Abstract
The metaphor of scaffolding has been found to be a useful and inspiring metaphor to describe the temporary and tailored support a teacher can give to a student. However, its interactive nature makes the measurement of scaffolding difficult; to our knowledge, no instrument for measuring scaffolding in classroom situations is available. In this chapter, two different instruments for analysing the scaffolding process in teacher small-group interactions are presented. We focus on what we see as the most important feature of scaffolding: its adaptive nature, or its contingency. We build on two contingency-oriented frameworks, the more qualitative and general model of contingent teaching and the more quantitative and detailed micro-level contingent shift framework. We present an analysis of 29 interaction fragments from a larger corpus of data on scaffolding.
In these interaction fragments, prevocational social studies teachers support students (7th and 8th grade; 12-15 years old) performing open-ended tasks. We present detailed analyses that demonstrate that both frameworks are suitable for exploring the variability in contingency in teacher – small-group interactions. The models are complementary in that the first, general model distinguishes different phases in scaffolding interactions and emphasises the importance of the first phase (that is, diagnostic strategies), while the second, micro-level model provides detailed insight into the extent to which teachers adapt their teaching in response to student actions and the nature of these adaptations.
The model of contingent teaching appears to be especially useful for practice and for teachers’ professional development. The contingent shift framework appears to be the most useful for scientific purposes.
6 This chapter is based on:
Van de Pol, J., Volman, M., Elbers, E., & Beishuizen, J. (2012). Measuring scaffolding in teacher – small-group interactions. In R. Gillies (Ed.), Pedagogy: New Developments in the Learning Sciences. Hauppage: Nova Science Publishers.
INTRODUCTION
Although there is keen interest in small-group work (Cohen, 1994; Johnson &
Johnson, 1974; Mercer & Littleton, 2007; O’Donnell, 2006), merely placing students in groups does not mean that learning will necessarily occur (Chiu, 2004; Gillies, 2006).
High-quality teacher support – not only in terms of group work or collaborative learning skills but also in terms of content – is therefore vital.
Not just one particular type of support is effective in small-group work; being contingent or adaptive to students’ understanding has been pinpointed as a crucial characteristic of effective support (Chiu, 2004; Van de Pol, Volman, & Beishuizen, 2010; Webb, 2009). Contingency, or a teacher’s adaptation of his or her support to a student’s understanding, is central to the concept of scaffolding (Wood, Bruner, &
Ross, 1976) which is, therefore, a powerful metaphor to describe and analyse the interaction between a learner and a more knowledgeable other. Since the introduction of this concept in 1976 by Wood et al., research on scaffolding has been ubiquitous.
The concept of scaffolding has been applied and studied in many different contexts, such as parent-child interactions (Mattanah, Pratt, Cowan, & Cowan, 2005; Pino-Pasternak, Whitebread, & Tolmie, 2010), teacher-student interactions (Mercer & Fisher, 1992; Nathan & Kim, 2009), student-peer interactions (De Guerrero & Villamil, 2000), computer-student interactions (Azevedo, Cromley, Winters, Moos, & Greene, 2005), and computer-student-teacher interactions, also known as distributed scaffolding (Tabak, 2004).
Regardless of the context, scaffolding is, by nature, interactive because it always involves at least two actors (e.g., a teacher and a learner). This interactive nature makes the measurement of scaffolding difficult. Especially in naturalistic classroom situations, such as small-group work, measuring scaffolding is a challenge because teachers must work with more than one student at a time, and tasks for small-group work are often not highly structured. To our knowledge, no measurement instrument has been developed to characterise and measure scaffolding in classroom situations while taking into consideration this interactive nature (Van de Pol et al., 2010). Such an instrument, however, is needed for several reasons. First, it will enable us to compare the results of different studies in different classroom contexts. Second, it enables research, and thus understanding, of the effects of classroom scaffolding on student learning. Third, a deeper understanding of what scaffolding looks like and how it can be measured will facilitate teacher training; scaffolding can be more easily explained, taught, and assessed. Therefore, the aim of this study was to develop instruments for characterising and measuring the process of scaffolding in teacher – small-group interactions that take into account the interactive nature of scaffolding. In this chapter,
we present the instruments we have developed, explain the usage of these instruments and, via the use of these instruments, illustrate the scaffolding process.
The Concept of Scaffolding
The metaphor of scaffolding is derived from construction work, where it represents a temporary structure that is used to construct a building. Scaffolding has been found to be a useful and inspiring metaphor for studying the temporary and tailored support a teacher can give to a student in the Zone of Proximal Development (Mercer & Littleton, 2007; Stone, 1998a; Stone, 1998b). Within Vygotsky’s sociocultural theory (1978), the Zone of Proximal Development (ZPD) represents the distance between what a student can do independently (the actual understanding) and what a student can do with the help of a more knowledgeable other (the potential understanding). The support that is given in the ZPD is often referred to as scaffolding (Mercer & Littleton, 2007).
Scaffolding can be characterised by three main features: (1) contingency, (2) fading, and (3) transfer of responsibility (Van de Pol et al., 2010). The first and most important feature is contingency, which represents the adaptive nature of the support that can be labelled as scaffolding. Not all forms of support can be called scaffolding;
the support needs to be adapted to or contingent upon the students’ understanding.
Fading and transfer of responsibility are closely related in that the scaffolding support should decrease over time to transfer the responsibility for the task or for learning to the student. However, fading and transfer of responsibility can only be accomplished effectively if they are performed in a contingent way. Therefore, we see contingency as the most crucial characteristic of scaffolding.
After Wood et al. introduced the concept of scaffolding in 1976, Wood, Wood, and Middleton (1978) adopted precisely this focus on contingency in their empirical study. They studied the effects of contingency on four-year-old children’s mastery of a construction task in one-to-one tutoring situations. Contingent tutoring was defined by two basic rules: when the learner fails, increase control; when the learner succeeds, decrease control7. These rules were called the contingent shift principle. All tutors’ turns were coded according to the degree of control they exerted, ranging from low control (general verbal encouragement) to high control (demonstration). Children’s turns were coded according to their understanding. In this way, every three-turn sequence of a teacher’s turn, the student’s turn and the teacher’s turn could be categorised as 7 Note that the degree of control is inherently connected to the degree to which a student is challenged; when a teacher increases control, the student is less challenged to think independently, and when a teacher decreases control, the student is more challenged to think independently.
contingent or noncontingent. A three-sequence turn was considered contingent if the teacher increased control in reaction to an incorrect response by a student and decreased control in reaction to a correct response by a student. Contingency thus depends on the teacher’s adaptation in reaction to the student’s understanding.
Gathering information about the students’ actual understanding is crucial in contingent teaching or scaffolding. A teacher needs this information to be able to judge whether to increase or decrease the level of control. The model of contingent teaching (Van de Pol, Volman, & Beishuizen, 2011) stresses the need for gathering information about the student’s actual understanding. This model is based on the work of Wood et al.
(1978) and the formative assessment literature, which promotes diagnostic strategies for contingent teaching (Ruiz-Primo & Furtak, 2007). The model of contingent teaching consists of four teaching steps (see Figure 1).
Figure 1. Model of contingent teaching
The steps of the model of contingent teaching are: (1) applying diagnostic strategies (discovering a student’s actual understanding); (2) checking one’s diagnosis with the student (summarising what the student said and asking whether this is correct to create common understanding or intersubjectivity); (3) applying intervention strategies (the actual support, adapted to the information gathered in steps 1 and 2); and (4) checking a student’s understanding (verifying whether the student learned something). Although this model focuses on the steps that a teacher can take in contingent teaching, the responses of the students are crucial to the model. A teacher needs the responses (demonstrations) of the students to determine his or her next step. Therefore, although the model adopts the teacher’s perspective in the scaffolding process, we argue that it does not neglect the interactive nature of scaffolding.
The contingent shift framework developed by Wood et al. (1976; 1978) and the model of contingent teaching both provide the means to describe and measure classroom scaffolding from an interactive perspective. Table 1 presents the characteristics of the two frameworks.
Table 1
Characteristics of the Model of Contingent Teaching and the Contingent Shift Framework
Model of contingent teaching Contingent shift framework
Type Qualitative Quantitative
Granularity Macro Micro
Focus Steps of contingent teaching Contingency
Unit of analysis Teacher turn Three-turn sequence of
teacher turn – student turn – teacher turn
The Model of Contingent Teaching: Research
Ruiz-Primo and Furtak (2007) explored the ways in which three science teachers elicited responses from students and how the teachers used the information they gathered. They found that, in most cases, teachers elicited information about their students’ understanding; however, they did not use this information in their guidance of student learning. A case study on the ways in which three social studies teachers organised and supported small-group work (Van de Pol et al., 2011) used the model of contingent teaching to reveal patterns of contingent teaching in three of each teacher’s lessons. The separate steps of contingent teaching were identified in interaction fragments (i.e., conversations between a teacher and a small group of students), and the interaction fragments, as a whole, were considered contingent when “the teacher was judged to use information gathered about the students or students’ understanding in his provision of support to the student(s)” (p. 5). As in the study by Ruiz-Primo and Furtak (2007), contingent teaching was found to be generally scarce, partially because the support was not adapted to the students’ understanding. However, in contrast to the findings of Ruiz-Primo and Furtak (2007), Van de Pol et al. found that teachers hardly elicited information from students (e.g., through diagnostic questions). Other studies have also demonstrated teachers’ failure to adequately diagnose students’
understanding prior to instruction. Lockhorst, Wubbels, and van Oers (2010) examined the educational dialogues between two excellent biology teachers in Montessori secondary schools (preuniversity track) and small groups of students or individual students. These dialogues demonstrated that these teachers rarely employed the diagnostic language genre, in which the teacher asks information-seeking questions to explore a student’s understanding.
Elbers, Hajer, Jonkers, Koole and Prenger (2008) analysed the one-to-one interactions between two mathematics teachers and their students at two secondary schools. The results indicated that these teachers rarely explored the exact nature of the students’ problems, but instead started to give explanations right away. Therefore, diagnostic strategies (step 1 in the model of contingent teaching) were rare, and
teachers started instruction by providing support (step 3).
Chiu (2004) studied one teacher’s interventions during small-group work. The students (9th grade) were working on algebra problems. The teacher’s evaluations of the students’ progress (i.e., diagnostic strategies) appeared to have the greatest effect on students’ subsequent time-on-task and progress in problem solving. Performing diagnostic strategies before providing students with support thus appeared to be crucial.
The model of contingent teaching appears useful for a qualitative and systematic description of the scaffolding process from a contingency perspective and, especially, for pointing out what steps teachers are and are not taking. However, these analyses are of a qualitative nature and remain rather general. Even if a teacher performs all steps in the model of contingent teaching, we cannot be sure whether the teacher is truly adapting his or her level of control to the students’ understanding. This understanding requires a micro-level analysis of the interactions between students and teachers.
The Contingent Shift Framework: Research
In parent-child studies, Wood’s contingent shift framework has been used regularly (e.g., Mattanah et al., 2005; Pino-Pasternak et al., 2010; Pratt, Green, MacVicar, &
Bountrogianni, 1992). Pino-Pasternak et al. (2010) for example studied parents’
interactions with their children while helping them with their mathematics homework.
The effects of authoritative parenting (i.e., “a balanced and contingent display of positive affect, responsiveness, and demandingness” p. 221) and scaffolding on the children’s performance were studied. The contingent shift framework was used to measure the degree to which parents used scaffolding; the rules of contingent teaching were applied to three-turn sequences composed of a parent’s turn, a child’s turn and a parent’s turn. The degree to which parents’ turns were contingent appeared to be a crucial and unique predictor of the child’s performance, above and beyond the effects of authoritative parenting. Pino-Pasternak et al. (2010) demonstrated that the contingent shift framework could be reliably applied to situations in which a parent helped a child on homework involving authentic and less-structured tasks. However, all of the studies mentioned here took place in one-to-one, parent-child settings; classroom settings include more diverse and complicated types of interactions.
The contingent shift framework has never been applied before in research on scaffolding in classroom settings. Our review of classroom scaffolding over the last decade has indicated that the majority of studies on scaffolding were descriptive and qualitative (Van de Pol et al., 2010). Sometimes, criteria were used to establish whether scaffolding took place. For example, Oh (2005) used the criteria developed by Maybin, Mercer, Stierer, and Norman (1992), i.e., there had to be evidence that the mentor’s
support was contingent, that the learner accomplished the task with the mentor’s situated help, and that the learner performed the task independently (see also Mercer
& Fisher, 1992). However, scaffolding has more often been described in more general terms.
When scaffolding was measured in a more quantitative way, often only the teacher’s support was taken into account. This approach does not adequately consider the interactive nature of the construct. One exception is the work of Nathan and Kim (2009). Although they did not refer to the work of Wood et al. (1976; 1978), their methodology was similar. The study examined one teacher’s regulation of conceptual reasoning and participation in mathematics whole-class interactions. The study explored not the degree of control exerted by the teacher, but rather the teacher’s adaption of the elicited cognitive complexity to the students’ understanding. The teacher tended to shift the level of cognitive complexity in response to students’ answers; he reduced the cognitive complexity in response to incorrect answers, but he increased the cognitive complexity in response to correct answers.
Although the contingent shift framework has never been applied in studies on scaffolding small-group work, the existing literature on small-group learning stresses the importance of contingency; it is the degree to which the support is contingent upon the students’ understanding and needs, as opposed to the type of support provided (e.g., specific strategies such as high vs. low content support), that is considered crucial (Chiu, 2004; Webb, 2009).
The contingent shift framework was originally intended to analyse interactions between two people; therefore, it was necessary to make some changes to adopt it to the analysis of teacher–small-group interactions. The unit of analysis consists of a teacher turn, a student turn, and a teacher turn. However, in a group of up to about five students, each student might give different responses indicating different levels of understanding. The teacher needs to react to meet the needs of each student.
Therefore, the unit of analysis needs reconsideration, as well as the contingency rules on deciding upon which student(s) the teacher is being contingent.
Furthermore, in the original contingent shift framework, two types of turns (i.e., claims and demonstrations) that, in reality, do not provide the same level of information regarding the students’ actual understanding were used interchangeably.
Koole (2010) distinguished between a claim of understanding (e.g., “I get it”) and a demonstration of understanding (e.g., an explanation or reason), based on Sacks (1992). A demonstration gives a teacher and a researcher much more information about a student’s understanding than a claim does; this information can be used as a tool to direct contingent instruction. If a student merely claimed not to understand a particular concept, the resulting contingent act, according to the original contingent
shift framework, would be to increase control. However, we argue that a teacher really needs a demonstration of a student’s actual understanding to be able to take contingent action in response. Therefore, we argue that the framework requires further adaptation to address the distinction between claims and demonstrations. These issues are addressed when describing and applying the contingent shift framework.
The Current Study
In this study, we sought to develop instruments for characterising and measuring the process of scaffolding in teacher–small-group interactions, taking into account the interactive nature of scaffolding. The model of contingent teaching was used to develop an instrument for qualitative analysis of the process of scaffolding in teacher–
small-group interactions. Wood’s contingent shift framework was adapted to develop a coding scheme for microanalysis of contingency in teacher–small-group interactions.
The research question that is explored in this study is the following: How can classroom scaffolding in teacher–small-group interactions be analysed from a contingency perspective that takes the interactive nature of scaffolding into account?
Two research questions are formulated:
1. How can the model of contingent teaching be applied to measure contingency in teacher – small-group interactions?
2. How can the contingent shift framework be applied to measure contingency in teacher – small group interactions?
To answer these questions, we analysed 29 interaction fragments from a larger corpus of data on scaffolding. In these interaction fragments, prevocational social studies teachers support 7th- and 8th-grade students working on open-ended tasks.
METHODS
Participants
Thirty prevocational social studies teachers, each with one 7th- or 8th-grade class, participated in this study. We recruited the teachers by placing a call for participation in several online teacher communities. Of the 30 teachers, 20 were men and 10 were women. The average teaching experience of the teachers was 9.7 years. The mean class size was 26.8 students, and the lessons lasted 53.2 minutes on average. Most
teachers and their classes were used to working on thematic projects and in small groups. On average, 22.5 percent of the children in each class were from immigrant families (i.e., one or both parents were born abroad).
In total, 768 students participated in this study. However, because only one fragment of an interaction between each teacher and a group was selected, the total number of students for this study was 73. The average number of students within one group was 4.1, and groups ranged from 2 to 6 students. All students were between 12 and 15 years old.
Materials Data
The data come from an experimental study with a between-subjects design and pre and postmeasurement data collection (see Van de Pol, Volman, Oort & Beishuizen, resubmitted). Thirty teachers participated in this experimental study, 17 in the scaffolding condition and 13 in the nonscaffolding condition. All teachers taught the same five-lesson curriculum on the European Union (EU). The teachers in the scaffolding condition additionally participated in a scaffolding intervention. In this intervention, teachers were trained according to the model of contingent teaching. In each session, they focused on one of the steps of contingent teaching, and in the last lesson (postmeasurement), they focused on all of the steps of contingent teaching. Each teacher’s first and last project lessons were videotaped (the premeasurement and postmeasurement sessions).
The data come from an experimental study with a between-subjects design and pre and postmeasurement data collection (see Van de Pol, Volman, Oort & Beishuizen, resubmitted). Thirty teachers participated in this experimental study, 17 in the scaffolding condition and 13 in the nonscaffolding condition. All teachers taught the same five-lesson curriculum on the European Union (EU). The teachers in the scaffolding condition additionally participated in a scaffolding intervention. In this intervention, teachers were trained according to the model of contingent teaching. In each session, they focused on one of the steps of contingent teaching, and in the last lesson (postmeasurement), they focused on all of the steps of contingent teaching. Each teacher’s first and last project lessons were videotaped (the premeasurement and postmeasurement sessions).