Effects of Tracking/Ability Grouping

The Effects Of Tracking/Ability Grouping On All Levels Of Students

Abstract

The research detailed in this paper provides a systematic description and analysis of classroom grouping practices in primary and secondary schools in England. Practices are compared to main findings in developmental and educational literature with regard to effective contexts for learning and recent ideas about pedagogy. The research is based on an analysis of 4924 pupil groupings from 672 Reception, Year 2 and Year 5 classes in 331 primary schools and 248 Year 7 and Year 10 classes in 47 secondary schools. The data came from ‘classroom mapping questionnaires’ that were completed by teachers at a particular point in the school. Completed questionnaires provided information about the nature and use of pupil groupings within their classrooms and focused on the number and size of groupings, type of working interaction between pupils, the presence of adults, grouping composition and the type of task that groupings were engaged with. Results showed that there were changes in grouping practices with pupil age. As pupils increased in age they were increasingly likely to experience whole class ability based sets (tracking) for core curriculum subjects and more formal row/pair seating arrangements.

Chapter One: Introduction

Statement of the Problem

All pupils in schools are grouped in some form or another. At a school level, pupils are organised into classrooms on the basis of decisions about age and ability mix (Dean, 1992; Dreeben, 1984). Classes can be viewed as nested contexts within a school ( Bronfenbrenner, 1979) and within classrooms there are further nested contexts. Within the classroom the teacher is responsible for making decisions regarding these units or nested contexts for instruction and learning, that is, the unit to which learning tasks and working interactions are co-ordinated. This unit we have termed the ‘grouping’ and it can consist of anything from a single pupil to a whole class of pupils. There are many possible grouping contexts and each has different implications for pupils’ learning. Groupings can be of different sizes and compositions, and can vary in the amount of adult support they receive, the curricula and tasks they are given and the degree and quality of interaction between pupils.

Decisions on the groupings to use for particular learning purposes should be partly dependent on the pupils themselves. The way children respond to the grouping contexts the teacher sets up and the interactive and learning benefits pupils take from them will depend on the skills and knowledge that they bring to these contexts. Of course, these skills and understandings will vary not only between pupils but also over time within pupils. During the 12 years that children spend in US schools, between the ages of 4 and 16, large and dramatic developments are apparent in children’s social, cognitive and communication skills. Thus particular combinations of grouping structure, learning task, working interaction and adult assistance may be more, or less, supportive of learning and development at different points in childhood. A pedagogic approach to the teaching of, and learning by, pupils of different ages must take account of these interactions between grouping and learning.

Previous studies have provided clear findings but they have not examined classroom groupings in any great depth and do not systematically examine within-class grouping practices relative to the age (or learning needs) of children. They have tended to focus exclusively on either infant or junior classrooms (and not on secondary classrooms—11–16 years) and thus do not consider grouping contexts in relation to children’s development or consider how teachers treat children differently at different ages. This paper therefore seeks to examine groupings used within primary and secondary school classrooms in terms of a number of core dimensions and further to analyse them in relation to the effective practices suggested by developmental and educational research. The five core dimensions examined are: the size and number of groupings in the class; the working arrangement between grouping members; adult support of groupings; grouping composition; and the curriculum area and task type undertaken by the grouping.

Purpose of the study

This paper aims to provide a naturalistic and multidimensional description, in terms of the five core dimensions, of the nature of grouping practices as employed in primary and secondary school classrooms. Changes in grouping practices with pupil age will also be examined relative to the practices reported by teachers of Reception, Year 2 and Year 5 classes at the primary school level and teachers of Year 7 and Year 10 classes, at the secondary school level. A further aim is to consider how the description relates to developmental research and theory about the conditions for effective learning.

Research Questions

What are the effects of tracking/ability grouping on all levels of students?

Significance Of Study

The current study goes beyond previous work by exploring the nature of the grouping context in relation to learning task, working interaction and adult involvement across different year groups in primary and secondary schools. In contrast to other papers in this volume, this study is deliberately descriptive of current practice. The study is informed by an ecological view of classroom life and current ideas concerning a pedagogic approach to teaching and learning.

Chapter Two: Review of literature

Galton and Williamson’s (1992) review identified four distinct types of classroom groupings: individuals, pairs, small and large groups and whole class. Yet their existence does not mean that teachers effectively co-ordinate their teaching and learning tasks with them. Research has examined the educational implications of grouping size for effective learning ( Kutnick, 1994) and emphasises that small groups are the most effective for learning ( Lou et al., 1996). However, this research did not relate grouping size to child development and has not examined actual grouping sizes used in classrooms. The number of groupings found in a classroom at any one time also has implications for the learning activity and adult involvement but has not been considered fully in relation to the age of pupils. An analysis of grouping size and number is also important in relation to the working arrangements of the grouping and the task that is undertaken.

A central consideration in the nature and use of groupings is the type of working arrangement between members of the grouping. A range of working arrangement types in classrooms has been described as: individualised work (children work on unique individual tasks designed for their specific needs); individuated work (children work on the same task but are expected to work alone); individuated work with talk (children work on the same task, are expected to do the work alone but are allowed to talk); peer interactive work (children either work on separate sub-components of one task or work together on a single task with a shared goal); and work with a teacher either as a whole class or as a group (Bennett & Dunne, 1992; Galton & Williamson 1992; Kutnick & Rogers, 1994).

The research literature on the ‘science’ of teaching and learning highlight the potential of different types of working interactions for learning. Piagetian, Vygotskyan, and socio-cultural theories have prompted much research into individual, adult-child and peer interactions in equal (cooperative and collaborative) and unequal (peer tutoring) relations. The general view is that learning is more likely and deeper when pupils are actively engaged in learning. Learning is often more active when pupils work in a shared social context involving interaction with others, whether with an adult or other pupils, rather than when working alone or just listening to the teacher (Wood, 1998; Rogoff, 1990; Doise & Mugny, 1984; Topping, 1994; Forman & Cazden, 1985; Slavin, 1990; Johnson & Johnson, 1987; Tolmie, Howe, Duchak, & Rattray, 1998).

Research suggests that from at least 6 to 7 years children can begin to engage in and benefit from ‘collaborative’ interaction, though younger children do engage in social learning through co-ordination and cooperation as well as imitation and instructed learning (Azmitia & Perlmutter, 1989; Tomasello, Kruger, & Ratner, 1993). Collaborative interaction requires the use of cognitive skills, perspective taking and particular conversational skills to compare other’s perspectives with one’s own ( Baines, 1996; Miller, 1987; Piaget (1928, trans. 1959); Tomasello et al., 1993). Theorists and researchers from a Vygotskyan and socio-cultural tradition emphasise that interaction with others is productive prior to the age 6–7 years and they usually focus on interactions between a child and more able other (adult or child) ( Rogoff, 1990; Hogan & Tudge, 1999) and peer interactive contexts where children scaffold each other’s learning, engage in the co-construction and elaboration of ideas or give and receive help ( Forman & Cazden, 1985; Forman, 1992; Webb, 1989). These interaction types may be described as interactants co-ordinating and co-operating in interaction rather than being specifically collaborative ( Baines, 1996; Tomasello et al., 1993).

There are indications that, for interactive group-work, smaller grouping sizes may be more suited to younger children or children with limited communication skills (Baines, 1996; Fuchs et al., 2000; Smith, 1994). Younger children will not have the conversational strategies, confidence or experience to enable them to easily interact in large groups. Large and even small group situations make it difficult for children to co-ordinate the taking of turns in conversation and keep interruptions to a minimum, unless regulated by an adult or more able other who can help them stay on one topic ( Dorval & Eckerman, 1984). Moreover, in situations where children are expected to plan and organise their interactions, large groupings will make this very difficult. By contrast, dyadic and triadic groupings provide a simpler context within which children can develop many of the conversational and social-cognitive strategies needed for interacting in larger groups. These contexts also reduce the ‘risk’ ( Doyle, 1980) to self esteem that is central to increasing the effectiveness of group work (see Blatchford, Kutnick, Baines, & Galton, this volume). There are, of course, settings where individual work is necessary and more productive than group-work ( Howe, Duchak-Tanner, & Tolmie, 2000; Murphy & Messer, 1998) allowing children time to reflect on ideas and knowledge. These situations may be used to prepare for group work or to accommodate thinking after group work.

In classroom settings, small groupings may be more effective when pupils are expected to interact on a task (Bosert et al., 1985; Nasasti & Clements, 1991). On the other hand, a few large groups may allow more control of behaviour and attention as well as enable the targeting of guidance and support. Equally, from a management point of view lots of small groupings may prove unwieldy for teachers, particularly when pupils have difficulties working together on tasks. However, there is little information available on relationships between working interaction type and group size relations at different points in pupils’ education.

The composition of groupings may have implications for teaching and learning. From a teaching perspective, homogenous ability grouping is efficient but may not be as beneficial for pupils’ learning. Most experimental research has focused on whether grouping members should be of similar or mixed ability. During peer interaction, the process of cognitive conflict, which theory suggests underlies collaborative group work, requires a difference in perspective and ability (Doise & Mugny, 1984; Piaget (1928, trans. 1959)). Furthermore, peer tutoring requires a large difference in ability ( Rogoff, 1990; Vygotsky, 1978). However, similar ability peers may be able to scaffold each other during interaction ( Forman & Cazden, 1985). Webb (1989), in her research on peer helping, recommends a compromise position such that high and middle ability pupils work together and low and middle ability pupils work together. This strategy overcomes the common problem of high ability pupils becoming annoyed with having to help low ability pupils, while maintaining differences in perspectives and knowledge to support learning such that pupils of all ability levels benefit. Homogenous ability grouping within classes is currently recommended by school inspectors in England and Wales ( OFSTED, 1995), despite a scarcity of research into the effectiveness of this form of grouping ( Hallam & Toutounji, 1996). Setting classes by ability is becoming ever more prevalent in US primary schools and has been the norm for pupils above 13 years of age in secondary schools despite evidence emphasising negative social, emotional and academic outcomes ( Benn & Chitty, 1996; Hallam & Toutounji, 1996). Yet little is known about its incidence in primary schools or the prevalence of within-class ability grouping in primary and secondary schools.

Adult support of groupings is essential with regard to guidance, instruction and support for learning, but also from a practical point of view in terms of effective management of behaviour and attention. The existence of an adult in a group will affect interactions between pupils and thus may be highly inappropriate under certain task and interactive activities. Theoretically driven research highlights the beneficial effects of an adult scaffolding an individual’s learning but holds little faith in the support or instruction of larger groupings where a teacher must know about, and adjust their teaching to, all children in their class (Tharp & Gallimore, 1991; Wood & Wood, 1996). When taught as a whole class, pupils will vary in their ability and learning needs, even when placed in classes ‘set’ by ability (where pupils are re-organised into new classes for one or more curriculum area on the basis of similarity in ability—also known as tracking in the US), and thus direct whole class teaching can be difficult. This is further compounded by the possibility that not all pupils will be actively engaged all of the time. Research by Galton and colleagues also illustrates the unfeasibility of adults scaffolding pupils’ learning. They found that while teachers spend most of their time interacting with pupils, each pupil on average received no more than 10 min of focused teacher attention and support per day (Galton, Simon, & Croll (1980) and Galton, Hargreaves, Comber, Wall, & Pell (1999).

Adult support may be more important with younger children as they may have difficulties remaining on-task or working independently as a group or alone, and may need guidance for the learning and elaboration of new knowledge and concepts. Despite the crucial nature of adult deployment in relation to the groupings used in classrooms, little research has examined adult support of groupings of different sizes and compositions and in relation to different curricula and tasks and across different year groups. However, teachers report that small groupings are preferable as they allow better quality input from an adult, allow better quality work from pupils, are better for pupils’ concentration, and are more manageable generally (Blatchford, Baines, Kutnick, & Martin, 2001a). But in large classes teachers feel driven to reduce the number of groupings and thus increase the size of groupings for ease of management and control.

Finally, the nature of the curriculum area and task that groupings work on is a key component of an analysis of the effective use of groupings. Particular curriculum subject cultures may lead to different teaching and learning practices (Goodson & Managan, 1995). Yet if activities are not appropriate to the particular type of grouping arrangement, then learning may be ineffective or threatening ( Bossert, Barnett, & Filby, 1985; Doyle, 1983; Galton & Williamson, 1992). Currently, in the US, it is suggested that in the ‘literacy hour’ different activities are conducted at the same time when pupils are working in groups. This may allow teachers to target their support to particular groupings but this has implications for the complexity of task types given to other groups where no adult support is available. This may be particularly problematic with young children that are less able to work independently.

The type of task undertaken has been conceived across many different dimensions (Alexander, Schallert, & Hare, 1991; Bloom, 1956; Norman, 1978; Pica, Kanagy, & Falodun, 1993). Norman (1978) suggested that tasks could either introduce new ideas, procedures or skills; demand that an individual discovers ideas; require the practising of new skills on familiar problems; require the application of old knowledge/skills to new areas or problems; or revise or recap knowledge/skill. In the experimental research literature there are clear indications that different types of task are more suited to different types of grouping and adult support of groupings ( Kutnick, 1994). For example, the learning of new skills and procedural knowledge may be best conducted by a tutor ( Howe et al., 2000; Rogoff, 1990); practice tasks may be better suited to working alone with adult support to clarify problems or as a whole class ( Jackson & Kutnick, 1996); tasks involving the application of skills to new areas may be best suited to collaborative group working situations as individuals have already honed the necessary skills but need to develop new conceptual understandings ( Howe et al., 2000). There is also an indication that the task needs to encourage ambiguity ( Doyle, 1980), whether in terms of outcome or process, in order to be effective for group work. Different tasks may be evident at different age levels, especially as during the early years children are learning and developing new skills and knowledge. Bennett et al. (1984) found that tasks used in infant classrooms primarily involved the practising of skills. Few studies have examined the types of tasks that are given to different types of groupings in primary and secondary classrooms. If teachers are to be effective in their use of grouping strategies, they must be aware of the potential for learning and have the ability to co-ordinate task types with groupings in their classrooms.

Chapter Three: Methods

Research Design and Data Collection

This paper aims to provide a naturalistic and multidimensional description, in terms of the five core dimensions, of the nature of grouping practices as employed in primary and secondary school classrooms. Changes in grouping practices with pupil age will also be examined relative to the practices reported by teachers of Reception, Year 2 and Year 5 classes at the primary school level and teachers of Year 7 and Year 10 classes, at the secondary school level. A further aim is to consider how the description relates to developmental research and theory about the conditions for effective learning.

Analysis of Data

The data used in this paper come from three separately funded but parallel studies that used the same methodology. One project, the Primary Classroom Groupings Project, examined grouping practices in Year 2 (6–7 years) and Year 5 (9–10 years) classrooms. The focus of the second study was on the effects of class size on pupil learning experiences and involved data on grouping practices in Reception (4–5 years) classes. The third study examined grouping practices in secondary schools (11–12 years and 14–15 years). All three projects involved the use of a ‘grouping mapping questionnaire’ to collect quantitative data on the nature of groupings as used in classes at a specified time and day. This paper only reports part of the total data collected—other articles summarise other sections of the data (Blatchford et al., 2001a; Kutnick et al., 2002; Blatchford, Kutnick, Clark, McIntyre, & Baines, 2001b).

The grouping mapping questionnaire

On an assigned day and sample time in the lesson, teachers were asked to make a quick note (on a predrawn map of their classroom) of the location of individual male and female pupils, the grouping that they were part of, the curriculum and task that groupings were working on, and the nature of the working interaction the grouping was engaged in. The teacher also noted the location of the adults working in the classroom. Later, at a convenient moment, teachers completed a questionnaire where they elaborated on the classroom map and provided further information on the class and themselves (such as whether the class was set by ability and class size). The advantage of this approach is that it utilises the benefits of a large scale questionnaire and an observational approach while avoiding many of the difficulties associated with these methods, such as the lack of detail and expense respectively. The grouping questionnaire completed by Reception class teachers was a simplified version of the questionnaire used with teachers of Years 2, 5, 7 and 10.

Sample

The primary classroom groupings project

Schools were contacted across 5 participating Local Education Authorities (LEAs). Of these 5 LEAs, 3 covered sub-urban areas in the south of England and 2 covered inner city areas, one in the South and one in the North West of England. LEAs with schools in rural areas were not approached because of the strong likelihood of mixed age classes which would have required further detailed analyses in areas that the project had not set out to address. Schools in the 4 LEAs in the south of England were contacted prior to sending questionnaires and schools in the fifth LEA were sent questionnaires directly. Teachers were asked to complete the questionnaires at a set time (one of five possibilities), on a particular day (Tuesday, Wednesday or Thursday) during weeks when they were not overburdened (i.e. weeks where Standard Assessment Tasks and OFSTED inspections were taking place and the first and last weeks of term were avoided). Times were selected to avoid school assembly, break-times and the lunch hour. Returned questionnaires were evenly distributed across the five completion times.

Just less than half of the schools approached agreed to participate in the project and over half of these returned questionnaires. From the 111 schools that participated, 187 teachers returned questionnaires. About half of these teachers were of Year 2 classes and half taught Year 5 classes. Some teachers taught mixed age classes (always consisting of pupils from one year above or below), although all had a predominance of Year 2 or 5 pupils. Over 1000 groupings were described in the questionnaires of the 187 classes and there was a fairly even contribution from both Year 2 and 5 classes.

The class size project: Reception class data on grouping practices

In the class size study, questionnaires were completed at 10:00 a.m. on a specified day of the week. Questionnaires were returned by 485 Reception class teachers (from 220 schools which were randomly sampled from 8 LEAs). These questionnaires provided data on over 2000 groupings. Further information on this project can be found in Blatchford, Moriarty, Edmonds, and Martin (2002) and Blatchford et al. (2001a).

Grouping practices in secondary schools project

The grouping practices in secondary schools project involved 47 schools in the collection of classroom mapping data, 36 were mixed sex schools, nine were all-girl schools and two were all-boy schools. Questionnaires were returned from 128 Year 7 and 120 Year 10 classes which were spread equally across the four subject areas of English, Mathematics, Science, and Humanities. Teachers were asked to complete questionnaires either towards the start, in the middle or towards the end of the lesson. Percentages of questionnaires returned were in the proportions 20%, 54% 26%, for the different timings of the lesson respectively; very close in number to those sent out. These questionnaires provided data on 1767 groupings. Further information on this project can be found in Blatchford et al. (2001b).

Chapter Four: Results

Results were analysed using Chi-square, Spearman’s correlation and, where appropriate, ANOVA with Tukey hsd post hoc tests. This section will briefly consider class level contextual data before examining the data on within-class groupings. While many comparisons between core dimensions are possible, here we have focused only on those that are relevant to changes in grouping practices over the years of primary and secondary school (see Blatchford, Baines, Kutnick, & Martin (2001a) and Blatchford, Kutnick, Clark, McIntyre, & Baines (2001b); Kutnick et al., 2002, for further analyses).

Streaming and setting classes by ability

All Reception classes and the vast majority of Years 2 and 5 classes were of mixed ability. Not one single class was streamed (i.e. where pupils are allocated to a class where they are taught all subjects on the basis of ability) and only a quarter of classes experienced any form of ‘setting’. Year 5 classes were more likely to be set by ability where as many as 44% were set for a subject compared to only 6% of Year 2 classes (χ2(1)=35.98;p<0.001). Setting was mainly for Mathematics or Mathematics and English—only once was it found for English alone. However, only 18% of Year 5 classes actually worked as a set when the questionnaire was completed. At the secondary school level setting was much more frequent. At Year 7, 42% of classes were set by ability and this increased to 70% by Year 10 (χ2(4)=20.38;p<0.001). Setting at Year 7 was primarily for Mathematics and Science where approximately 50% of classes were set by ability and 35% of classes studying English and Humanities (e.g. History, Geography, Art) were set. At Year 10 setting was even more prevalent for English, Mathematics and Science (69%, 100% and 83% respectively) but much less likely for Humanities subjects (19%) (χ2(3)=48.40;p<0.001).

Classroom seating and working arrangements

Small group seating was the predominant classroom layout at both Year 2 and 5 levels (59%) (this was not coded at the Reception level). However, there were differences between Year 2 and 5 classrooms in the next most popular class layout (χ2(3)=25.0;p<0.001). While at Year 2 there was more large group seating than at Year 5 (37% vs 14% respectively), at Year 5 there was more traditional row/pair class seating (0% vs 15% respectively). By secondary school, pair and row seating was the predominant classroom layout (67% at Year 7 and 75% at Year 10) with the remaining classrooms having layouts consisting of a combination of small, large, row and paired seating.

Number and size of groupings within classrooms

The average number of groupings in classes increased with pupil age. Typically, Reception classes contained fewer groupings than Year 2 classes, and Year 2 classes fewer than Years 5, 7 and 10 (F(4,915)=28.9; p<0.001). By contrast, groupings decreased in size with age, with Reception and Year 2 children working in the largest groupings and Year 7 and 10 pupils working in the smallest (F(4,4902)=61.5; p<0.001).

Grouping composition

Group composition varied across the age groups (χ2(4)=433;p<0.001). While pupils of all ages, even Reception, were most likely to be in similar ability groupings, by Year 10 this had reached the point where virtually all pupils were grouped according to similarity in ability. The levels of low and middle ability groupings remained fairly constant across all year groups but during the secondary stage, levels of high ability groupings increased and mixed ability grouping decreased.

Curriculum area and task type

Age differences were evident in the types of task given to pupils (χ2(12)=2913;p<0.001). In Reception and Year 2 classrooms, practice and revision tasks were the most common task type while in secondary classrooms pupils were most likely to be working on tasks involving the application of existing knowledge. There was an equal balance between these two task types at Year 5. Thus, as pupil age increased, groupings were more likely to be given tasks where they were expected to apply their existing knowledge to new areas and less likely to be practising or revising their skills. Tasks involving the introduction of new knowledge remained relatively constant over primary and secondary school levels.

Working interaction type in relation to grouping size

Both the primary (Years 2 and 5 only) and secondary school data show that working alone was common in all group sizes except groupings of 11 or more and in large groupings at the secondary school stage . In primary classrooms, peer interactive work was most often conducted in small groups but, proportionally, dyads and triads were more likely to involve peer interactive work than other types of working arrangement (χ2(10)=4231;p<0.001). At the secondary level, individuated work was most often experienced in pairs and, proportionally, triads and small groups were more likely to be associated with peer interaction than other working arrangements (χ2(10)=6102;p<0.001). At both primary and secondary school stages, groups of 11 or more pupils (usually whole classes) were most likely to be listening to and interacting with the teacher than other group sizes.

Adult presence in relation to the number and size of groupings

The relationship between the number and size of groups and adult presence is revealing about the effect adult support has on the organisation of the class, the way teachers allocate their time among the groupings, and the function of different grouping sizes. The number of adults in classes increased with the number of groupings at Reception (r=0.36;n=476;p<0.001) and Year 2 (r=0.24;n=92;p<0.05) but not at Years 5, 7 or 10. At both primary and secondary school levels, adults were most likely to be present with very large groupings and large groups of 7–10 (Primary—χ2(5)=3772;p<0.001; Secondary—χ2(5)=4558;p<0.001). Small groups, triads and especially dyads were least likely to have an adult present and were thus the main grouping size where pupils work independently. However, as age increased adults were decreasingly likely to support individuals (χ2(4)=128.9;p<0.001). Thus in Reception classes individual pupils were more likely to have an adult present than not, but by the Secondary stage very few individuals, dyads, triads, small or large groups were supported by an adult.

Task type in relation to grouping size

In primary classrooms most types of learning task were conducted in small groups and thus there was no distinct relationship between grouping size and task type. However, some trends are apparent if the data are looked at within rather than across grouping sizes. At the Reception level, large groupings of 7–10 pupils were proportionally more likely to be engaged on tasks involving the introduction of new information (χ2(15)=1463;p<0.001). At Years 2 and 5, dyads were more likely to work on tasks involving the application of skills than other task types and very large groupings were most likely to be gaining new knowledge than working on other task types (χ2(15)=691.3;p<0.001).

The findings at the secondary school stage indicate a clearer relationship between task type and grouping size (χ2(15)=884.9;p<0.001). Groupings of 11 or more pupils were most likely to be working on tasks involving the introduction of new knowledge. The application of existing knowledge was most connected to dyads (the most common grouping size at this level) but also small groups. Practice and revision tasks were most often conducted in dyads and very large groupings invo