|
Article Excerpt
It is well known that the knowledge and understanding of school-aged children can be facilitated by collaborative work with peers (Damon & Phelps, 1989; Doise & Mugny, 1984; Howe, Tolmie, & Mackenzie, 1995). At the same time, it has become clear that the collaborating peers should have differing ideas about the topics they are considering (Doise & Mugny, 1984; Howe, Tolmie, & Mackenzie, 1995). Similarity of outlook is less productive. The topics should, in addition, require genuine discovery or conceptual grasp (Damon & Phelps, 1989; Howe, Tolmie, Duchak-Tanner, & Rattray, 2000; Rogoff, 1990). Collaboration has been less successful with the practising of skills that have been acquired but not yet perfected. Based on such findings, hypotheses have been proposed about the forms of interaction that precipitate growth. For instance, the combination of differing preconceptions and discovery/conceptual learning suggests a need for the discussion of opposing ideas. Research has confirmed this, either documenting the value of opposition directly (e.g. Howe & Tolmie, 1998; Howe, Tolmie, Greer, & Mackenzie, 1995), or endorsing forms like explanation (Webb, 1989) and transaction (Berkowitz & Gibbs, 1983; Kruger, 1993) that opposition is known to promote (Howe & McWilliam, 2001). What remains uncertain, however, are the mechanisms by which collaborative interaction operates upon and transforms the knowledge of individuals. It is this issue that the three studies to be reported later were intended to address.
Peer collaboration and individual growth
To date, proposals about the mechanisms that underpin collaboratively induced growth have presupposed Vygotsky's (1978) view that knowledge develops 'first, on the social level, and later, on the individual level' (p. 57). As a result, the emphasis has been upon collective insights created by collaborating peers that surpass what any could manage independently, and that are subsequently assimilated by each individual (e.g. Rogoff, 1990; Tomasello, Kruger, & Ratner, 1993). Consistent evidence has been obtained (e.g. Doise & Mugny, 1984; Howe, Tolmie, Anderson, & Mackenzie, 1992; Williams & Tolmie, 2000). However, there are also exceptions, with growth being observed despite collaborative performances that were non-progressive and/or unrelated to eventual change. An example appears in work by Howe, Tolmie, and Rodgers (1992), which involved pre-testing children about motion down an incline, engaging them in a collaborative task, and post-testing them 4 weeks afterwards. Pre- to post-test advances were observed and these were dependent upon opposition during collaboration. However, the advances showed no association with the collaborative performance, which was significantly worse than performance at pre-test. Equivalent results have been obtained by Howe, Rodgers, and Tolmie (1990), Howe, Tolmie, Greer, et al. (1995), and Howe et al. (2000) in other areas of science, and by Wegerif, Mercer, and Dawes (1999) for Raven's progressive matrices.
At present, nothing is known about the interpretation of collaboratively induced growth that is independent of collective insight. However, it may be significant that when independence occurs, the benefits from peer collaboration are typically delayed. The Howe, Tolmie, and Rodgers (1992) study is relevant here, through its use of immediate post-tests administered within 24 hours of the collaborative task in addition to the 4-week post-tests mentioned above. Immediate post-test performance was no better than pre-test, suggesting that the advances detected at 4 weeks had taken place subsequently. Similar results have been reported by Howe et al. (1990), Scanlon, Issroff, and Murphy (1998), Tolmie, Howe, Mackenzie, and Greer (1993) and Wegerif et al. (1999). For instance, in research concerned with floating and sinking, Tolmie et al. found progress between post-tests held 4 and 11 weeks after peer collaboration, with the extent of progress once more associated with the discussion of opposing ideas. The indications are then that to understand how collaboration can have benefits that are independent of collective insights, it may be profitable to focus on the reasons why delayed effects occur.
In principle, the delayed effects of peer collaboration could be encompassed within Vygotsky's (1978) image of social and then individual development. Children might, after all, consult post-collaboratively with parents, teachers or peers, and during consultation achieve collective insights. In practice, this seems unlikely. Systematic investigation by Howe, Tolmie, and Rodgers (1992) yielded little evidence for consultation. Moreover, when consultation occurred, it was not associated with pre- to post-test change. Howe et al. concluded that on balance of probabilities the delayed effects depend on individual cognitive activity, implying a significant departure from the Vygotskyan perspective. The studies that follow attempt to clarify the form that this individual activity takes.
Delayed effects and incubation
The starting point was the fact that if the delayed effects of peer collaboration result from individual cognition, they are instances of what cognitive psychologists call incubation. Cognitive psychology recognizes incubation (Yaniv & Meyer, 1987) when individuals: (a) engage in intense and ultimately unsuccessful work, (b) abandon the task for an interval, at least in terms of discernible activity, and (c) resume work and achieve greater success. Incubation has been recorded autobiographically as contributing to the work of, for example, Darwin, Feynman, Piaget, Poincare and Proust, although the status of these accounts has been challenged (Weisberg, 1986). More recently, incubation has been observed in laboratory problem solving (see Dorfman, Shames, & Kihlstrom, 1996), primarily using anagrams, rebus problems (e.g. What does 'you--just--me' represent; 'just between you and me'), and remote associations (e.g. What links 'democrat, favour, girl'; 'party').
The laboratory and autobiographical contexts differ markedly from the ones within which peer collaboration has been explored. The laboratory problems do not include the discovery/conceptual tasks that benefit from collaboration. The intervals over which incubation is charted are seldom more than a few minutes, not the weeks associated with collaboration. Although discoveries and/or conceptual insights are central to the autobiographical data and the time frames can be substantial, there are obvious differences between breakthroughs in knowledge and their (usually partial) reconstruction by children. Moreover, the fact that the peer collaboration literature has focused on children rather than adults is a further contrast. Nevertheless, although with the exception of Kirkwood (1979), incubation has not even been identified in collaborative contexts, there is no doubt that the delayed effects discussed above have all the relevant features. As a result, it seems reasonable to ask whether the processes proposed in cognitive psychology to account for incubation can be extended to collaboratively induced growth.
Within cognitive psychology, there are three broad approaches to explaining incubation, the first being that delays distract individuals, and by virtue of this boost their ability to cope with tasks once these are resumed. One distraction account calls upon fatigue dissipation during the interval, permitting resumption with renewed vigour. Although dissipation may have relevance in certain contexts (but see Dorfman et al., 1996; Smith & Blankenship, 1989, 1991), it could scarcely underpin the delayed effects of peer collaboration. As noted, gains have been observed between 24 hours and 4 weeks after collaboration (Howe, Tolmie, & Rodgers, 1992), and between 4 weeks and 11 weeks (Tolmie et al., 1993). Such intervals are well beyond those needed for recovery from fatigue. From the collaborative perspective, a more compelling distraction account is the suggestion (Anderson, 2000; Smith & Blankenship, 1989, 1991; Woodworth & Schlosberg, 1954) that individuals can be thwarted initially by unhelpful mental sets, with the interval allowing the retrieval of productive strategies. The latter depends upon alternative strategies being available, and should therefore be supported by previous experience with a variety of ideas. As a result, the approach is consistent with the evidence cited earlier that peer collaboration is optimized when participants have differing views about the topic they are considering (Doise & Mugny, 1984; Howe, Tolmie, & Mackenzie, 1995).
Although set breaking may contribute to the delayed effects, the implication that initial sets will be unhelpful suggests that it is unlikely to provide a full account. Howe, Tolmie, and Rodgers (1992) and Tolmie et al. (1993) detected relevant as well as irrelevant ideas during their first post-tests, with change to the second post-tests being a shift in balance towards relevancies. To explain the shift, it may be necessary to call upon one or both of the remaining approaches to incubation. One of these approaches emphasizes the processing during the interval of information that was available from the outset. On some versions (e.g. Smith & Blankenship, 1991), the processing revolves around the forgetting of inappropriate material. However, the children in Howe et al. and Tolmie et al.'s research did not know a priori which initial ideas were acceptable, suggesting that more active mechanisms are needed to explain why relevant ideas proved most durable. It would be more consistent with Howe et al. and Tolmie et al. to follow the 'private work' version of the information processing account (e.g. Campbell, 1960; Dorfman et al., 1996; Olton & Johnson, 1976; Wallas, 1926), and suggest that collaboration stimulates reflective (and essentially private) appraisal of collaboratively generated ideas. This results in the rejection of inadequate ideas and confirmation (and/or supplementation) of acceptable alternatives. On the other hand, when individuals work on tasks that remain incomplete, they do not simply restrict their attention to available information; they are also primed to respond differently and more productively to events that occur subsequently (Hayes-Roth & Hayes-Roth, 1979; Patalano & Seifert, 1997). Such priming underpins the third approach to incubation (Dorfman et al., 1996; Olton, 1979; Seifert, Meyer, Davidson, Patalano, & Yaniv, 1995; Yaniv & Meyer, 1987), for this approach emphasizes the likely interval between priming from initial failure, and experiences that, by hypothesis, are primed for. In principle, a parallel situation could apply with tasks that are attempted collaboratively, with priming arising from the collaborative interaction, and the delayed effects resulting from time-lagged access to pertinent events.
In sum, the incubation literature suggests that the delayed effects of peer collaboration could result from: (a) elimination over time of unhelpful mental sets, (b) engagement in post-collaborative reflective appraisal, and (c) experience of relevant events that, thanks to collaboration, can be used productively. The processes are not mutually exclusive: the second and third could occur after set breaking, and the third could supplement ideas that are reflectively appraised. The issue is whether any apply with peer collaboration in practice. Resolving the issue would not only make a significant contribution to clarifying the mechanisms by which collaboration transforms individual knowledge; it should also, given that incubation has not previously been studied in collaborative contexts, inform the generalizability of cognitive models. Recognizing this two-fold importance, the studies that follow took each of the processes in turn, and explored their relevance to the collaboratively induced understanding, by 9- to 12-year-old children, of the factors relevant to floating and sinking. The focus upon floating and sinking was motivated partly by its reliable association with delayed learning gain (Howe et al., 1990--two studies; Tolmie et al., 1993), and partly by the wide range of preconceptions that children are known to hold, that is, 250 distinct ideas reported by Howe (1998). Such diversity renders the topic ideal for exposing patterns of change.
STUDY 1
Study 1 addressed the possibility that the delayed effects of peer collaboration stem from mental sets that require significant intervals to be broken. The study involved individual pre-tests, peer collaboration, immediate individual post-tests, and delayed individual post-tests. Collaboration for half of the children (WeakSet condition) involved a task that was expected (Howe, Tolmie, Greer, et al., 1995; Tolmie et al., 1993), to have modest set-inducing powers. Collaboration for the other half (StrongSet condition) involved a task that was expected to have strong set-inducing powers. If intervals are needed to break mental sets, these should, therefore, be more significant for the StrongSet children than the WeakSet. As a result, there should be: (a) less progress from pre-test to immediate post-test with the StrongSet children than with the WeakSet, and (b) more progress from immediate post-test to delayed post-test with the StrongSet children than with the WeakSet.
Method
Participants
The pre-test sample comprised 73 children who attended a state primary school in Glasgow. Twenty-four children (15 boys; 9 girls) were from the school's Primary (P)5 class (age range = 9;0 to 10;11, M = 9;8). Twenty-five children (12 boys; 13 girls) were from the P6 class (age range = 10;0 to 11;0, M = 10; 7). Twenty-four children (13 boys; 11 girls) were from the P7 class (age range = 10;11 to 12;4, M = 11;5). (1) Absence from school meant that 12 children (8 boys; 4 girls; 4 from each age group) missed the collaborative task. Their data were discounted.
Materials
Materials for the pre-test, immediate post-test and delayed post-test comprised 12-item response booklets, sets of objects placed in envelopes, and a fish tank half-filled with water. The first six items in the booklets instructed children to consider six objects in turn, tick float, sink or not sure to indicate what each object would do if immersed in water, and write 'all the reasons'...
|
