively utilize cooperative group experiences. There are data to support the use of short term training in modeling or direct instruction to develop cooperative behavior (Sagotsky, Wood-Schneider, & Konop, 1981). Generalization of effects in this study was found for older elemen- tary school students but not for younger students. Interventions designed to facilitate group skills need to be tested with special populations.

Finally, there is a need to clarify the goals of cooperative learning within a classroom, and to link training and rewards to those goals. Coop- erative groups have been used to enhance cognitive development, develop social skills, and integrate children with diverse academic and social needs. In establishing goals, the interaction of outcomes must also be considered. For example, the failure of the group to reach its academic goals may affect the social cohesion of the group, particularly for less able students. What- ever goals are established, monitoring student performance and developing training procedures to assist students to engage in effective behaviors are at least prerequisites for successful use of this technology.

Behavior Analysis and Cooperative Learning

This review has discussed a number of areas in which behavior analy- sis is well suited to contribute to the literature in cooperative learning and assist in the development of a more powerful technology for all learners, including those with learning challenges. In general, most demonstrations of cooperative learning procedures have not provided fine-grained meas- ures of social interaction within groups. In addition, the design of both group and individual contingencies has not been as systematic as behavior analyses could provide. The following areas seem well suited for the de- velopment of behavior analytic research programs.

Definition of contingencies

The bulk of research conducted in the area of cooperative learning has not come from a behavior analytic tradition. In many cases the de- scriptions of contingencies and rewards are not precise, and in some cases are omitted entirely from research reports. An important deficiency in this literature is the lack of individualization of rewards, or demonstrations that the rewards employed are in fact potentially reinforcing to the students.

66 Cosden and Haring

In most studies reviewed, the group or individual rewards employed con- sisted of grades on assignments, with no manipulation of other events dur- ing the school day that could potentially function as reinforcement. Behavior analysts could make a substantial contribution to this literature by increasing the technical adequacy of the definition of contingencies em- ployed, and by demonstrating procedures to select consequences that can be reasonably expected to function as reinforcers. This is an important first step in increasing our understanding of how the complex, multiple contin- gencies in cooperative learning affect academic performance, social inter- action, and social acceptance of all s tudents including those with disabilities.

Linkage between group contingencies and individual contingencies

A linkage between group contingencies and individual contingencies seems essential for cooperative learning (Slavin, 1983). An over-reliance on group contingencies, without individual accountability, and without a linkage between the two can lead to situations in which more able group members dominate interactions and simply give answers to less able mem- bers in order to complete the task. Group contingencies based on individual performance in testing situations assures that the learning of class material by all students is promoted.

Similarly, individual accountability assures that measures of individ- ual learning are taken. In many studies of cooperative learning, individual assessment of learning is not conducted (e.g., Johnson & Johnson, 1981). If the group contingency is based on all members of the group achieving individual goals, an incentive is created for members to help each other learn.

Not all individual accountability procedures, or group contingencies, will produce equal outcomes. Individual accountability procedures that sim- ply grade and reward each individual will probably not serve to increase learning to the same extent as when a direct linkage is created between individual performance and the group contingency. For example, if com- pleting a group assignment leads to a group reward (e.g., recess) while in- dividual accountability consists of test scores on unit tests and grades, there is a lack of linkage between the group contingency and the individual con- tingency. However, if the group contingency (e.g., access to a longer recess) is based on a sum of individual performances (assessed independently) then a linkage is created between individual learning and the instructive behavior of the group members. Behavior analysis provides methodologies to

Cooperative Learning 67

compare and contrast the effects of variation in contingency design on per- formance and social interaction within groups.

Direct instruction of social interaction within groups

Some group members, particularly those who are academically at risk for failure, may lack the social interaction skills necessary to maximally benefit from cooperative learning procedures (Morrison & MacMillan, 1983). Skills such as listening to others, accepting feedback, offering feed- back in a constructive manner, staying on task, accepting praise, and ap- propriately giving praise are important skills to engage in cooperative learning. These are precisely the skills that many students with learning disabilities and mild behavior problems lack in instructional contexts. Thus, specific skill training programs and behavioral interventions designed to be implemented within the context of cooperative learning groups would be an important contribution to the literature.

More able students in cooperative groups have an increased instruc- tional responsibility to model the processes that they use to solve problems for less able members. One of the major functions of group contingencies is to create a reward structure to increase the frequency of offering elabo- rated explanations, because, as Webb's (1982; 1989) research has shown, these response classes seem to be responsible for the success of this ap- proach on academic performance. This leads to other social skills training issues. Even students who are gifted academically do not necessarily have the social interaction repertoires and leadership skills to effectively provide elaborated explanations. Thus, the effectiveness of cooperative learning groups might be substantially enhanced if more able group members are taught and reinforced for offering elaborations. Skills such as explaining algorithms for problem solving, suggesting sources of information for re- ports, demonstrating solutions to problems, and verbalizing the reasons why facts are related, are potentially separate response classes that can be taught, modeled, and reinforced. Indeed, contingencies applied directly to these responses may be as powerful as the contingencies typically used to promote cooperative learning.

Promoting full inclusion

A current policy trend in special education is to promote the fuller inclusion of students with special learning needs within regular education classes (e.g., Sailor, Anderson, Halvorson, Doering, Filler, & Goetz, 1990). This policy trend has been recently re-defined to include the inclusion of

68 Cosden and Haring

students with moderate and severe disabilities, as well as students with mild disabilities in regular education academic instruction. Cooperative learning models, with the ability to individualize goals and tasks for students within groups, offer a means of achieving this goal. However, there have been few demonstrations of mechanisms and training activities to integrate stu- dents with disabilities into cooperative groups.

Research in this area will require two related approaches. First, stu- dents with disabilities will need interventions to teach appropriate social interaction skills in groups. These skills, for students with more severe dis- abilities, should include recognizing group members, greeting group mem- bers, and developing skills to partially participate in group activities. For mildly disabled students, asking appropriate questions, asserting ideas, and attending to others may also require instruction. Nondisabled group mem- bers will require a small repertoire of skills that are individualized for the needs of the students being integrated. For example, some students may require familiarization with nonsymbolic communication means, or sugges- tions for how to naturalistically prompt the student with disabilities to par- ticipate in the group. Research by Kohler and Greenwood (1990) indicates the importance of collateral peer supportive behaviors in maintaining the efforts of peers to engage in instructive activities, particularly when one member of the group has learning problems.

Summary

Cooperative learning models are becoming an increasingly important aspect of instruction in typical classrooms. Although the current research in this area shows the potential power of cooperative models in creating valued outcomes such as increased achievement, improvements in attitudes towards learning, and increased acceptance of students from diverse back- grounds, there is much that remains unknown about the factors that control the effectiveness of this instructional technology. We have argued that the tools of behavior analysis can be productive in enhancing the understanding of the factors that control the success of groups as well as contributing procedures to further enhance the potential of these procedures.

REFERENCES

Ames, C. (1981). Competitive versus cooperative reward structures: The influence of individual and group performance factors on achievement attributions and affect. American Educational Research Journal, 18, 273-287.

Cooperative Learning 69

Bryan, T., Donahue, M., & Pearl, R. (1981). Learning disabled children's peer interactions during a small-group problem-solving task. LearnhTg Disability Quarterly, 4, 13-22.

Bryan, T., Donahue, M., Pearl, R., & Sturm C. (1981). Learning Disabled children's conversational skills - - the "TV talk show". LearnhTg Disability Quarterly, 4, 250-259.

Conway, R., & Gow, L. (1988). Mainstreaming special class students with mild handicaps through group instruction. Remedial and Special Education, 9, 34-41.

Cosden, M., & Goldman, S. (1990). Learning handicapped students' interactions during a microcomputer-based group learning activity. Journal of Special Education Technology, 10, 220-232.

Cosden, M., Pearl, R., & Bryan, T. (1985). The effects of cooperative and individual goal structures on learning disabled and nondisabled students. Exceptional Children, 52, 103-114.

Cotton, J. & Cook, M. (1982). Meta-analyses and the effects of various reward systems: Some different conclusions from Johnson et al. Psychological Bulleth~, 92, 176-183.

Damon, W., & Phelps, E. (1989). Critical distinctions among three approaches to peer education. International Journal of Educational Research, 13, %19.

Dembo, M., & McAuliffe, T. (1987). Effects of perceived ability and grade status on social interaction and influence in cooperative groups. Journal of Educational Psychology, 79, 415-423.

Dudley-Marling, C. (1985). The pragmatic skills of learning disabled children: A review. Journal of Learning Disabilities, 18, 193-199.

Gaylord-Ross, R., & Haring, T. (1987). Social interaction research for adolescents with severe handicaps. Behavior Disorders, 12, 264-275.

Gelb, R., & Jacobson, J. (1988). Popular and unpopular children's interactions during cooperative and competitive peer group activities. Journal of Abnormal ChiM Psychology, 16, 247-261.

Gettinger, M., (1985). Effects of teacher-directed versus student-directed instruction and cues versus no cues for improving spelling performance. Journal of Applied Behavior Analysis, 18, 167-178.

Harris, A., & Covington, M. (1987). Cooperative team failure: A double threat for the low performer? Paper presented at the Annual Meeting of the American Educational Research Association, Washington, D.C.

Johnson, D., Johnson, R., Warring, D., & Maruyama, G. (1986). Different cooperative learning procedures and cross-handicap relationships. Exceptional Children, 53, 247-252.

Johnson, D., Maruyama, G., Johnson, R., Nelson, C., & Skon, L. (1981). The effects of cooperative, competitive, and individualistic goal structures on achievement: A meta-analysis. Psychological Bulletin, 89, 47-62.

Johnson, D., & Johnson, R. (1982). The effects of cooperative and individualistic instruction on handicapped and nonhandicapped students. Journal of Social Psychology, 118, 257-268.

Johnson, D., & Johnson, R. (1985). Mainstreaming hearing-impaired students: The effects of effort in communicating on cooperation and interpersonal attraction. Journal of Psychology, 119, 31-44.

Johnson, R., Johnson, D., DeWeerdt, N., Lyons, V., & Zaidman, B. (1983). Integrating severely adaptively handicapped seventh grade students into constructive relationships with nonhandicapped peers in science class. American Journal of Mental Deficiency, 87, 611-618.

Johnson, R., & Johnson, D. (1981). Building friendships between handicapped and nonhandicapped students: Effects of cooperative and individualistic instruction. American Educational Research Journal, 18, 415-423.

Kerr, N. (1983). Motivation losses in small groups: A social dilemma analysis. Journal of Personality and Social Psychology, 45, 819-828.

Kerr, N., & Bruun, S. (1983). Dispensability of member effort and group motivation losses: Free rider effects. Journal of Personality and Social Psychology, 44, 78-94.

Kohler, F. W., & Greenwood, C. R. (1990). Effects of collateral peer supportive behaviors within the classwide peer tutoring program. Journal of Applied Behavior Analysis, 23, 307-323.

70 Cosden and tlaring

Lancioni, G. E., Hoogeveen, F. R., Smeets, P. M., Boelens, H., & Leonard, S. N. (1989). Errorless discrimination of reversible letters: Superimposition and fading combined with an intervening response. The Psychological Record, 39, 373-385.

Lloyd, J., Crawley, E. P., Kohler, F., & Strain, P. (1988). Redefining the applied research agenda: Cooperative learning, prereferral, teacher consultation, and peer-mediated interventions. Journal of Learning Disabilities, 21, 43-52.

Madden, N. A., & Slavin, R. E., (1983). Effects of cooperative learning on the social acceptance of mainstreamed academically handicapped students. Journal of Special Education, 17, 171-182.

Morrison, G. M., & MacMillan, D. L. (1983). Defining, describing, and explaining the social status of mildly handicapped children: Update and discussion. In J. M. Berg (Ed.), Perspectives and progress in mental retardation: Volume 1 - Social, psychological, and educational aspects (pp. 51-60). Baltimore, MD: University Press.

Peterson, P. L., Janicki, T. C., & Swing, S. R. (1979). Ability X treatment interaction effects on children's learning in large group and small group approaches. American Educational Research Journal, 18, 453-473.

Pfiffner, L. J., & O'Leary, S. G. (1987). The efficacy of all-positive management as a function of the prior use of negative consequences. Journal of Applied Behavior Analysis, 20, 265-272.

Pigott, E. H., Fantuzzo, J. W., & Clement, P. W. (1986). The effects of reciprocal peer tutoring and group contingencies on the academic performance of elementary school children. Journal of Applied Behavior Analysis, 19, 93-98.

Roberts, R. N., Nelson, R. O., & Olson, T. W. (1987). Self-instruction: An analysis of the differential effects of instruction and reinforcement. Journal of Applied Behavior Analysis, 20, 235-242.

Sagotsky, G, Wood-Schneider, M., & Konop, M. (1981). Learning to cooperate: Effects of modeling and direct instruction. Child Development, 52, 1037-1042.

Sailor, W., Anderson, J. L., Halvorson, A. T., Doering, K., Filler, J., & Goetz, L. (1990). The comprehensive local school: Regular education for all students with disabilities. Baltimore, MD: Paul H. Brookes.

Salomon, G., & Globerson, T. (1989). When teams do not function the way they ought to. International Journal of Educational Research, 13, 89-99.

Sharan, S. (1980). Cooperative learning in small groups: Recent methods and effects on achievement, attitudes and ethnic relations. Review of Educational Research, 50, 241-271.

Slavin, R. (1983). When does cooperative learning increase student achievement? Psychological Bulletin, 94, 429-445.

Slavin, R. (1987). Ability grouping and student achievement in elementary schools: A best evidence synthesis. Review of Educational Research, 57, 293-336.

Slavin, R. (1991). Cooperative learning and group contingencies. Journal of Behavioral Education, L 105-115.

Slavin, R., Madden, N., & Leavey, M. (1984). Effects of cooperative learning and individualized instruction on mainstreamed students~ Exceptional Children, 50, 434-443.

Smith, K., Johnson, D., & Johnson, R. (1981). Can conflict be constructive? Controversy vs. concurrence seeking in learning groups. Journal of Educational Psychology, 73, 651-663.

Tateyama-Sniezek, K. (1990). Cooperative learning: Does it improve the academic achievement of students with handicaps? Exceptional Children, 56, 426-437.

Webb, N. (1982). Group composition, group interaction, and achievement in cooperative small groups. Journal of Educational Psychology, 74, 475-484.

Webb, N. (1989). Peer interaction and learning in small groups. International Journal of Educational Research, 13, 21-39.

Webb, N., Ender, P., & Lewis, S. (1986). Problem solving strategies and group processes in small group learning computer programming. American Educational Research Journal 23, 245-261.

Webb, N., M., & Kenderski, C. M. (1984). Student interaction and learning in small group and whole class settings. In P. L. Peterson, L. C. Wilkinson, & M. Hallinan (Eds.), The

Cooperative Learning 71

social context of instruction: Group organization and group processes (pp. 153-170). New York: Academic Press.

Wilkinson, L. (1988-89). Grouping children for learning: Implications for kindergarten education. In E.Z. Rothkopf (Ed.) Review of research in education (pp. 203-223). Washington D.C.: American Educational Research Association.