The Influence of the Tutee in Learning by Peer Tutoring

The Influence of the Tutee in Learning by Peer Tutoring Rod D. Roscoe (roscoe@pitt.edu) Michelene T. H. Chi (chi@pitt.edu) Learning Research and Development Center, 3939 O’Hara Street Department of Psychology, University of Pittsburgh Pittsburgh, PA 15260 USA Abstract Previous research has demonstrated that students can learn by tutoring other students. Tutors are thought to learn because they generate instructional explanations and monitor their own understanding while teaching. We analyzed verbal data from tutorial sessions to explore how the tutees influence this process. We found that tutors were primarily responsible for introducing topics, but the tutees stimulated more thorough discussions of topics. We also found that tutee questions influenced tutor explanations and metacognition. Tutor responses to “deep” questions were more likely to contain inferences and self-monitoring than responses to “shallow” questions. In sum, tutees had a significant and positive influence on the tutors’ learning activities and opportunities. Introduction Peer tutoring and cross-age tutoring are popular and cost- efficient educational interventions in which students provide instruction for other students. One reason for the widespread use of these interventions is their effectiveness – with training, students seem quite capable of successfully teaching each other and younger pupils (e.g. Cohen, Kulik, & Kulik, 1982; Greenwood, Carta, & Hall, 1988). Another reason for the popularity of peer and cross-age tutoring programs is the robust finding that the tutors also benefit academically from the teaching experience (e.g. Allen & Feldman, 1973; Annis, 1983; Cloward, 1967; Cohen et al., 1982; Greenwood et al., 1998; Morgan & Toy, 1970; Rekrut, 1992). Based on such findings, some researchers have advocated reciprocal tutoring programs in which the participating students take turns being the tutor and tutee. In general, these programs are educationally effective (e.g. Fantuzzo, King, & Heller, 1992; Fantuzzo et al., 1989; Fuchs et al., 1997; King, Staffieri, & Adelgais, 1998; Palincsar & Brown, 1984). Why do students learn by tutoring? Some evidence suggests that tutors learn by generating instructional explanations, which facilitates integration and organization of knowledge. For example, Coleman, Brown, & Rivkin (1997) found that when students were told to teach a peer by explaining, they learned better than students told to teach by summarizing and better than students who did not teach. Similarly, Fuchs et al. (1997) showed that training students to give each other conceptually-rich explanations during reciprocal tutoring was more effective than classroom instruction and reciprocal tutoring without such explanations. Additional evidence indicates that tutoring may also encourage students to engage in metacognitive self-monitoring, which helps learners to detect and repair missing knowledge and misconceptions. For example, King et al. (1998) trained reciprocal tutors to give quality explanations and to ask each other questions that stimulated critical thinking and self-monitoring. They found that these explaining and metacognitive activities resulted in better learning than explaining activities alone. Explaining and self-monitoring have also been shown to improve learning in solo studying (e.g. Chi, 2000; Chi, deLeeuw, Chiu, & LaVancher, 1994) and collaborative learning (e.g. Coleman, 1998; Webb, Troper, & Fall, 1995), which further highlights the efficacy of these activities. In this paper, we explore the hypothesis that tutees influence the learning activities of the tutors in important ways. In other words, tutors might be able to learn by explaining and self-monitoring, but tutees may affect how and whether these activities occur. One way that tutees may guide the tutorial session is by choosing which topics are discussed and in how much detail, thus creating or limiting opportunities to think about the underlying ideas. Another powerful way in which tutees may influence the learning activities of the tutor is through the kinds of questions they ask. As described above, King (e.g. King, 1994; King et al., 1998) has shown that when students construct and ask each other questions based on high-level question stems (i.e. questions prompting for comparisons, justifications, causes- and-effects, evaluations, etc.), they produce better explanations and learn more effectively. Coleman (1998) has demonstrated very similar findings in collaborative learning settings with students using high-level explanation prompts. Research on naturalistic tutoring has shown that tutees do occasionally ask “deep” questions in tutoring sessions, although the majority of questions are “shallow” (Graesser & Person, 1994). These deep questions, although they may be rare, should stimulate deeper responses. In order to address these hypotheses about the influence of the tutee on tutor learning, we analyzed tutor learning in a non-reciprocal and naturalistic (i.e. little or no training) tutoring context. This design allowed us to be more sensitive to the benefits and processes of tutoring. In reciprocal tutoring, by definition, students learn from both teaching and being taught, and thus it is almost impossible to assess the specific contribution of tutoring activities to learning in these settings. Similarly, it is possible that when tutoring programs are highly structured (i.e. training on when and how to explain, ask and answer questions, etc.), important aspects of spontaneous tutoring behaviors that positively or negatively impact learning may be obscured.