Optimizing small group discourse in classrooms: Effective practices and theoretical constraints

International Journal of Educational Research 63 (2014) 107–115

Contents lists available at SciVerse ScienceDirect

International Journal of Educational Research journal homepage: www.elsevier.com/locate/ijedures

Optimizing small group discourse in classrooms: Effective practices and theoretical constraints Christine Howe * Faculty of Education, University of Cambridge, 184 Hills Road, Cambridge CB2 8PQ, UK

A R T I C L E I N F O

A B S T R A C T

Article history: Received 9 November 2012 Received in revised form 5 March 2013 Accepted 19 March 2013 Available online 25 April 2013

Acknowledging that small group activities are prominent features of science classrooms, this article addresses two questions about the discourse that occurs while such activities are in progress. The first is whether small group discourse actually matters as regards student learning, in other words whether there are forms of discourse that, if they occur in small groups, promote knowledge gain. With reference to the author’s past research, this question receives a clear, affirmative answer. The second question relates to the prevalence of productive forms of small group discourse in science classrooms, and here the focus is a systematic review of research that others have conducted. Although a sizeable body of material is identified that describes relevant discourse, virtually none of it takes productive forms as the yardstick and addresses their prevalence. This state of affairs is attributed to tacit theories of learning, which locate key processes within whole-class discourse orchestrated by teachers and physical activities (not discourse) that occur at the small group level. Moreover, these theories are likely to be held by practitioners as well as researchers. The implication is that if classroom-based discourse is to be improved in small group settings, it is not, for science, fundamentally a question of establishing relevant strategies. Rather it is acceptance that, far from being tangential to the teaching and learning process, small group discourse is a resource that should be harnessed appropriately. It is suggested that this message might apply beyond the science context. ß 2013 Elsevier Ltd. All rights reserved.

Keywords: Classroom discourse Small group Science

1. Introduction While the whole class is the basic organizational unit for purposes of teaching, classes are often divided into smaller groups for specific activities. The frequency with which small groups are used depends upon numerous factors, of which culture is probably fundamental (Alexander, 2001; Osborn, 2001): at present group work appears to be particularly prevalent in North America, Northern Europe, and Australasia, although there is evidence for growing interest in other parts of the world. Within cultures, one of the key sources of variation is school subject area, for small group activity seems to play an especially significant role in science. For instance, based on a survey of 331 English primary schools and 248 English secondary schools, Baines, Blatchford, and Kutnick (2003) report that 46% of science teaching takes place in small group contexts compared with only 15% of mathematics teaching and 24% of language teaching. This extensive use of group work within science is almost certainly a consequence of the emphasis on practical work: schools do not typically have the resources to provide apparatus on a one-to-one basis, and the need to share necessitates group work. However, practical

* Tel.: +44 1223 767724; fax: +44 1223 767602. E-mail address: [email protected]. 0883-0355/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijer.2013.03.011

108

C. Howe / International Journal of Educational Research 63 (2014) 107–115

work in science typically involves using apparatus to create effects or to examine causes, and it is inconceivable that such activities could be conducted within small group contexts without social interaction (and therefore discourse) amongst students. Occupying nearly half of the total teaching time and involving discourse, small group activity in science is therefore a key forum for classroom-based discourse. As a consequence, it was selected as the focus for the material to follow. Specifically, two questions are addressed relating to science. The first is whether the discourse that occurs during group work actually matters, in other words whether there are forms of discourse that, if they occur, support student learning. This question has been the focus of my own research for over 20 years, with learning interpreted as both conceptual mastery and procedural skill. Therefore the first of the two sections that follow is structured around a review of my previous research. The second question relates to the prevalence of productive forms amongst the small group discourse that occurs routinely in science classrooms. Here the focus, which occupies the second of the sections to follow, is a systematic review of research that others have conducted. The basic conclusion is that whilst a definitive answer can be given to the first question, the second question remains unresolved. Moreover, this lack of progress does not reflect failure to sample the small group discourse that occurs in science classrooms: numerous studies have analyzed such samples. Rather, it is a consequence of analyses that typically by-pass conceptions of productivity. The explanation that is offered for this seemingly curious state of affairs implies a need for significant theoretical change. Furthermore, the need is not simply to permit research gaps to be filled, but also (and crucially) to support optimal practices in classrooms. It is suggested that this is a message that might apply beyond the science context. 2. Productive discourse during group work in science Analyses of pedagogic discourse are traceable to classical Greece, in particular via Plato to Socrates. Moreover one theme that occurs repeatedly in such analyses is the power of discourse that involves contrasting ideas. Appearing initially in Greek texts, it is detectable in the pedagogical writings of, for instance, Rousseau, Mill, Dewey and Piaget. While Bakhtin seldom addressed education explicitly (and application of his ideas to classrooms is perilous, Matusov, 2004), the theme is unmistakable in such claims as ‘The importance of struggling with another’s discourse, its influence in the history of an individual’s coming to ideological consciousness, is enormous’ (Bakhtin, 1981, p. 348). By the 1980s, benefits from contrasting ideas during peer interaction were receiving empirical support, primarily from research with Piagetian conservation and perspective taking tasks (e.g. Doise & Mugny, 1984; Perret-Clermont, 1980) and with tasks requiring the resolution of ethical or legal dilemmas (e.g. Damon & Killen, 1982; Roy & Howe, 1990). This led me to wonder about group work in science, especially given the close association that Piagetians theorize between the knowledge underpinning conservation and perspective taking and conceptions of physical and biological reality. Certainly, evidence was emerging in the 1980s that students approach science education with a wide range of preconceptions about the phenomena they are studying (e.g. Driver, Guesne, & Tiberghien, 1985), suggesting that contrasting ideas within small groups were highly probable. However, these preconceptions were often found to diverge markedly from the target science, in extreme cases proving contradictory. Even if the goal is merely progress towards targets rather than complete mastery, it seemed implausible that discussion alone would turn out to be beneficial. Nevertheless, together with colleagues, I resolved to examine the issue, and the paragraphs to follow summarize what we found.1 In other words, the first of the two questions flagged above is addressed through ascertaining the relevance to learning of small group discourse around contrasting ideas. Reflecting the emphasis on preconceptions within background research, the majority of my studies address the conceptual dimension of science, e.g. students’ understanding of the properties of objects relevant to floating and sinking (Howe, Rodgers, & Tolmie, 1990; Tolmie, Howe, Mackenzie, & Greer, 1993), the direction and speed of object motion (Howe, Tolmie, Anderson, & Mackenzie, 1992; Howe, Tolmie, & Rodgers, 1992; Howe, Tolmie, & Mackenzie, 1995), and the characteristics of containers that determine the rate at which hot water cools (Howe & Tolmie, 2003; Howe, Tolmie, Greer, & Mackenzie, 1995). Jointly, the studies cover the age range from late primary school (8–12 years) to undergraduate level. In all studies, groups (dyads, triads, or foursomes) worked on tasks that required them to formulate joint predictions about outcomes. For instance, groups made predictions about whether an empty metal box or a solid rubber ring would float or sink in a tank of water, and whether a heavy lorry rolling down a slope with a rough surface and then onto the floor would travel a greater, similar or lesser distance along the floor than a light car rolling down a smooth surface. Having agreed predictions, groups were invited to test these using apparatus that was provided, and to formulate joint interpretations of why things turned out as they did. Tasks were designed to optimize the expression and discussion of student ideas, e.g. through requiring each student to write personal predictions on cards before formulating joint predictions (so that nobody could ‘hide’) and repeatedly instructing groups to ‘make sure everybody says what they think’. Sometimes task instructions were presented via computers, but usually they were presented via workbooks, which group members took turns to read out loud. After a brief introduction, groups worked through the tasks on their own with minimal intervention from teachers or researchers, and typically took about 1 h to reach completion. In all studies, participating students were individually pre-tested to establish their preconceptions prior to the group tasks, sometimes by responding orally in one-to-one interviews, and sometimes by completing written tests in whole-class

1 While this is the first time that all of the studies reviewed here have been brought together, the studies described in the first two paragraphs have been summarized previously (Howe, 2010). The text of these two paragraphs is substantially the same as the earlier summary.

C. Howe / International Journal of Educational Research 63 (2014) 107–115

109

settings. Some studies took groups where students were known, from pre-test responses, to hold contrasting views about the concept under investigation, and made comparisons with groups where students were known to have similar views. Mostly the studies considered group discourse directly, through analysis of videotapes recorded while the tasks were ongoing. Without exception, the results provide strong evidence for the power of contrasting ideas. In particular, the students who worked in groups where initial ideas differed and/or were observed to express differences during group work demonstrated significantly greater conceptual understanding when individually post-tested a few weeks later than during the initial pretest. Their pre- to post-test progress also significantly outstripped the progress detected with students who worked in groups with similar ideas and/or failed to express differences. When post-tested a few weeks after the group tasks, the latter students had sometimes made no progress whatsoever, despite completing the tasks and talking while they did this. Furthermore, contrasting opinions triggered change regardless of whether the contrast occurred because some group members had better understanding than others (as with the metal box floats because it is light for its size vs. the metal box floats because it is light), or whether it occurred despite equivalent (and generally poor) understanding (as with the metal box floats because it is flat vs. the metal box floats because it is shiny). Nevertheless, despite its power, discourse around contrasting opinions cannot be sufficient to guarantee growth. Differences must be resolved in a productive fashion, and it soon became apparent that the resolutions achieved during group interaction (i.e. agreed solutions) were usually irrelevant. It was only at undergraduate level that positive correlations were detected between the quality of ideas converged upon (or even expressed) during group work and pre- to post-test growth: at primary and secondary school level, the groups that converged upon poor solutions were as likely to support growth as the groups whose agreed solutions were good, so long as their opinions differed. Furthermore, there was strong evidence that with school-aged participants, resolution typically involved individual (rather than group-based) re-appraisal once group work was complete, i.e. what O’Donnell and Dansereau (1992) term ‘post-group reprocessing’. For instance, Howe, Tolmie, and Rodgers (1992) found that when students were post-tested immediately after the group tasks, the conceptual quality of their ideas about motion down an incline was, on average, no better than the conceptual quality of the ideas expressed at pre-test. Yet at further post-tests about four weeks later pre- to post-test progress was detectable, and strongly associated with contrasting opinions during group work. Likewise, Tolmie et al. (1993) found that the ideas about floating and sinking produced at post-tests conducted 11 weeks after group work were superior to the ideas produced at post-tests conducted after only four weeks, with 11-week progress again predicted by the expression of differences within group discussion. Interestingly, Mugny and Doise (1978) report similar results with research that uses Piagetian perspective taking tasks. More recent research (Howe, McWilliam, & Cross, 2005) indicates that post-group reprocessing involves the productive use of post-group events, with such use ‘primed’ by group discourse. Specifically, events involving floating and sinking were demonstrated to 9–12-year-olds, e.g. evidence that all other things being equal, big things are more likely to float than small things. Demonstrations were provided on three occasions at fortnightly intervals, without accompanying instruction or even discussion. Students who witnessed the demonstrations after participating within group tasks paid more attention to them than control students who did not undertake group work. Furthermore, using pre- to post-test change once more as the index of learning, students who witnessed the demonstrations after group work made significantly greater progress than the control group. In addition, strong relations were detected between post-test performance and unresolved contradiction within small group discourse, e.g. asserting that ‘Big things float’ and ‘Small things float’ without reconciling the difference: even amongst the students who experienced group work plus demonstrations, progress after unresolved contradiction was ten times the magnitude of progress when unresolved contradiction did not occur. These relations have now been confirmed with datasets addressing motion down an incline and rates of cooling, as well as with a further dataset relating to flotation (Howe, 2009). Thus, drawing everything together, one message from my research is that the form of small group discourse is highly relevant to student learning: the more that students discuss contrasting ideas, the more progress they make towards conceptual mastery. On the other hand, it is less important whether the discussion is itself progressive, for the post-group reprocessing that small group discourse triggers can result in growth. The studies summarized above address the conceptual dimension of science. However, while conceptual mastery is a significant component of classroom science, it is not the sole component. Students are, for instance, also expected to learn standard procedures, perhaps most fundamentally the principles and practices of controlled experimentation. Thus, my colleagues and I have also conducted four studies that address small group discourse around the design and implementation of experimental investigations (Howe & Tolmie, 1998, 2003; Howe, Tolmie, Duchak-Tanner, & Rattray, 2000; Tolmie & Howe, 1994). The studies involved students in the late primary to early secondary age range (9–14 years), and in each case the group task was to determine the effects of one specified (‘focal’) factor, using apparatus that allowed this factor and four non-focal factors to be manipulated or held constant. There was rotation of which factors were focal and which were non-focal. For instance, two studies addressed the factors relevant to shadow size, and the apparatus comprised a screen, a lamp whose brightness could vary, triangles of differing sizes and colours, and a slider that could hold the triangles and move them to varying distances between the lamp and screen. Thus, lamp brightness, triangle size, triangle colour, triangle-lamp distance, and triangle-screen distance could all be manipulated or held constant, with appropriate choices depending upon which factor was focal. The central aim of the studies was to examine the effectiveness of differing forms of expert support, and therefore many aspects of the design and results are irrelevant here. What are important are video-recordings made while the students worked on the tasks in small groups (of between three and five members), and relations between group discourse and

110

C. Howe / International Journal of Educational Research 63 (2014) 107–115

change from pre-tests prior to the group tasks to post-tests a few weeks afterwards where the students designed experiments individually. On average, pre- to post-test change was positive, indicating progress. The two earliest studies (Howe & Tolmie, 1998; Tolmie & Howe, 1994, addressing shadow size and water pressure) coded discourse into categories that covered contrasting ideas and much more: request contribution, suggest appropriate factor, suggest inappropriate factor, suggest appropriate apparatus setting, suggest inappropriate setting, suggest prediction, suggest observation, suggest conclusion, relate to external situation, relate to previous tests, relate to current factors, relate to principles, accept contribution, reject/query contribution. However far from finding strong positive associations between category frequencies and pre- to post-test growth, virtually all correlations were non-significant, and the few significant correlations were negative. Similar discourse categories were used in the two more recent studies (Howe & Tolmie, 2003; Howe et al., 2000, concerned respectively with rates of cooling and shadow size), with some differences in the results. Here relations between discourse categories and experimental skill were consistently non-significant rather than sometimes significantly negative. Moreover, these studies highlighted how discourse during earlier tasks could sometimes create frameworks that boosted the quality of subsequent experiments. In particular, when groups discussed and agreed which factors affect outcomes and which are irrelevant, and in subsequent sessions tested consensual views, the sense of ownership, created through discourse, promoted higher quality experiments. Nevertheless, the relation between group discourse and experimental skill was indirect and temporally displaced, and this apart, the studies confirm the basic message from their two predecessors: small group discourse when conducting experiments is irrelevant to the progress that students make in mastering key procedures. What was in fact relevant to progress (in all four studies) was the adequacy of the tests conducted during group work, i.e. manipulating the focal factor and holding the non-focal factors constant. In other words, it was what the students did that mattered, not what they said. In sum then, my studies provide strong evidence that small group discourse around contrasting ideas can promote conceptual understanding in science, but do not indicate comparable benefits for procedural skill, where physical activity (i.e. testing) appears to be more important than discourse. However, even with the conceptual dimension, it is unclear whether the benefits could be realized in classrooms, i.e. whether conceptual understanding in science could be promoted through small group discourse during routine teaching. The reason is that although all of the studies discussed so far were school-based, none were conducted in classrooms. Moreover, in all cases, researchers rather than teachers introduced the procedures. When authentic classrooms are busy and multi-faceted, it is uncertain that small group discourse around contrasting ideas would even be noticed, let alone have the force it is capable of. Some reassurance can be obtained from the Thinking Together programme that Neil Mercer and colleagues have developed, for the programme is fully embedded in routine practice (see, e.g. Mercer & Littleton, 2007). Furthermore, data obtained during programme evaluation point towards the power of contrasting ideas. Within Thinking Together, the significance of small group discourse around contrasting ideas is introduced to teachers via professional development sessions, and teachers are also shown strategies for boosting this form of discourse in classrooms, which they are encouraged to use. Efficacy of implementation is assessed via videorecordings, and related to student learning as revealed on pre- and post-tests. One assessment took place with 9–10-year olds in a science context (Dawes, 2004; Mercer, Dawes, Wegerif, & Sams, 2004): compared with control students who did not participate in Thinking Together, students whose teachers implemented the programme: (1) exchanged opinions more frequently during group work; (2) progressed more from pre- to post-test. Yet while results from Thinking Together are encouraging, they do not show that student progress was actually caused by opinion exchange, even though causation is highly plausible. Recent research that my colleagues and I have conducted relating to science was partially intended to address this, while emulating the authenticity of the Thinking Together programme and synthesizing some of its features with group tasks like those used in our earlier research. Howe et al. (2007) describe the research in detail, but essentially it involved two extended (3+ weeks) programmes of teaching. The programmes related respectively to evaporation and condensation, and force and motion. As is normal in classrooms, the programmes included whole-class teaching and practical demonstration in addition to small group activities. However, group work was prominent, and addressed conceptual understanding and (to a lesser extent) skilled experimentation. Twenty-four primary school teachers implemented the programmes with their 10–12-year-old students, and administered individual pre- and post-tests at the beginning and end of each programme. Researcher involvement during implementation was restricted to recording classroom discourse, and rating the overall quality of group work. While the pre- and post-tests covered ‘fair testing’ (i.e. controlled experimentation), their emphasis, reflecting the programme’s focus, was upon conceptual understanding. Therefore, it was enormously reassuring to find that discussion of contrasting opinions during small group activity was the single most important predictor of knowledge gain. Furthermore, gain was detected not simply between programme pre- and post-tests, but also after an 18-month interval (Tolmie et al., 2007) when the students had moved from primary school to secondary. At the same time, the results also suggest a word of caution: there were negative correlations between the extent of teacher intervention and the extent of learning gain, implying that the more teachers leave groups alone, the more students express contrasting opinions and/or find these beneficial. 3. Monitoring small group discourse in science Thus research that takes place in classrooms and research that is one degree removed both attest to the relevance of small group discourse for learning in science, at least when the focus is upon conceptual mastery. As regards conceptual

C. Howe / International Journal of Educational Research 63 (2014) 107–115

111

understanding, students who discuss contrasting opinions around the phenomena they are studying make greater progress than students who do not hold such discussions. Taking the point as established, the issue to be addressed in this section is the extent to which small group discourse around contrasting opinions is embedded in science education as normally practised. At first sight, it might appear as if the issue has already been affirmatively, albeit implicitly, resolved, via the research discussed above. After all, many features that are associated through the research with opinion exchange occur naturally in classrooms. For instance, as noted earlier, students embark on science with a wide range of preconceptions, and therefore any classroom group, even if randomly formulated, will comprise students with contrasting ideas. Moreover the tasks used in the research emulate many properties of typical classroom exercises. However, while natural occurrence is true of some features, it is not true of all: in my research the tasks were embellished with procedures designed to draw differences out, and in the classroom-based investigations with which the previous section ended, teachers were trained in supportive strategies. For this reason, it would be inadvisable to draw inferences about routine practice from the results: the issue needs to be investigated directly, via studies that monitor the small group discourse that occurs in science. While I have not conducted such studies myself, I have recently completed a systematic review of publications relating to classroom discourse in general (Howe and Abedin, in press). A substantial body of work was found to be concerned with small groups in science, and because Howe and Abedin do not pinpoint this work specifically or analyze it in depth, it provides the focus for the discussion that follows. As detailed in Howe and Abedin’s report, the review procedures began with the formulation of two sets of keywords, one set relating to discourse (e.g. conversation, dialogue, discourse, interaction) and the other set relating to classrooms (e.g. classroom, instruction, small-group, student). The keywords were cued into the Google Scholar search engine, and also used to examine four electronic databases: EBSCO, ERIC, JSTOR and PsycINFO. The specification was for any publication associated with at least one keyword from each set, which had appeared during the 40 years up to 2011. The reference lists of items identified via the electronic resources were then examined in case further publications were to be found, with this process iterated until no new items emerged. The search procedures generated 1532 publications, which were then assessed against seven criteria: (1) the item had to be peer-reviewed, as a straightforward index of threshold quality; (2) the publication had to be obtainable in full-text form; (3) the publication had to be a primary report of empirical research—articles summarizing some of the primary reports were identified, but to avoid duplication these were not included in the review sample; (4) the research had to be conducted in primary or secondary school classrooms (or whatever the equivalents were termed in other countries), since the focus was upon compulsory schooling; (5) the primary focus of the research had to be dialogue, with dialogue defined in its broad everyday sense; (6) the publication had to be in English, although the reported research did not have to be conducted in an English-speaking country; (7) the results should not have been covered already via another, more comprehensive publication in the review sample, once more to avoid duplication—a single study could only be included more than once if different results were covered. Initially, the abstracts alone were assessed, resulting in the exclusion of 857 publications for failure to comply with one or more of the criteria. The full text of the remaining items was then examined, resulting in the exclusion of a further 450 publications. Thus, 225 studies remained as Howe and Abedin’s review sample, and 59 of these studies were conducted in science classrooms. Science was by far the most prevalent of the subject areas that were covered (lack of clarity over subject area with a further 68 studies means that there may actually have been more than 59 science-based investigations; the total of 59 covers the clear-cut cases). Of the 59 studies, 27 were ruled out as being concerned with whole-class interaction and/or teacher–student interaction. Moreover, while the remaining 32 studies all addressed group work amongst students, eight were attempts to modify discourse rather than chart its naturally occurring features. Thus, 24 studies remained for detailed analysis. The studies were conducted in Australia, Brazil, Canada, Germany, Greece, Hong Kong, Spain and the United States, with the United States accounting for 11 studies and Canada accounting for five. While publication dates ranged between 1986 and 2008, 17 studies were published between 1996 and 2003. No study addressed the early primary school age group, but there was a reasonably even spread across late primary school (7 studies), early secondary (8 studies) and late secondary (9 studies). Sixteen studies addressed physics, seven studies addressed biology, and one study addressed chemistry. Inevitably, the 24 studies overlap with the studies that Bennett, Hogarth, Lubben, Campbell, and Robinson (2010) consider in their review of ‘small group discussions in science teaching’. However, the overlap is far from perfect. This is partly because Bennett et al. restrict their sample to research with secondary school students, and partly because they use ‘small group’ and ‘science’ as the search terms when Howe and Abedin’s keywords are much broader. Moreover, Bennett et al. are concerned with the effectiveness of small group discourse, i.e. the issue discussed in the previous section, which as Bennett et al. acknowledge none of their sampled studies address satisfactorily. The issue of present concern is the extent to which small group discourse in science classrooms routinely approximates what the research discussed above has shown to be effective. In fact, while all 24 sampled studies provide detailed information about small group discourse, the prevalence of effective forms is almost entirely overlooked. In particular, only Mortimer and Machado (2000) and Tao (2001) could be said to have examined exchanges of opinion in contexts where conceptual growth was the focus. Moreover, while exchanges were found to be commonplace, both pieces of work were conducted with students in the late stages of secondary schooling. Thus, they do not provide information about the full age range. In addition, both studies were small in scale: Mortimer and Machado’s work addressed a single (non-randomly selected) group, while Tao’s analysis was based on nine groups. Exchanges of opinion were coded in two further studies (Alexopoulou & Driver, 1997; Roychoudhury & Roth, 1996), again with relatively senior students. However, Alexopoulou and

112

C. Howe / International Journal of Educational Research 63 (2014) 107–115

Driver’s analysis focused on agreement, disagreement and requests for evidence, and gender differences over the usage of these three forms. Although disagreement and requests for evidence most probably triggered exchanges of opinion, this was not considered explicitly. Roychoudhury and Roth did analyze exchanges of opinion, and demonstrated that they are most frequent when groups are symmetric, i.e. when all members are involved. However, their data were collected during the design and implementation of experiments, when according to my studies of group-based experimentation, opinion exchange has no positive function. Seven of the remaining studies were also concerned with patterns of dialogue, but their focus was upon variations across group members over individual contributions. The studies did not address the relations amongst contributions, as implied by opinion exchange. The theme of gender differences provides the strongest link amongst the studies, with four addressing the theme directly (Conwell, Griffin, & Algozzine, 1993; Guzzetti & Williams, 1996; Kurth, Anderson, & Palincsar, 2002; Rennie & Parker, 1987). The general finding is that typically girls play supportive rather than dominant roles in science group work, in other words watching, listening and facilitating. It is boys who are most likely to make proposals, and (non-discursively) to control apparatus during practical activities. A fifth study (Hogan, 1999) attempts to characterize the roles that students play during group work. When the roles include contributing ideas, completing tasks, mediating group interaction and remaining reticent, they resonate clearly with the analyses of gender. The final two studies in the group are concerned respectively with the emotional and physical characteristics of classrooms that promote student contribution (Olitsky, 2007; Roth, McGinn, Woszczyna, & Boutonne´, 1999). While it would overplay the degree of relatedness to suggest that these studies too are concerned with gender, both reports stress strong cultural underpinnings to the variations that are highlighted. Culture is obviously also crucial with gender. No matter how much they relate to each other though, it is difficult to regard any of the studies as germane to the central issue. Contributing to small group discourse is not the same as participating: students must speak in order to contribute, but they can participate through being actively engaged in silent listening. Therefore, gender disparities over levels of contribution do not necessarily signify gender disparities over access to productive forms of discourse. With the remaining 13 studies in the sample, small group discourse was regarded not so much as a potential mechanism of learning, but as a window through which learning can be observed. In particular, it was regarded primarily as a context in which student mastery can be externalized and therefore opened to formative assessment. Sometimes observations were made on one occasion only, but typically a series of observations is reported with analyses focusing on change in understanding over time. In eight cases, the emphasis was upon conceptual understanding, particularly the match or (more often) mismatch between the notions students call upon when explaining phenomena and the interpretations that curricula target (Ash, 2008; Finkel, 1996; Gilbert & Pope, 1986; Jime´nez-Aleixandre & Pereiro-Munoz, 2002; Lajoie, Lavigne, Guerrera, & Munsie, 2001; Meyer & Woodruff, 1997; Palincsar, Anderson, & David, 1993; Roth & Roychoudhury, 1992). The remaining five studies highlight the epistemological status of science knowledge (Herrenkohl, Palincsar, DeWater, & Kawasaki, 1999; Jime´nez-Aleixandre, Rodrı´guez, & Duschl, 2000; Keys, 1997; Richmond & Striley, 1996; Tao, 2003). Key questions relate to what discourse reveals about students’ grasp that concepts are theories, theories are provisional, and data are relevant to theory evaluation. One recurring theme across the 13 studies is that knowledge changes gradually, with unorthodox ideas co-existing with received ideas long before the latter become ascendant (e.g. Ash, 2008; Herrenkohl et al., 1999; Palincsar et al., 1993; Roth & Roychoudhury, 1992). A second, undoubtedly related, theme is that progress occurs in specific areas before being generalized (Finkel, 1996; Meyer & Woodruff, 1997). A third is the close tie-up between conceptual misunderstanding and lack of procedural skill, especially over controlled experimentation (Keys, 1997; Richmond & Striley, 1996). This third set of studies has two striking features. First, the underpinning assumptions are precisely the reverse of what would seem appropriate from the research of mine that was outlined earlier. As noted, the studies in the third set focus on the potential of small group discourse to disclose student learning, rather than to operate as the mechanism by which learning occurs. Moreover, the key indices of learning are the ideas that students, collectively or individually, adopt during group discussion. By contrast, my research shows very clearly that small group discourse can operate as the learning mechanism, yet warns against placing too much emphasis upon within-group ideas. Such ideas may be better, worse or equivalent to the ideas that students begin with, but whatever the case they are likely to be transformed post-group. The results of transformation rather than within-group ideas should be regarded as indexical of growth. Second, the small group context is not typically even treated as the locus of learning within the third set of studies. Rather, it is what teachers say (perhaps in interaction with students) that drives student learning, not what small groups say when teachers are absent. Thus, many of the studies (and indeed some work placed in the first and second sets too) emphasize how the teacher’s ‘voice’ is re-presented during small group activity, and how it does or does not exert authority over group discussion (Finkel, 1996; Herrenkohl et al., 1999; Jime´nez-Aleixandre & Pereiro-Munoz, 2002; Kurth et al., 2002; Mortimer & Machado, 2000; Palincsar et al., 1993; Roth & Roychoudhury, 1992). 4. Discussion and conclusions If the perspective outlined in the previous paragraph is widespread, it is scarcely surprising to find so few studies monitoring small group discourse in science classrooms in order to check the incidence of productive features. Instead of treating such discourse as a phenomenon whose forms may be more or less productive, the implication is perceived

C. Howe / International Journal of Educational Research 63 (2014) 107–115

113

irrelevance for learning no matter what forms are adopted. At first, this implicit dismissal of small group discourse is bewildering, when there is so much evidence indicating its significance. However, the dismissal becomes comprehensible once one remembers that the use of small group activities in science is probably not the result of reflections on discourse but rather a reaction to physical shortage: inadequate apparatus for one-to-one usage. Thus from the practitioner point of view, the success of group work as a forum for learning is likely to be measured against the extent to which joint activity using apparatus is associated with positive results, and it is possible that this attitude also permeates a significant proportion of classroom-based research. In other words, both practitioners and educational researchers typically interpret successful science teaching as dependent upon teacher-orchestrated wholeclass discourse and practical activity within small groups. Certainly, research-informed guidance for practitioners is often couched in these terms (e.g. Association for Science Education, 2004; Harlen & Qualter, 2004; Sherman, 1999), as in the United Kingdom are policy directives (Department of Education & Science, 1989; Learning & Teaching Scotland, 2000). Moreover, my studies show how easy it would be for classroom observers (whether teachers or researchers) to underestimate the benefits of small group discourse. For one thing, such discourse is irrelevant when science procedures are the focus: in my work, it was not simply that mastery of experimentation was unrelated to small group discourse, but also that mastery was strongly associated with physical action, i.e. successful experimentation through joint activity. In addition, while small group discourse is strongly and consistently relevant for conceptual understanding, its benefits are not straightforwardly detectable. They depend upon contrasting opinions rather than constructing correct solutions, which may in fact be misleading. Furthermore, they may not be apparent until some interval after the group work is complete. Finally, they can be undermined through direct intervention from teachers. Nevertheless, despite the lack of transparency, the potential benefits from small group discourse are both genuine and extremely robust. Thus, it is to be hoped that future research will fill the gap that has been identified, through charting the incidence of productive forms during routine interaction in classrooms. Furthermore, to the extent that productive forms turn out to be rare, there is one very good reason for changing practices to increase their frequency. As stressed already, there is evidence that students approach science education with a wide range of preconceptions, some highly divergent from curriculum targets. In addition though, there is also evidence that these preconceptions are extremely durable, surviving many years of formal teaching. In other words, the traditional combination of teacher-orchestrated whole-class discourse and practical activity within small groups is not universally successful at promoting conceptual change. While it is inconceivable that small group discourse could substitute for teacher-led instruction, it could perhaps nudge understanding in the desired direction, creating greater receptivity to what teachers say. Thus it could (and my research suggests would) persuade students to use lightness for size as the criterion for floating rather than mere lightness, paving the way for teaching about density. Similarly, it should persuade students to regard warming as a process of gradual heat transfer rather than instant heat attraction, setting the scene for teaching about conduction. If the optimization of small group discourse were to be accepted as a goal worth pursuing, the literature already provides clear guidance as to the steps that should be taken in classrooms to bring this to fruition. The Thinking Together programme and the Howe et al. (2007) project discussed earlier can be interpreted in these terms, and their message is endorsed and amplified through work reported in Blatchford, Baines, Rubie-Davies, Bassett, and Chowne (2006), Galton, Hargreaves, and Pell (2009), Kutnick, Ota, and Berdondini (2008), and Osborne, Erduran, and Simon (2004). While less concerned with the relation between discourse and learning, this work addresses small group discourse in terms that are entirely compatible with the themes developed here, and presents complementary strategies that could be employed in classrooms. However, use of these strategies depends upon acceptance amongst practitioners that small group discourse is significant, and there is likely to be resistance. As noted already, it is not simply educational researchers who, bypassing small group discourse, highlight teacher-orchestrated whole-class discourse and practical activity as the locus of growth; it is policy makers and practitioners. Indeed, the research focus was argued earlier to stem from the presumptions within policy and practice. In short, change is required in tacit theories of how learning occurs and the forms of discourse that support this before the strategies known to improve small group discourse in science classrooms are likely to be implemented. Barriers to implementation have been identified previously: teachers have been found to be concerned about losing control, mistiming input, or monitoring inadequately (Brown & Hirst, 2007; Buzzelli & Johnston, 2001; Emanuelsson & Sahlstro¨m, 2008; Franke, Webb, & Chan, 2009; Jones & Tanner, 2002; Pierce & Gillies, 2008; Wood, 1999). However, the barrier identified here is much more fundamental: the attributed theories of learning devalue implementation before it has even been attempted. Moreover, while the emphasis here has been upon science (and an explanation for tacit theorizing has been proposed that is specific to science), it is possible that the barrier exists in some form in other contexts. Based on an extensive questionnaire and interview survey of primary school teachers and students, Fisher and Larkin (2008) demonstrate general underestimation of the power of discourse to support classroom learning. Clearly, the wider implications of these findings warrant more detailed exploration in the future, but they resonate clearly with what has been proposed here for science. This, to summarize, is that optimizing small group discourse involves maximizing the exchange of contrasting ideas around conceptual material. Moreover, establishing whether current practices require improvement and effecting improvement should this be required probably rest as much on changes in tacit theorizing as on the adoption of supportive practices.

C. Howe / International Journal of Educational Research 63 (2014) 107–115

114

Acknowledgements The Economic and Social Research Council of Great Britain was the primary sponsor of the research reported in the early part of this article; the Faculty of Education supported the systematic review reported later. Their contribution is gratefully acknowledged. References2 Alexander, R. (2001). Culture and pedagogy: International comparisons in primary education. Oxford: Blackwell. *Alexopoulou, E., & Driver, R. (1997). Gender differences in small group discussion in physics. International Journal of Science Education, 19, 393–406. *Ash, D. (2008). Thematic continuities: Talking and thinking about adaptation in a socially complex classroom. Journal of Research in Science Teaching, 45, 1–30. Association for Science Education. (2004). Primary Science Review83. Baines, E., Blatchford, P., & Kutnick, P. (2003). Changes in grouping practices over primary and secondary school. International Journal of Educational Research, 39, 9–34. Bakhtin, M. (1981). The dialogic imagination: Four essays. Austin: University of Texas Press. Bennett, J., Hogarth, S., Lubben, F., Campbell, B., & Robinson, A. (2010). Talking science: The research evidence on the use of small group discussions in science teaching. International Journal of Science Education, 32, 69–95. Blatchford, P., Baines, E., Rubie-Davies, C., Bassett, P., & Chowne, A. (2006). The effect of a new approach to group work on pupil–pupil and teacher–pupil interactions. Journal of Educational Psychology, 98, 750–765. Brown, R., & Hirst, E. (2007). Developing an understanding of the mediating role of talk in the elementary mathematics classroom. Journal of Classroom Interaction, 42, 18–28. Buzzelli, C., & Johnston, B. (2001). Authority, power, and morality in classroom discourse. Teaching and Teacher Education, 17, 873–884. *Conwell, C., Griffin, S., & Algozzine, B. (1993). Gender and racial differences in unstructured learning groups in science. International Journal of Science Education, 15, 107–115. Damon, W., & Killen, M. (1982). Peer interaction and the process of change in children’s moral reasoning. Merrill-Palmer Quarterly, 28, 347–367. Dawes, L. (2004). Talking and learning in classroom science. International Journal of Science Education, 26, 677–695. Department of Education and Science. (1989). Science in the national curriculum. London: HMSO. Doise, W., & Mugny, G. (1984). The social development of the intellect. Oxford: Pergamon. Driver, R., Guesne, E., & Tiberghien, A. (1985). Children’s ideas in science. Milton Keynes: Open University Press. Emanuelsson, J., & Sahlstro¨m, F. (2008). The price of participation: Teacher control versus student participation in classroom interaction. Scandinavian Journal of Educational Research, 52, 205–223. *Finkel, E. A. (1996). Making sense of genetics: Students’ knowledge use during problem solving in a high school genetics class. Journal of Research in Science Teaching, 33, 345–368. Fisher, R., & Larkin, S. (2008). Pedagogy or ideological struggle? An examination of pupils’ and teachers’ expectations for talk in the classroom. Language and Education, 22, 1–16. Franke, M. L., Webb, N. M., & Chan, A. G. (2009). Teacher questioning to elicit students’ mathematical thinking in elementary school classrooms. Journal of Teacher Education, 60, 380–392. Galton, M., Hargreaves, L., & Pell, T. (2009). Group work and whole-class teaching with 11- to 14-year-olds compared. Cambridge Journal of Education, 39, 119–140. *Gilbert, J. K., & Pope, M. L. (1986). Small group discussions about conceptions in science: A case study. Research in Science and Technological Development, 4, 61–76. *Guzzetti, B. J., & Williams, W. O. (1996). Gender, text, and discussion: Examining intellectual safety in the science classroom. Journal of Research in Science Teaching, 33, 5–20. Harlen, W., & Qualter, A. (2004). The teaching of science in primary schools (4th ed.). London: David Fulton. *Herrenkohl, L. R., Palincsar, A. S., DeWater, L. S., & Kawasaki, K. (1999). Developing scientific communities in classrooms: A sociocognitive approach. The Journal of the Learning Sciences, 8, 451–493. *Hogan, K. (1999). Sociocognitive roles in science group discourse. International Journal of Science Education, 21, 855–882. Howe, C. (2009). Collaborative group work in middle childhood: Joint construction, unresolved contradiction and the growth of knowledge. Human Development, 39, 71–94. Howe, C. (2010). Peer groups and children’s development. Oxford: Blackwell. Howe, C., & Abedin, M. Classroom dialogue: A systematic review across four decades of research. Cambridge Journal of Education, in press. Howe, C., McWilliam, D., & Cross, G. (2005). Chance favours only the prepared mind: Incubation and the delayed effects of peer collaboration. British Journal of Psychology, 96, 67–93. Howe, C., Rodgers, C., & Tolmie, A. (1990). Physics in the primary school: Peer interaction and the understanding of floating and sinking. European Journal of Psychology of Education, V, 459–475. Howe, C., & Tolmie, A. (1998). Computer support for learning in collaborative contexts: Prompted hypothesis testing in physics. Computers and Education, 3/4, 223–235. Howe, C., & Tolmie, A. (2003). Group work in primary school science: Discussion, consensus and guidance from experts. International Journal of Educational Research, 39, 51–72. Howe, C., Tolmie, A., Anderson, A., & Mackenzie, M. (1992). Conceptual knowledge in physics: The role of group interaction in computer-supported teaching. Learning and Instruction, 2, 161–183. Howe, C., Tolmie, A., Duchak-Tanner, V., & Rattray, C. (2000). Hypothesis testing in science: Group consensus and the acquisition of conceptual and procedural knowledge. Learning and Instruction, 10, 361–391. Howe, C., Tolmie, A., Greer, K., & Mackenzie, M. (1995). Peer collaboration and conceptual growth in physics: Task influences on children’s understanding of heating and cooling. Cognition and Instruction, 13, 483–503. Howe, C., Tolmie, A., & Mackenzie, M. (1995). Collaborative learning in physics: Some implications for computer design. In C. O’Malley (Ed.), Computer-supported collaborative learning (pp. 51–68). Berlin: Springer-Verlag. Howe, C., Tolmie, A., & Rodgers, C. (1992). The acquisition of conceptual knowledge in science by primary school children: Group interaction and the understanding of motion down an incline. British Journal of Developmental Psychology, 10, 113–130. Howe, C., Tolmie, A., Thurston, A., Topping, K., Christie, D., Livingston, K., et al. (2007). Group work in elementary science: Towards organisational principles for supporting pupil learning. Learning and Instruction, 17, 549–563. *Jime´nez-Aleixandre, M. P., & Pereiro-Munoz, C. (2002). Knowledge producers or knowledge consumers? Argumentation and decision making about environmental management. International Journal of Science Education, 11, 1171–1190. *Jime´nez-Aleixandre, M. P., Rodrı´guez, A. B., & Duschl, R. A. (2000). ‘Doing the lesson’ or ‘doing science’: Argument in high school genetics. Science Education, 84, 757–792. Jones, S., & Tanner, H. (2002). Teachers’ interpretations of effective whole-class interactive teaching in secondary mathematics classrooms. Educational Studies, 28, 265–274.

2

Review sample asterisked.

C. Howe / International Journal of Educational Research 63 (2014) 107–115

115

*Keys, C. W. (1997). An investigation of the relationship between scientific reasoning, conceptual knowledge and model formation in a naturalistic setting. International Journal of Science Education, 19, 957–970. *Kurth, L. A., Anderson, C. W., & Palincsar, A. S. (2002). The case of Carla: Dilemmas of helping all students to understand science. Science Education, 86, 287–313. Kutnick, P., Ota, C., & Berdondini, L. (2008). Improving the effects of group working in classrooms with young school-aged children: Facilitating attainment, interaction and classroom activity. Learning and Instruction, 18, 83–95. *Lajoie, S. P., Lavigne, N. C., Guerrera, G., & Munsie, S. D. (2001). Constructing knowledge in the context of BioWorld. Instructional Science, 29, 155–186. Learning and Teaching Scotland. (2000). Environmental studies: Science. Edinburgh: Scottish Executive. Matusov, E. (2004). Bakhtin’s dialogic pedagogy. Journal of Russian and East European Psychology, 42, 3–11. Mercer, N., Dawes, L., Wegerif, R., & Sams, C. (2004). Reasoning as a scientist: Ways of helping children to use language to learn science. British Educational Research Journal, 30, 367–385. Mercer, N., & Littleton, K. (2007). Dialogue and the development of children’s thinking: A sociocultural approach. London: Routledge. *Meyer, K., & Woodruff, E. (1997). Consensually driven exploration in science teaching. Science Education, 81, 173–192. *Mortimer, E. F., & Machado, A. H. (2000). Anomalies and conflicts in classroom discourse. Science Education, 84, 429–444. Mugny, G., & Doise, W. (1978). Socio-cognitive conflict and structure of individual and collective performances. European Journal of Social Psychology, 8, 181–192. O’Donnell, A., & Dansereau, D. F. (1992). Scripted cooperation in student dyads: A method for analyzing and enhancing academic learning and performance. In R. Hertz-Lazarowitz & N. Miller (Eds.), Interaction in cooperative groups: The theoretical anatomy of group learning (pp. 120–141). Cambridge: Cambridge University Press. *Olitsky, S. (2007). Promoting student engagement in science: Interaction rituals and the pursuit of a community of practice. Journal of Research in Science Teaching, 44, 33–56. Osborn, M. (2001). Constants and contexts in pupil experience of learning and schooling: Comparing learners in England, France and Demark. Comparative Education, 37, 267–278. Osborne, J., Erduran, S., & Simon, S. (2004). Enhancing the quality of argumentation in school science. Journal of Research in Science Teaching, 41, 994–1020. *Palincsar, A. S., Anderson, C., & David, Y. M. (1993). Pursuing scientific literacy in the middle grades through collaborative problem solving. The Elementary School Journal, 93, 643–658. Perret-Clermont, A.-N. (1980). Social interaction and cognitive development in children. London: Academic Press. Pierce, K. M., & Gillies, C. (2008). From exploratory talk to critical conversations. In N. Mercer & S. Hodgkinson (Eds.), Exploring talk in schools (pp. 37–53). London: Sage. *Rennie, L., & Parker, L. (1987). Detecting and accounting for gender differences in mixed-sex and single-sex groupings in science lessons. Educational Review, 39, 65–73. *Richmond, G., & Striley, J. (1996). Making meaning in classrooms: Social processes in small-group discussion and scientific knowledge building. Journal of Research in Science Teaching, 33, 839–858. *Roth, M.-W., McGinn, M. K., Woszczyna, C., & Boutonne´, S. (1999). Differential participation during science conversations: The interaction of focal artifacts, social configurations, and physical arrangements. The Journal of the Learning Sciences, 8, 293–347. *Roth, W.-M., & Roychoudhury, A. (1992). The social construction of scientific concepts or the concept map as conscription device and tool for social thinking in high school science. Science Education, 76, 531–557. Roy, A. W. N., & Howe, C. J. (1990). Effects of cognitive conflict, socio-cognitive conflict and imitation on children’s socio-legal thinking. European Journal of Social Psychology, 20, 241–252. *Roychoudhury, A., & Roth, W.-M. (1996). Interactions in an open-inquiry physics laboratory. International Journal of Science Education, 18, 423–445. Sherman, S. J. (1999). Cooperative learning and science. In S. Sharan (Ed.), Handbook of cooperative learning methods. (pp. 226–244). Westport, CT: Praeger. *Tao, P.-K. (2001). Developing understanding through confronting various views: The case of solving qualitative physics problems. International Journal of Science Education, 23, 1201–1218. *Tao, P.-K. (2003). Eliciting and developing junior secondary science students’ understanding of the nature of science through a peer collaboration instruction in science stories. International Journal of Science Education, 25, 147–171. Tolmie, A., Christie, D., Howe, C., Thurston, A., Topping, K., Donaldson, C., et al. (2007). Classroom relations and collaborative groupwork in varying social contexts: Lessons from Scotland. Paper presented at annual meeting of American Educational Research Association, Chicago. Tolmie, A. K., & Howe, C. J. (1994). Computer-directed group activity and the development of children’s hypothesis testing skills. In H. C. Foot, C. J. Howe, A. Anderson, A. K. Tolmie, & D. Warden (Eds.), Group and interactive learning (pp. 139–144). Southampton: Computational Mechanics Publications. Tolmie, A., Howe, C., Mackenzie, M., & Greer, K. (1993). Task design as an influence on dialogue and learning: Primary school group work with object flotation. Social Development, 2, 183–201. Wood, T. (1999). Creating a context for argument in mathematics class. Journal for Research in Mathematics Education, 30, 171–191.