Exploring the characteristics of small groups within Science and English secondary classrooms
Sarah MacQuarriea*, Christine Howeb, and JamesBoylea
aDepartmentof Psychology, University of Strathclyde
b Faculty of Education, University of Cambridge
Contact:
*now based at Centre for Rural Childhood, University of the Highlands and Islands.
Exploring the characteristics of small groups within Science and English secondary classrooms
Abstract
Studies of primary education within the UK have shown that small groups can feature within classrooms; however equivalent research within secondary education remains scarce. Research has established effective group work approaches, yet secondary teachers may encounter difficulties employing approaches tied to parameters embedded within primary education. This problem is compounded as minor adjustments to the conditions surrounding group work are known to have consequences for its efficacy within classrooms. This paper reports naturalistic systematic observation of group work practice within 23 science and English secondary classrooms in Scotland. Pupils completed tasks according to whether they were situated within group work or conventional classes. Forms of dialogue known to be conducive to learning were prevalent whilst pupils worked in groups. The change in pupils’ behaviours does not appear to stem from the content of teachers’ talk. Teachers’ behaviour suggested they approach small groups as smaller structures equivalent to a whole-class set-up.
Key words: Group work, secondary classrooms, teachers’ behaviour, interaction, Science and English teaching.
Introduction
Research continues to demonstrate that group work involving small numbers of pupils (hereafter ‘group work’) can be employed as a productive teaching strategy (Baines, Blatchford, & Chowne, 2007; Blatchford, Baines, Rubie-Davies, Bassett, & Chowne, 2006; Gillies, 2004; Howe et al., 2007) signifying that awareness of its characteristic features is of great importance. The need is especially marked at secondary level, for overwhelmingly research evaluating classroom groups has focused on primary education. Group work research needs to acknowledge that pupils within secondary education face a myriad of specific challenges, implying that secondary education needs to be studied in its own right. Accordingly, this paper reports the findings of a naturalistic classroom observation study conducted within Scottish secondary classrooms.
Under guidance from education authorities, School Boards and national initiatives, all schools make strategic decisions about how to group pupils. The most frequently used teaching and learning arrangements are whole-class teaching (where a teacher engages with an entire class), group work (where the consensus amongst sources suggests as few as two and as many as six pupils work as a subgroup on a shared task) and individual learning (Alexander, 2000; Galton, Hargreaves, Wall, & Comber, 1999).This paper investigates how teachers and pupils in secondary school operate in groups in comparison with conventional teaching (equivalent to whole-class and individual learning approaches). Group work has been documented as a regular feature of classrooms in the UK (Blatchford, Kutnick, & Baines, 1999; Blatchford, Kutnick, Clark, MacIntyre, & Baines, 2001; Galton et al., 1999; Howe & Tolmie, 2003; MacQuarrie, 2006). Yet a recent review indicates a scarcity of systematic research considering how small groups operate and what factors influence such interaction (Bennett, Lubben, Hogarth, & Campbell, 2004). Paucity of research is also evident regarding teachers’ behaviour within lessons that make use of group work. Therefore a representative profile of classroom group work would help determine whether the strategies employed by secondary teachers incorporate features that research has associated with effective group work (Cohen, 1994; Slavin, 1996).
Background research into productive group work helped frame the parameters of the current study (such as Cohen, 1994; Howe et al., 2007). A wide body of literature has examined the gender and ability composition of classroom groups and heterogeneous arrangements are argued to be particularly effective, as they maximise the chance of contrasting perspectives during group discussion. Research from Piaget (1962) onwards indicates that productive outcomes depend on this form of interaction (see Howe, 2010). While not denying the importance of difference, research alsoindicates that it is preferable for pupils to cooperate rather than compete (Roseth, Johnson, & Johnson, 2008; Slavin, 1996). Studies acknowledging the role of the teacher within group work have concentrated on teacher to pupil interaction, and include investigation of the relationship between external assistance and resulting group behaviour (Webb et al., 2008; Webb, Nemer, & Ing, 2006), whereas the nature of the support given by teachers at a broader classroom level has yet to be examined. It follows that evidence concerning teachers’ roles within groups, and their management of lessons that use groups, is limited. Consideration of classroom dialogue – accessed by observing pupil and teacher talk – can help address this discrepancy. In primary school typically pupils are seen as receivers of knowledge and teachers are viewed as experts or bearers of such knowledge (Edwards & Mercer, 1987, p. 25; Edwards & Westgate, 1994, p. 106)and as studies considering primary to secondary school transfer have reported more rather than less continuity between the two settings, it is a reasonable assumption that these roles are likely to predominate in the secondary school (Galton, Morrison, & Pell, 2000; West, Sweeting, & Young, 2010). Hence to accommodate group work teachers would need to alter their approach to support and contextualise pupils’ learning, thus signalling a change in the strategies and behaviour that characterise their teaching.
Two projects that implemented interventions within schools are particularly relevant. The SPRinG project (Social Pedagogic Research into Group work) was conducted in England with pupils aged 5-14, followed by a partial replication with pupils aged 9-12 in Scotland (SCOTSPRinG). The relational approach to group work - which encompasses many of the principles of cooperative learning advocated by Johnson and Johnson’s (2005, 2009) - emphasises the inclusion of both teachers and pupils within interactions, but stipulates that pupils must be given sufficient time to develop the skills necessary for group interaction. Following extended participation, increases were evident in pupils’ academic achievements (Baines et al., 2007; Howe et al., 2007; Thurston et al., 2008) and encouraging findings were obtained regarding changes in pupil behaviours during interaction with their peers (Blatchford et al., 2006).Although these studies demonstrate that conditions for productive groupwork can be tailored for different disciplines, accounts of naturally occurring practice do not necessarily show comparable findings(Baines, Blatchford, & Kutnick, 2003). Consequently, reports from teachers that group work was especially likely within science and English subjects (in contrast to mathematics[1], history and modern languages), provided parameters for the present study(MacQuarrie, 2006).
Overview of the present study
The main objective of this research was to provide a description of secondary school teachers’ and pupils’ behaviour during lessons that make use of group work, enhanced through comparison with conventional instruction.
With so few studies evaluating group workwithin secondary education, make it difficult to form specific hypotheses about what takes place. However, based on the background literature, two issues were felt to merit specific investigation. The first relates to the structural features of groups looking at the relationship between pupils’ behaviour and the instructional methodsused by the class teacher. Observation (Blatchford et al., 2001; Roth et al., 2006) and self report measures(Assessment of Achievement Programme, 2003, 2005; Scottish Survey of Achievement, 2006)involving secondary schools note that small groups are used as a learning activity but feature in tandem with other instructional approaches. Further research suggests that pupils mainly work individually, irrespective of being seated in a variety of formats within classrooms(Baines et al., 2003). Indeed, being seated as groups and assigned a group task does not necessarily lead to productive interaction, (Barron, 2003; Oliveira & Sadler, 2008; Wegerif, 1997; Yetter et al., 2006). Thus, seating and task arrangements have importance when considering the nature of group work, as does understanding of how group members interact and contribute to the completion of group tasks. Differentiation is considered a tactic by which learning can be made accessible to all pupils in the classroom (Department for Children Schools and Families, 2008; Scottish Office Education and Industry Department, 1996). Thus teachers may distribute different tasks to different classroom subgroups as some evidence suggests teachers appear to understand that guidance and recommendations encourage the adoption of such processes to accommodate different learning styles or pupils’ varying subject understanding (Tobin & McInnes, 2008). However, as indicated earlier the scarcity of research considering how small groups operate within secondary education suggests it would be unwise to assume that uniform approaches will be observed. The observation materials used in the present study were sufficiently flexible to allow classroom interaction to be recorded independently of pupils’ seating and task arrangements. A systematic approach of recording observations was utilised. This method was employed whilst adopting a naturalistic approach - observation undertaken within classrooms avoided interfering with regular practice.
The second area of investigation focuses on pupils’ talk as particular forms of dialogue have been demonstrated as having greater potential for learning when pupils engage in group work (Barnes & Todd, 1977; Littleton et al., 2005; Mercer, Dawes, Wegerif, & Sams, 2004). For example, when pupils counter statements such action not only stimulates the understanding shared by group members but also goes some way to solidify an individual’s own beliefs and understanding. Solicited helping behaviours, where an individual receives help having requested it, are argued to be fundamental for encouraging group activity and maintaining on-task behaviour (Terwel, Gillies, van den Eeden, & Hoek, 2001). These include explanations, which have been consistently identified as helping both provider and recipient (Webb, 1982, 1991; Webb & Mastergeorge, 2003)and which provide opportunities for group members to align their learning and knowledge as a consequence of helping (Gillies, 2002; Oortwijn, Boekaerts, Vedder, & Fortuin, 2008). Thus pupils may seek assistance from their group rather than their teacher (Gilles, 2003). As regards the discussion of opposing perspectives, it undoubtedly can occur when small group work is governed by specific conditions (Howe & Tolmie, 2003),indicating that the standard of interaction suggested by theory can be approached(Piaget, 1962). In contrast, features contributing to effective groups are rare within authentic primary classrooms and talk is neither task enriching nor task related (Galton et al., 1999). In light of the above this study will examine the nature of pupil interaction and evaluate teachers’ behaviour by comparing observation recorded within conventional teaching with that achieved when pupils work in groups.
Method
Design
For pupils; a between-subjects design incorporated teaching approach (group work or conventional), subject (science or English), gender (male or female), topic stage (beginning, middle or end of topic) and year (first or third)[2]. For teachers; a mixed design incorporated a within-subjects factor - observation at the introduction and conclusion of lessons (lesson stage) and between-subjects factors – teaching approach (group work or conventional), subject (science or English), gender (male or female), topic stage (beginning, middle or end of topic) and year (first or third). Topic stage was treated as a between-subjects factor as difficulties encountered during data collection meant that on some occasions obtaining three recordings per participating class was not feasible. 5 instances of substitution occurred. Forexample, if a class was recorded within two group work lessons, but an additional group work observation was not feasible,a third observation of the classon occasion occurred within a conventional lesson.
Observations were recorded during science and English classes, because a preliminary survey of mathematics, English, modern languages, history and science teachers indicated that group work occurs with particularly high frequency within science and English (MacQuarrie, 2006). Within the current study first and third year classrooms were identified as suitable targetsas they are not subject to pressures such as external examinations or subject choice decisions. Access to teaching years where pupils encounter such challenges may be restricted. Therefore to produce a representative picture of small group work thestudy was restricted to two specific subjects and teaching years.
Sampling
Participants attended eight secondary schools located in West Central Scotland. These schools varied (0-5% to 30-40%) in their percentage of free school meal entitlement (FSE), where higher values indicate low family income and socio-economic status(Scottish Executive, 2009b). Threeschools (providing 8 (35%) of the 23 participating classes) were based within the 10-15% entitlement category, close to the national average for Scottish secondary pupils(Scottish Executive, 2008, 2009b).
[Insert Table 1 about here]
The overview of the sample in Table 1 means that clear guarantees of representativeness are beyond the scope of this naturalistic study. Contextual factors influenced the number of pupils observed within each classroom. Equivalent numbers of observations were recorded with regard to each topic stage and gender. Significant differences within the number of observations recorded are signalled with an asterisk in Table 1. Some differences relate to the naturalistic approach of the study and others require explanation. One point to be noted in relation to teaching year & subject is that within third year two sets of classes – chemistry and physics – were categorised as Science (in comparison with first year combined science lessons). This effectively doubles the volume of observations recorded of third year science teaching. The mean number of observations recorded was computed by dividing the total number of pupil observations (443 science, 252 English) with the number of lessons observed for each subject (34 lessons science, 16 lessons English), giving separate descriptives for science (M = 13) and English (M = 15.75).
An overview of visits made to participating teachers is presented in Table 2. Since each class was paired with a single teacher, the number of participating teachers was effectively equivalent to the number of participating classes (23). As noted earlier 5 instances of substitution occurred (where a teacher was recorded partly in both group work and conventional lessons).
[Insert Table 2 about here]
Significant differences regarding the number of teacher observations are signalled with an asterisk in Table 1. As noted earlier, first year teaching involves general subjects, whereas third year science involves discrete science thereby doubling the number of observations collected in third year science lessons. In contrast, significant differences regarding gender and lesson type indicate difficulties in obtained balanced samples.
Observation approach
Observation followed a naturalistic approach, where an observer coded lessons in vivo, but did not disturb classroom interaction. Teachers were asked not to make any particular efforts on behalf of the researcher. Teachers volunteered on the understanding that when observations were recorded within group work lessons, a further teacher who was content to have similar observations recorded within conventional lessons would also be sought. Systematic observations utilised a time-sampling procedure, facilitated by the use of a grid, with rows corresponding to observation windows and columns to observation categories. Observation grids and categories are presented in Appendix 1.
Identical grids each comprised of eight consecutive windows of 60 seconds in length were used to observe teachers at the introduction and conclusion of lessons. The duration of each observation window allowed statements used by teachers to be accurately recorded. The first of three sections concentrated on descriptive statements relating to lesson content, the second section focused on measures used to aid interpretation of material, and the final section examined dialogue. It seemed likely that teachers would model the dialogue or interaction that they wished to see within groups (Webb et al., 2008), therefore aspects of dialogue from the SCOTSPRinG project shown to be effective within pupil groupings were incorporated within the observation categories(Howe et al., 2007). Behaviours exhibited by teachers were recorded once within each observation window. Multiple coding, which would have required the researcher to distinguish between the boundaries of one statement and the next, could not be reliably assessed within a single observation window.
Pupils were also observed for eight windows. The duration of each window approximated 15 seconds, based on the approach utilised within SCOTSPRinG research (Howe et al., 2007). A specific sampling technique was employed: one pupil per small group was observed, each successive observation would attend to a different group, and whenever possible operate selection would be based on gender – observation of a male pupil would be followed by the observation of a female pupil. This sampling technique operated in the context of the variable nature of lessons, which influenced the opportunities to make observations (e.g., pupil behaviour would not be recorded when a teacher read poetry or gave a demonstration). Consequently a pre-determined number of pupil observations could not be set. An overview of classroom interaction was obtained by observing a variety of pupils during each classroom visit, rather than repeatedly recording specific pupils’ behaviour within and across lessons.
Sections one to three of the pupil observation grid each employed mutually exclusive coding categories. Pupils were coded as exhibiting behaviours within a specific category, when that behaviour had the longest duration within an observation window. The contents of each observation grid were summed, so that frequencies within each observation category were computed and used within the analysis. This approach was used previously within the SPRinG group work intervention and the SCOTSPRinGresearch (Baines et al., 2007; Howe et al., 2007;Thurston et al., 2008). The final section focused on the dialogue produced by pupils and multiple dialogue behaviours were permitted here, in contrast to the mutually exclusive categories used within the previous sections. Essentially, this approach followed that used successfully in the recent SCOTSPRinG transition project (Thurston et al., 2010).
Practicalities meant that the data for this study had to be recorded by one researcher; therefore reliability checks on the observation methodology were made in advance. In vivo coding of lessons, with two observers present in the same classroom, would have been impractical and ethically complicated; instead appropriate video resources were used. Inter-rater agreement over teacher observations was computed using Cohen’s Kappa and for this purpose observations were treated as a binary variable and classed in relation to their presence or absence. This approach avoided biases that may have resulted from teachers showing variability in their behaviour. Intraclass correlation coefficients (McGraw & Wong, 1996) were more appropriate for use with the interval-scale pupil data. The researcher and an observer trained for this purpose coded the recordings, with agreement relating to each of the categories ranging between .7 and 1.0(Landis & Koch, 1977; Wragg, 1999). Mean agreement between raters over the pupil observation categories (.89) and teacher observation categories (.84) was high.