Teacher peer coaching with digital video. Evaluation of a four-year professional development program

Teacher peer coaching with digital video. Evaluation of a four-year professional development program Paper presented at the Annual Meeting of the American Educational Research Association symposium Teacher Autonomy in Using Digital Video for Professional Learning of Practicing Teachers Dr Niels Brouwer Tuesday, April 14, 2009 2:15 pm 3:45 pm San Diego Marriott Hotel and Marina, Marriott Hall, Salon 1 Dr C.N. Brouwer ILS Graduate School of Education Radboud University Nijmegen P.O. Box 9103 6500 HD Nijmegen The Netherlands Phone: +31 24 3615739 Email: n.brouwer@ils.ru.nl

2 CONTENTS p. Introduction 3 Context 3 Research questions 7 Research design 12 Findings 14 What did teachers learn? 15 How did teachers learn? 19 Conclusions and discussion 30 References 32 FIGURES 1. Visual teacher learning model 9 2. Research model 11 3. Relationships between background variables, characteristics of the learning environment, learning activities and perceived learning value 29 TABLES 1. Project goals and criterion variables 5 2. Program overview 6 3. Composition of cohorts 14 4. Response to questionnaires 14 5. Response to plenary meeting evaluation forms 15 6. Base line and personal learning goals 16 7. Learning outcomes 18 8. Interpretation format for learning outcomes 20 9. Comparison of base line, personal learning goals and learning outcomes 22 10. Patterns of development 23 11. Characteristics of the learning environment 24 12. Learning activities 24 13. Presence during plenary meetings 25 14. Perceived learning value 27

3 Introduction Peer coaching is a valuable approach to professional development (PD) for teachers. In a review of 29 peer coaching projects, Ackland (1991, p. 23-24) distinguishes between coaching by experts and reciprocal peer coaching and notes three common characteristics of such projects: Peer coaching is distinct from evaluation [in the sense of teacher appraisal]. [ ] 2. Peer coaching includes observation followed by feedback. [ ] 3. Peer coaching focuses on improving instructional techniques. In this paper, results are presented from an evaluation of a professional development project in which digital video (DV) was used to support reciprocal peer coaching by four cohorts of ten to twenty experienced teachers. Teachers observed and filmed each other during teaching, selected personally significant fragments from the available footage and viewed and discussed these together with their peers, both in pairs and in plenary meetings. The use of digital video is a potentially significant addition to reciprocal peer coaching. We carried out a literature review into the use of digital video for PD (Brouwer, 2007) with three guiding questions: - What do teachers learn when using DV? - How do teachers learn when using DV? - In what conditions do teachers learn when using DV? On the basis of 21 empirical studies carried out worldwide, we concluded that for teachers, imaging teaching competence, be it one s own or colleagues teaching competence, has the potential of influencing professional action. Moreover, indications were found that this influence extends over time.we therefore expect that learning with the help of visual media can have a significance for the development of teachers professional competence and motivation in the longer term. Context From 2004 to 2008, we had the opportunity to support and study a reciprocal peer coaching project in the Urban Gymnasium Nijmegen. This school with about 1.000 pupils is situated in the city of Nijmegen, which lies in the east of the Netherlands near the German border. It provides secondary education preparing for university entry. Over the course of four years, roughly half of the teaching staff participated in the peer coaching project. From the first year onwards, a steadily increasing use was made of digital video to support the participating teachers in observing each other s lessons. The goals of the project were formulated as the consequence of a quality assurance initiative by the deputy school leader. He organized for 36 class mentors to observe during one school day all the lessons of the class entrusted to them. Both these pupils and the mentor teachers filled out questionnaires about how varied and motivating they found the instructional formats in the lessons attended and observed. Analysis and discussion of these data led to one important conclusion, shared by the teaching staff and the school s management team, i.e. We need to - increase the variation of learning activities in order to activate learners, - offer more opportunities for differentiation and - increase pupils autonomy as learners (Scheepens, 2004). These three goals became the focus of the reciprocal peer coaching project whose evaluation is presented in this paper.

4 Three important principles were adhered to in designing and implementating the project. In the first place, teachers participated voluntarily and could choose freely with which colleague they wanted to form a pair. These pairs worked together during a whole school year in observing, filming, discussing and planning lessons together. Secondly, every participating teacher was respected and recognized for participating, regardless of employment status, teaching experience or anything else. This principle implied that any problem, challenge, question or perspective a participant would deem important was valued in its own right. Finally, all participants were free to choose their own personal learning goals. Although in the beginning the school management was sometimes concerned that this might allow for serious departures from the project goals, this risk never materialized. We summarized these three principles as: - volunteering, - egalitarianism and - ownership and voice. Table 1 shows how the project goals were operationalized in terms of criterion variables. These three lists were used by the participating teachers for selecting personal learning goals at the beginning of and throughout the school year. They were also used in questionnaires for monitoring the PD project. Table 2 shows the typical program as carried out during a school year by each cohort of participants. The Urban Gymnasium Nijmegen gave all participants one lesson unit of release time per week to enable them to visit, observe, film and discuss each other s lessons. With regular intervals, all participants came together in plenary meetings, where background materials about instructional issues and approaches was given (documented both in a reader and an electronic learning environment), coaching skills were trained, experiences were shared and video clips of each other s lessons were discussed. They received technical support from the school s audiovisual department and the author. During the four years of its existence, the project was coordinated by two teachers and the author, who consulted regularly with a steering committee consisting of the deputy school leader and one other member of the management team.

Table 1. Project goals and criterion variables No. 1. Content EXPLAINING I. INCREASING VARIETY IN CLASSROOM ACTIVITY FORMATS Criterion variable explain subject matter stepwise 2. PRECISE INSTRUCTIONS give precise instructions for exercises and assignments 3. HOMEWORK coordinate classroom activity and homework 4. INSTRUCTIONAL DIALOGUE use instructional dialogue to engage pupils 5. INDEPENDENT WORK have pupils work by themselves 6. GROUP WORK have pupils cooperate in groups 7. PEER COMMENT have pupils comment each others work 8. PEER ASSESSMENT have pupils assess each others work 9. PUPIL PRESENTATION have pupils present to the group 10. DEBATE SUBJECT MATTER have pupils debate subject matter together 11. GAMES engage pupils with subject matter in the form of games 12. ICT use ICT in the classroom II. INCREASING OPPORTUNITIES FOR DIFFERENTIATION No. Content Criterion variable 13. WORK OUTSIDE CLASSROOM have pupils work also outside the classroom 14. ADDITIONAL MATERIAL have pupils study additional material 15. ADDITIONAL ASSIGNMENTS have pupils work on additional assignments 16. OWN PACE have pupils work at their own pace 17. OWN TOPICS have pupils work on topics of their own choice 18. COMPUTER WORK have pupils work with the computer III. PROMOTING SELF-DIRECTED LEARNING No. Content Criterion variable 19. PUPIL PLANNING have pupils plan their own work 20. PUPIL AS CO-TEACHER have pupils present part of the lesson 21. INTEREST IN SUBJECT have pupils develop an interest in my school subject 22. REFLECT ON FUTURE SUBJECT have pupils reflect on what they want to do with my school subject in the future

Table 2. Program overview Plenary meetings Themes Training and learning Pair work 1. August / September Observation and feedback Peer coaching Formulating personal learning needs and goals. Planning activities in pairs 2. October 3. November 4. January 5. February / March Principles of direct instruction. Activating learners Electronic learning environment. Video use Focusing personal learning goals Cooperative learning Role play and feedback / discussion Lesson planning Filming lessons Homework. Self-directed learning Viewing and discussing video recordings of each other s teaching Lesson planning Filming lessons Educative ICT uses Differentiation Sharing views and discussing practices concerning topics of own choice. Viewing and discussing video recordings of each other s teaching Filming lessons 6. April 7. June Sharing views and discussing practices concerning topics of own choice. Viewing and discussing video recordings of each other s teaching Program evaluation Poster presentations about each other s work and learning Monitoring Entry questionnaire Group discussion of progress Pupil consultation Exit questionnaire Work in classrooms: observing each other / peer coaching / preparing and trying out instructional approaches together

Research questions The research questions guiding this evaluation were: 1. What did the participating teachers learn? More specifically: What effects on their thoughts and actions did they attribute to participation? 2. How did the participating teachers learn? More specifically: Which factors, notably the learning environment created for peer coaching and the teacher learning activities undertaken in that environment, contributed to the participants professional learning? Teacher learning is here conceived in terms of both thought and action, i.e. - discovering and developing new ideas for teaching as well as - experimenting with new behaviors and changing teaching practices. Conceptually, this study is based on the literature review mentioned above. From this review, the Visual Teacher Learning model was derived. This VTL model builds on earlier work by Clarke & Hollingsworth (2002) and specifies the following nine factors as influencing teacher learning with digital video. 1. teachers personal characteristics, biography and learning goals 2. the change environment in which they work on their professional development 3. the learning activities they undertake for this purpose in cooperation with colleagues 4. the perceptual processes involved in these activities 5. the kind(s) of feedback they receive 6. how they interpret this feedback 7. the reflection processes involved in the foregoing 8. consequences for professional action and finally 9. their personal development in the teaching profession The studies analyzed in the review shed some light on how these factors may be related, in other words which circumstances can influence teachers learning with DV, in which ways this learning evolves and which impact it can have on their development in the profession. From the perspective of research design, this formulation refers to relevant independent, intervening and dependent variables. The Visual Teacher Learning model represents possible relationships between the nine factors mentioned. The model is meant as a conceptual map to serve theorizing and research about teachers learning with DV. The reasoning behind the Visual Teacher Learning model is as follows. Personal characteristics and biography are background factors which may influence whether, how and what a teacher learns when participating in a preservice teacher education curriculum or a professional development program in which his or her personal and/or others teaching competence is visualized. As relevant characteristics can in any case be seen a person s gender, age, teaching experience and work load. In addition, it is to be expected that a person s professional motivation and significant experiences in working as a teacher play a role. These background factors will probably influence what kinds of personal learning goals a teacher is given and/or chooses in the context of preservice teacher education or workplace learning.

8 What and how a teacher can learn is not only determined by his or her background, but also by the learning environment that a teacher education program or an employer offers. Here, facilitation is an important factor in different respects. Naturally, the availability of work time is an influence, but also the composition of the collegial learning group, the opportunities provided for interaction, the types of activities undertaken and the facilities available for personal and electronic collaboration. These factors determine in which context teachers can undertake learning activities. The learning activities themselves, the ways in which they are structured and the collegial interactions that they invite immediately influence individual learning. As the studies reviewed indicate, it is important how teachers visual learning is framed. Framing refers in particular to how perception is influenced by the nature of the visual material (trigger videos, modeling videos and/or action videos (cf. Fortkamp & Van den Berg, without year)), its source (whether other-viewing or self-viewing is involved (cf. Tochon, 1999)) and the viewing instructions given or chosen. Individual learning is also considerably influenced by the feedback teachers receive from various sources. Sources can be colleagues (peer feedback), the self as when a person alone views a recording of a lesson given by himself or herself (self feedback), learners (learner feedback or more generally client feedback) and/or a teacher educator or PD facilitator (facilitator feedback). It is to be expected that the perception of visual material and the feedback a teacher receives will together influence his or her interpretation of it. As made explicit in the Cognitive Development Process Model developed by Chan & Harris (2005), the interpretive process involves a series of cognitive activities. These cognitive activities can be conceived of as a normatively guided assessment of intentions and decisions concerning the use, elaboration and change of a teacher s personal repertoire of teaching activities. This assessment involves what Korthagen (et al., 2001, p. 44) describes in his ALACT model as the phases of Creating alternative methods of action (phase 4), Trial (phase 5) and Action (phase 1 of a new reflection cycle). Reflection can take place in more or less conscious ways. It is therefore to be expected that when a teacher changes his or her actions, this need not per se or predominantly be a consequence of explicit reflection. Teachers actions constitute a plastic entity with a dynamic of its own and a powerful influence on a person s development as a professional (Brouwer, 1989, p. 59-104; Brouwer & Korthagen, 2005, p. 154-157).

9 Figure 1. Visual Teacher Learning model CHANGE ENVIRONMENT Entry characteristics Biography Goals Perception - Framing - Self-viewing - Other-viewing Reflection Interpretation - Examining - Valueing Feedback - Peer - Self - Learner - Facilitator Enactment Ongoing development of - teaching competence - professional motivation TEACHERS LEARNING ACTIVITIES

10 The Visual Teacher Learning model served as the basis for the research model shown in Figure 2, which was employed in this study. The latter model specifies in which independent, intervening and dependent variables we were interested. The independent variables are divided into three types or model components. As background variables we included age, gender and teaching experience as well as participants school subjects and whether they participated once or more often. Secondly, we included characteristics of the learning environment, notably the composition of the peer coaching pairs, whether participants felt they had enough autonomy for deciding about what and how to teach in their classrooms and their percentage of presence during plenary meetings. Thirdly, we wanted to know how the teachers viewed their own teaching skills at the beginning of their participation and which goals they set for their personal learning. From our literature review it was clear that process variables play an important intervening role in how results of teacher learning with digital video come about. The model therefore specifies the learning activities undertaken, notably how many lessons participants observed, filmed, viewed on video and discussed. In addition, it specifies which forms of feedback were given and received during peer coaching and how the participating teachers experienced this feedback. Finally, the research model includes the learning outcomes of the project. Just as the base line data and the personal learning goals, these were conceptualized in terms of the criterion variables in Table 1. The arrows in the research model denote which possible relationships between the model components might explain the learning processes and outcomes involved.

11 Figure 2. Research model Independent Variables Intervening Variables Dependent Variables Background variables Participation Age Gender Teaching experience School subject Learning activities Number of lessons - observed - filmed - viewed - discussed Learning outcomes Criterion variables - thought - action Learning environment Composition of pairs Autonomy in classroom Presence during plenary meetings Perceived learning value Viewing and giving and receiving feedback on own and colleague s lesson recordings Base line and personal learning goals Criterion variables - thought - action

12 Research design On the basis of the research model in Figure 2, quantitative as well as qualitative data were collected from all four cohorts. For quantitative data, entry and exit questionnaires were administered at the beginning and end of each school year, respectively. In the first cohort, only an exit questionnaire was administered, because the criterion variables and the other measures were developed during the first project year. In the second, third and fourth cohorts, repeated measures of the criterion variables were taken. The other variables were contained in the exit questionnaire only. The criterion variables were measured as follows. At the beginning of the school year, the participating teachers were asked to state in the form of yes-or-no questions whether they felt they already had a command of the teaching skills specified. At the same time and also with yes-or-no questions, they were asked to state whether they selected these skills as personal learning goals for the school year ahead. These questions permitted an operationalization of their professional action and thought, respectively. At the end of the school year, the teachers were asked to state, again in the form of yes-or-no questions, about which teaching skills they felt they had acquired new insights and which skills they felt they had developed during the past school year. The exact wordings used in combination with the descriptions in Table 1 were: for teacher thought at the beginning of the school year: Through peer coaching with video I want to get to know (more) ways to in my teaching. for teacher thought at the end of the school year: Peer coaching with video has helped me get to know (more) ways to in my teaching. for teacher action at the beginning of the school year: In my teaching, I usually. for teacher action at the end of the school year: Peer coaching with video has helped me (more often) in my teaching. Besides teacher questionnaires, also pupil questionnaires were administered in classes selected by participants in the second and third cohorts, in which they felt they had attempted most intensely to apply what they learned during the peer coaching project. The following qualitative data were collected: - field notes by the author, - verbal evaluation forms completed by the participating teachers at the end of plenary meetings and - video recordings of - lessons given by participants - plenary meetings, in which they shared experiences, performed role plays and discussed video fragments from their lessons. The data basis for this paper is restricted to the teacher questionnaires and the plenary meeting evaluation forms. The data used in this paper were analyzed as follows.

13 For the quantitative data, descriptive statistics were produced. In addition, possible associations and relationships between variables were explored by means of crosstabulations and correlations, respectively. The plenary meeting evaluation forms contained two categories, in which the participating teachers wrote comments about the progress they made in the pair work with their colleague and about the organisation of this pair work. In two additional categories, they gave comments on how the plenary meetings were conducted and voiced their needs and suggestions for these meetings. On the qualitative data collected in this way, content analysis was performed as follows. All the entries made by the teachers were categorized under headings denoting similar topics. Then, for each topic summaries of their remarks were written for each cohort separately. Finally, from these summaries the findings pertaining to all cohorts were derived.

14 Findings In this paper, findings are reported for all teachers who participated in the four consecutive cohorts from the school year 2004-2004 through the school year 2007-2008. Table 3 shows the composition of the four cohorts. The total number of participants is 45. However, two women and eleven men participated twice. These teachers came to be known as veterans. Six veterans participated in the first and second cohorts. The other seven veterans participated in the fourth cohort, three from the first cohort, one from the second and three from the third. These figures mean that in all, 45 participants are responsible for 58 participations. Please note that all of the following analyses are based on data concerning these 58 participations. In other words, the unit of analysis is participation, not participant. From table 3, it can be seen that both the total number of participants per cohort and the number of new participants per cohort has decreased over the four project years. Table 3. Composition of cohorts School year 2004-2005 2005-2006 2006-2007 2007-2008 Total of participants Participations 20 18 10 10 45 Women 7 7 5 2 18 Men 13 11 5 8 27 Veterans 0 6 0 7 13 New participants 20 12 10 3 Table 4 shows the response to the questionnaires. This response is sufficiently high to make adequate statements about the project, except perhaps for the last cohort. Table 4. Response to questionnaires 2004-2005- 2006-2007- Total 2005 2006 2007 2008 entry entry exit entry exit entry exit Number of participants 20 18 18 10 10 10 10 58 Response 14 17 17 9 9 5 5 45 % response 70.0% 94.4% 94.4% 90.0% 90.0% 50.0% 50.0% 77.6% Nonresponse 6 1 1 1 1 5 5 % nonresponse 30.0% 5.6% 5.6% 10.0% 10.0% 50.0% 50.0% 22.4% Table 5 shows how many plenary meetings were evaluated by means of written forms and the response achieved. The number of meetings was reduced by two in the last cohort. Most meetings were evaluated in the third cohort and least in the second and fourth. For those meetings evaluated, the response was generally high.

15 Table 5. Response to plenary meeting evaluation forms School year 2004-2005 2005-2006 2006-2007 2007-2008 Number of plenary 7 7 7 5 meetings Frequency of 3:7 = 43% 1:7 = 14% 4:7 = 57% 1:5 = 20% evaluation Response per meeting no. 3: 13:20 = 65% 4: 11:20 = 55% 4: 15:18 = 83% 2: 6:10 = 60% 3: 9:10 = 90% 3: 7:10 = 70% (absolute and relative) 5: 12:20 = 60% 5: 7:10 = 70% 6: 6:10 = 60% What did teachers learn? In table 6, the descriptions in the left-hand column refer to the criterion variables in table 1. The shaded rows show the percentages of teachers who felt at the beginning of their participation that they routinely practiced the teaching skills distinghuished. The white rows contain the percentages of teachers who chose the teaching skills mentioned as personal learning goals. Percentages at the right of the columns headed yes and no exceed the middle mark of 50. The modes in the right-hand column 1 for yes and 0 for no correspond of course with the percentages. Of the three project goals, increasing variety in classroom activity formats lay closest to the participants existing practice. By far the most teachers found themselves proficient at giving direct instruction (criterion variables 1 through 5). Still, a majority wished to delve more into this approach. Clearly fewer teachers felt they had a command of cooperative learning techniques (criterion variables 6 through 8). Almost all of them wanted to get more acquainted with this approach. Less current activity formats (criterion variables 9 through 12) were practiced in the classroom by almost a third of the participants. A small majority wanted to develop insights into these activity formats. Differentiation turned out to be a knotty subject for most participants. Having pupils work with additional materials, also outside the classroom (criterion variables 13 through 18), was routine for a minority, although working at your own pace and at topics of your own choice was a pupil experience allowed for by more teachers. Computer use in the classroom (criterion variable 18, cf. 12) was not standard practice. A majority of participants was willing to study these options further. Self-directed learning did not occur habitually in our respondents teaching practice, at least in the sense that pupils were allowed to co-determine how lessons were carried out (criterion variables 19 and 20). Trying to develop intrinsic interest in the school subject (criterion variables 21 and 22), however, was an ambition for most teachers, probably within the setting of direct instruction and as we saw above, less so of cooperative learning. These issues had the interest of most participants.

16 Table 6. Base line and personal learning goals Criterion variables Valid N % yes % no mode I. Increasing variety in classroom activity formats 1. EXPLAINING 21 52 48 1 28 96 4 1 2. PRECISE 23 52 48 1 INSTRUCTIONS 27 74 26 1 3. HOMEWORK 19 32 68 0 28 82 18 1 4. INSTRUCTIONAL 22 73 27 1 DIALOGUE 28 79 21 1 5. INDEPENDENT 23 70 30 1 WORK 28 86 14 1 6. GROUP WORK 26 89 11 1 29 38 62 0 7. PEER COMMENT 26 81 19 1 30 27 73 0 8. PEER ASSESSMENT 25 84 16 1 30 23 77 0 9, PUPIL 24 63 37 1 PRESENTATION 29 52 48 1 10. DEBATE SUBJECT 24 67 33 1 MATTER 28 32 68 0 11. GAMES 28 82 18 1 30 30 70 0 12. ICT 12 67 33 1 14 14 86 0 II. Increasing opportunities for differentiation 13. WORK OUTSIDE 24 63 37 1 CLASSROOM 30 23 77 0 14. ADDITIONAL 26 85 15 1 MATERIAL 29 14 86 0 15. ADDITIONAL 22 59 41 1 ASSIGNMENTS 29 35 65 0 16. OWN PACE 26 81 19 1 29 48 52 0 17. OWN TOPICS 26 73 27 1 30 67 33 0 18. COMPUTER WORK 27 67 33 1 29 24 76 0 III. Promoting self-directed learning 19. PUPIL PLANNING 24 63 37 1 27 33 67 0 20. PUPIL AS CO- 26 81 19 1 TEACHER 31 16 84 0 21. INTEREST IN 26 58 42 1 SUBJECT 29 79 21 1 22. REFLECT ON 24 58 42 FUTURE SUBJECT 30 50 50

17 Table 7, laid out analogously to table 6, shows the learning outcomes as measured at the end of the school year. Concerning the first project goal, increasing variety in classroom activity formats, a clear majority of participants felt that they had improved their skills in direct instruction and that they had (further) developed relevant insights (criterion variables 1, 4 and 5). This, however, was less the case for giving precise instructions and coordinating classroom activity and homework (criterion variables 2 and 3). Still, the overall picture is that the teachers did maintenance work for and polished existing skills in this area. Moreover, a clear majority felt they had improved their skills in and knowledge about cooperative learning (criterion variable 6). Almost half the teachers made progress regarding having pupils comment and assess each other s work (criterion variables 7 and 8). Finally, the less current activity formats remained less used (criterion variables 9, 10 and 12), excepting game forms. A majority of teachers got to know and made (more) use of games in their lessons (criterion variables 11). Concerning the second project goal, increasing opportunities for differentiation, about a third of the participants developed knowledge and skills (criterion variables 13 through 18). Having pupils work at their own pace (criterion variable 16), especially, increased. In the area of self-directed learning, about a quarter of the teachers allowed pupils more influence on how lessons were carried out (criterion variables 19 and 20). A small majority felt they succeeded (better) in rousing interest in their school subjects and in having pupils reflect on the potential use of the subject for their futures (criterion variables 21 and 22).

18 Table 7. Learning outcomes Criterion variables Valid N % yes % no mode I. Increasing variety in classroom activity formats 1. EXPLAINING 43 49 51 0 44 50 50 2. PRECISE 41 63 37 1 INSTRUCTIONS 42 22 78 1 3. HOMEWORK 41 22 78 0 41 24 76 0 4. INSTRUCTIONAL 41 59 41 1 DIALOGUE 40 53 47 1 5. INDEPENDENT 44 82 18 1 WORK 44 71 29 1 6. GROUP WORK 43 79 21 1 44 80 20 1 7. PEER COMMENT 43 51 31 1 45 49 51 0 8. PEER ASSESSMENT 42 40 60 0 43 44 56 0 9, PUPIL 42 40 60 0 PRESENTATION 43 39 61 0 10. DEBATE SUBJECT 31 23 77 0 MATTER 30 27 73 0 11. GAMES 31 61 39 1 30 60 40 1 12. ICT 14 21 79 0 14 14 86 0 II. Increasing opportunities for differentiation 13. WORK OUTSIDE 44 23 77 0 CLASSROOM 43 33 67 0 14. ADDITIONAL 43 30 70 0 MATERIAL 43 33 67 0 15. ADDITIONAL 42 26 74 0 ASSIGNMENTS 43 33 67 0 16. OWN PACE 43 44 56 0 44 39 61 0 17. OWN TOPICS 43 35 65 0 44 29 71 0 18. COMPUTER WORK 44 29 71 0 44 32 68 0 III. Promoting self-directed learning 19. PUPIL PLANNING 44 26 74 0 44 25 75 0 20. PUPIL AS CO- 41 32 68 0 TEACHER 42 29 71 0 21. INTEREST IN 42 56 44 1 SUBJECT 39 64 36 1 22. REFLECT ON 43 16 84 0 FUTURE SUBJECT 42 9 91 0

19 How did teachers learn? During preliminary analyses on the data of the second cohort, indications surfaced that the participants were learning far more and/or different things than they had formulated as learning goals for themselves at the beginning of the school year. We explored this possibility by means of a systematic comparison of their scores on base line, personal learning goals and learning outcomes. This was done by crosstabulating the measurements of each of the criterion variables with those at the beginning. The numbers of valid cases involved in these analyses did not allow interpretable values of chi square to be calculated, but the raw frequencies emerging in the cells did provide some illuminating insights. Table 8 shows an interpretation format for the crosstabulations performed. The cells in the upper part of this table specify in general terms what it means when a respondent has scored one of four possible combinations of yes or no on a criterion variable at the beginning and at the end of the school year. The middle part of the table does the same for the criterion variables concerning teacher thought and the lower part for those concerning teacher action. Using this format one can interpret the frequencies resulting from the crosstabulations as patterns according to which teachers insights and skills may develop. When a teacher scored no both at the beginning and at the end of the school year, we interpreted this as a pattern of standstill, i.e. no personal learning goal was set and no learning outcome was achieved (pattern A). When a teacher scored yes at the beginning of the school year and no at the end, we interpreted this as a pattern of failure, i.e. a personal learning goal was set, but no learning outcome was achieved (pattern B). When a teacher scored no at the beginning of the school year and yes at the end, we interpreted this as a pattern of serendipitous learning, i.e. no personal learning goal was set, but a learning outcome was achieved nevertheless (pattern C). When a teacher scored yes both at the beginning and at the end of the school year, we interpreted this as a pattern of success, i.e. a personal learning goal was set and a learning outcome was indeed achieved (pattern D).

20 Table 8. Interpretation format for learning outcomes General Exit questionnaire Thought Exit questionnaire Action Exit questionnaire No Yes No Yes No Yes No A Standstill No personal learning goal set, No outcome achieved C Serendipity No personal learning goal set, Outcome achieved No A Standstill No need for insight felt, No new approach discovered Insight remains out of reach C Serendipity No need for insight felt, New approach discovered Discovery made No A Standstill No routine, No change of action Skill remains out of reach C Serendipity No routine, Change of action New skill developed Entry questionnaire Entry questionnaire Entry questionnaire Yes B Failure Personal learning goal set, No outcome achieved D Success Personal learning goal set, Outcome achieved Yes B Failure Need for insight felt, No new approach discovered Intended insight remains out of reach D Success Need for insight felt, New approach discovered Intended insight achieved Yes B Failure Routine exists, No change of action No development of existing skill D Success Routine exists, Change of action Further development of existing skill

21 Using the interpretation format in table 8, it was possible to discern different patterns of development for different (groups of) criterion variables by inspecting the frequencies appearing in the respective crosstabulation cells. These frequencies are listed for each development pattern in the columns of table 9. As in tables 6 and 7, the white rows refer to teacher thought and the shaded rows to teacher action. Most successes (pattern D) can be noted in the area of the first project goal, increasing variety in classroom activity formats. About two thirds of the teachers succeeded in having pupils work more for themselves (criterion variable 5). Also, a large majority succeeded in improving giving precise instructions and implementing cooperative learning (criterion variables 2 and 6). However, the latter successes were also to a considerable extentunintended or serendipitous learning outcomes (pattern C). The skills of explaining subject matter stepwise and using instructional dialogues (criterion variables 1 and 4) apparently developed according to a different combination of patterns. Here, teachers apparently faced the possibilities of either failure (pattern B) or success (pattern D). Five skills turned out to be hard to acquire or develop (pattern A), i.e. coordinating classroom activity and homework, having pupils comment and assess each others work, having pupils debate subject matter together and using ICT in the classroom (criterion variables 3, 7, 8, 10 and 12). A mixture of patterns was found for having pupils present to the group and using games in lessons (criterion variables 9 and 11). Fewer successes can be noted for the second project goal, increasing opportunities for differentiation. The teachers succeeded moderately in having pupils work more at their own pace (criterion variable 16). They succeeded however to a small extent only in having them work on subject matter with computers (criterion variable 18). Also in the area of differentiation, some teaching skills turned out hard to acquire or develop, notably having pupils work outside the classroom, having them study additional materials and work on additional assignments and topics of their own choice (criterion variables 13, 14, 15 and 17). Standstill (pattern A) predominated in the development of all of these skills. The third goal of promoting self-directed learning also involved tough challenges. For the skills needed for having pupils plan their own work, present part of the lesson and reflect on the use of the school subject for their futures (criterion variables 19, 20 and 22), the patterns of standstill (A) and failure (B) predominated. The only skill in this area with which the teachers had success (pattern D) involved promoting interest in their school subject (criterion variable 21). Even so, intended or unintended learning occurred also in these areas. The above account refers mainly to the development of teacher action. However, it should be noted that the development of teacher thought did not automatically follow the same patterns.

22 Table 9. Comparison of base line, personal learning goals and learning outcomes Criterion variables N Valid % A. Standstill B. Failure C. Serendipity D. Success I. Increasing variety in classroom activity formats 1. EXPLAINING 18 25 % 33 % 17 % 17 % 33 % 24 33 % 4 % 46 % 0 % 50 % 2. PRECISE 20 28 % 5 % 20 % 35 % 40 % INSTRUCTIONS 22 31 % 0 % 27 % 32 % 41 % 3. HOMEWORK 16 22 % 50 % 13 % 19 % 19 % 22 31 % 18 % 60 % 4 % 18 % 4. INSTRUCTIONAL 19 26 % 16 % 10 % 11 % 63 % DIALOGUE 23 32 % 9 % 43 % 13 % 35 % 5. INDEPENDENT 20 28 % 10 % 5 % 20 % 65 % WORK 23 32 % 0 % 17 % 13 % 70 % 6. GROUP WORK 23 32 % 0 % 22 % 9 % 70 % 44 61 % 9 % 9 % 46 % 36 % 7. PEER COMMENT 23 32 % 9 % 30 % 9 % 52 % 26 36 % 42 % 8 % 35 % 15 % 8. PEER ASSESSMENT 22 31 % 5 % 45 % 9 % 41 % 24 33 % 58 % 4 % 21 % 17 % 9. PUPIL 21 29 % 19 % 29 % 14 % 38 % PRESENTATION 25 35 % 32 % 28 % 16 % 24 % 10. DEBATE SUBJECT 21 29 % 19 % 53 % 14 % 14 % MATTER 25 35 % 56 % 16 % 12 % 16 % 11. GAMES 24 33 % 8 % 33 % 13 % 46 % 25 35 % 32 % 12 % 36 % 20 % 12. ICT 9 13 % 33 % 33 % 11 % 22 % 11 15 % 82 % 9 % 0 % 9 % II. Increasing opportunities for differentiation 13. WORK OUTSIDE 21 29 % 24 % 48 % 9 % 19 % CLASSROOM 25 35 % 52 % 12 % 20 % 16 % 14. ADDITIONAL 23 32 % 13 % 57 % 4 % 26 % MATERIAL 24 33 % 58 % 17 % 25 % 0 % 15. ADDITIONAL 20 28 % 25 % 50 % 20 % 5 % ASSIGNMENTS 24 48 % 46 % 21 % 17 % 16 % 16. OWN PACE 23 32 % 9 % 52 % 4 % 35 % 24 33 % 50 % 25 % 4 % 21 % 17. OWN TOPICS 24 33 % 13 % 58 % 8 % 21 % 25 35 % 60 % 12 % 20 % 8 % 18. COMPUTER WORK 25 35 % 20 % 36 % 12 % 32 % 24 33 % 50 % 12 % 21 % 17 % III. Promoting self-directed learning 19. PUPIL PLANNING 22 31 % 14 % 52 % 19 % 15 % 24 33 % 46 % 21 % 21 % 12 % 20. PUPIL AS CO- 21 29 % 14 % 67 % 5 % 14 % TEACHER 26 36 % 68 % 8 % 24 % 0 % 21. INTEREST IN 21 29 % 15 % 30 % 25 % 30 % SUBJECT 22 31 % 9 % 32 % 9 % 50 % 22. REFLECT ON 21 29 % 29 % 48 % 9 % 14 % FUTURE SUBJECT 25 35 % 36 % 52 % 8 % 4 %

23 The patterns of development described above were also summarized for each project goal by calculating means for the percentages in table 9. The result can be seen in table 10. Again, the white rows concern teacher thought and the shaded rows concern teacher action. Table 10. Patterns of development Project goal A Standstill B Failure C Serendipity D Success 1. Increasing variety mean 17.25 % 25.83 % 15.08 % 41.92 % in classroom activity SD 14.08 % 13.58 % 7.02 % 17.52 % formats mean 28.50 % 23.25 % 19.00 % 29.25 % 2. Increasing opportunities for differentiation 3. Promoting selfdirected learning SD 25.48 % 17.13 % 14.50 % 17.04 % mean 17.33 % 50.17 % 9.50 % 23.00 % SD 6.02 % 7.27 % 5.47 % 9.81 % mean 52.67 % 16.50 % 17.83 % 13.00 % SD 4.85 % 5.06 % 6.62 % 6.98 % mean 18.00 % 49.25 % 14.50 % 18.25 % SD 6.36 % 13.18 % 7.92 % 6.80 % mean 39.75 % 28.25 % 15.50 % 16.50 % SD 21.19 % 16.13 % 7.09 % 19.82 % Table 10 shows that the project goals share some commonality in how the teachers developed. For all three goals, successes in behavior change (pattern D) occurred to a smaller extent than changes in ideas. However, the reverse is true for serendipitous learning (pattern C). This occurred more for action than for thought. When standstill (pattern A) occurred, this was the case more for teacher action than for teacher thought. Failure (pattern B) on the other hand, prevailed more over thought than over action. Now that we know about the learning outcomes of the peer coaching project, we turn to the characteristics of the learning environment, the learning activities that took place within it and their value for professional learning as perceived by the participating teachers. As stated in the introduction, when the project began, the steering committee saw it as an important principle that participation should be voluntary and that teachers should be free to choose with which colleague they wished to cooperate in a pair. At the same time, the committee was uncertain whether pairs should be composed of teachers of the same or similar school subjects. The questions in table 11 were meant to find out how the participants viewed these issues. In addition, we were interested in how they felt about working with a colleague with more or less or the same amount of teaching experience in comparison to themselves. Table 11 shows that the teachers found it quite helpful that they could decide freely with whom they were going to work together. They preferred working with a colleague in the same or a similar school subject. Cooperating with a colleague in a different subject received a lower rating, although not a negative one. Similarly, cooperation with a colleague with the same amount of teaching experience or more experience was preferred. Cooperation with a colleague with less experience was not the favorite option. Finally, we wanted to know how

24 the teachers felt about working with a colleague who pursued clearly different learning goals. On average, they considered this slightly helpful for their own learning. Tabel 11. Characteristics of the learning environment Content of variable FREE CHOICE OF COLLEAGUE SAME SUBJECT SAME DOMAIN DIFFERENT SUBJECT MORE EXPERIENCE AS MUCH EXPERIENCE LESS EXPERIENCE DIFFERENT LEARNING NEEDS Item: For my own learning, I found it very unhelpful [1] < > very helpful [5] that I could choose myself with whom I was going to cooperate in a pair. the colleague in my pair teaches the same subject as I do. [when applicable] the colleague in my pair teaches a subject in the same domain as mine. [when applicable] the colleague in my pair teaches a completely different subject from mine. [when applicable] the colleague in my pair has more teaching experience than I do. [when applicable] the colleague in my pair has about as much teaching experience as I do. [when applicable] the colleague in my pair has less teaching experience than I do. [when applicable] the colleague in my pair and I generally wanted to develop different teaching skills. Valid N Mode Mean SD 37 5 4.19 0.91 17 3; 5 4 1 22 3; 4; 5 3.86 1.04 26 3 2.77 0.71 19 3 3.32 0.89 20 3 3.45 0.89 16 3 2.75 0.58 30 3 3.50 0.94 Concerning the learning activities in the project we were interested in knowing how frequently the intended observations of lessons actually took place. Also, since the project started as a peer coaching project, not especially as a video project, we wanted to know how frequent the use of video actually was. The results for the questions in table 12 show that on average, the participants succeeded in observing colleagues lessons and discussing them within their pair almost twice per month. During the school year, between one and two lessons per teacher were filmed. The teachers viewed the video recordings of their lessons one and a half times on average. Table 12. Learning activities Content of Item: Valid N Mode Mean SD variable NUMBER OF During school weeks I observed on 44 2 1.78 0.96 LESSONS OBSERVED average of my colleague s lessons per month. NUMBER OF This school year of my lessons have 40 2 1.65 1.63

25 OWN LESSONS FILMED NUMBER OF LESSONS DISCUSSED NUMBER OF SELF- VIEWINGS been recorded on video. During school weeks I discussed on average lessons per month with my colleague. I viewed the video recordings of my lesson(s) times. 44 1 1.78 1.11 33 1 1.48 0.91 The participants presence during plenary meetings was generally high, as shown by the percentages in table 13. Table 13. Presence during plenary meetings School year 2004-2005 2005-2006 2006-2007 2007-2008 Average presence during plenary meetings 76% 83% 80% 83% The content analysis of all the available plenary meeting evaluation forms made it possible to produce the following list of conditions that, when fulfilled, the participating teachers considered helpful for their learning. - It is clear from the beginning what results should ensue from the meeting. - The background materials about instructional theory (in a reader and/or an electronic learning environment) should remain succinct. - The meeting is characterized by a good atmosphere and mutual trust in the group. - The meeting has a good pace, i.e. it is neither too slow nor too hurried. - There is opportunity for activity by the partcipants themselves. Not all time should be spent on listening. - Viewing and discussing video recordings of each other s lessons should be alternated with other activities. - Concrete cases, video recordings of each other s lessons and observation points to guide viewing and discussion should be recognizably relevant for the theme or topics on the agenda of the meeting concerned. - Part of the time in plenary meetings should be spent on work within pairs. The content analysis of the available plenary meeting evaluation also resulted in a list of conditions favorable for productive work within pairs. - The school s timetable should offer not only opportunities for visiting, observing and filming each other s lessons, but also for following up these activities within a short time span with a discussion of the lesson at hand. - Post-lesson conferences and discussions take place on the basis of clear observation points. - The work atmosphere between colleagues is open and characterized by mutual trust.

26 - Participants know how to work with digital video hardware. - The sound quality of video recordings is at least adequate. An important characteristic of the peer coaching project studied here was the use of digital video. This is why we focused a number of questions on how the participants experienced the use of video in connection with feedback on their teaching competence. Table 14 makes clear that both viewing video recordings of lessons and receiving and giving feedback on them with the help of video were valued quite positively by the teachers. On average, they valued most the feedback they received from the colleague in their pair on the video recording of their own lessons. However, they found giving feedback to the colleague in their pair almost equally helpful for their learning. Viewing the video recordings of one s own lessons was also experienced as quite helpful. When one compares the first four questions in table 14, it turns out that the viewing of lesson recordings had slightly more influence on the participants reflection on their own teaching than on their changing it. In this respect, self-viewing was slightly more powerful than other-viewing. Such a difference is not in evidence for the influence on changing one s own teaching. Also the feedback from the colleague in one s pair had more influence on the participants reflection on their own teaching than on their changing it.

27 Table 14. Perceived learning value Content of variable USE OF SELF- VIEWING FOR REFLECTION USE OF SELF- VIEWING FOR CHANGE OF TEACHING USE OF OTHER- VIEWING FOR REFLECTION USE OF OTHER- VIEWING FOR CHANGE OF TEACHING INFLUENCE OF FEEDBACK ON REFLECTION INFLUENCE OF FEEDBACK ON CHANGE OF TEACHING USE OF RECEIVING FEEDBACK USE OF GIVING FEEDBACK Item: Valid N Mode Mean SD Viewing my OWN lesson(s) on video was very unhelpful [1] < > very helpful [5] for REFLECTING on my way of teaching. Viewing my OWN lesson(s) on video was very unhelpful [1] < > very helpful [5] for CHANGING my way of teaching. Viewing video recordings of OTHERS lessons was very unhelpful [1] < > very helpful [5] for REFLECTING on my way of teaching. Viewing video recordings of OTHERS lessons was very unhelpful [1] < > very helpful [5] for CHANGING my way of teaching. The feedback from the colleague in my pair and/or from other colleagues on the video recordings of my OWN lesson(s) was very unhelpful [1] < > very helpful [5] for REFLECTING on my way of teaching. The feedback from the colleague in my pair and/or from other colleagues on the video recordings of my OWN lesson(s) was very unhelpful [1] < > very helpful [5] for CHANGING my way of teaching. For my own learning, I found it very unhelpful [1] < > very helpful [5] that I received feedback from the colleague in my pair. For my own learning, I found it very unhelpful [1] < > very helpful [5] that I had to give feedback to the colleague in my pair. 27 5 4.30 0.72 31 4 3.55 1.00 39 4 3.92 0.66 37 4 3.65 0.79 32 4 4.16 0.68 31 4 3.84 0.64 44 5 4.48 0.66 43 4 4.30 0.64 We concluded the analyses for this paper with an exploration of possible relationships between the interval variables in the research model. From a correlation matrix for these variables, those correlations were selected which were significant at the.05 level and surpassed the level of r >.30. The resulting relationships apart from an obvious correlation between age and teaching experience (r =.074) are shown in figure 3. From top to bottom, these results indicate that lessons observed by the teacher pairs were generally also discussed by them (r =.80). Next, filming lessons and viewing the recordings of one s own and colleagues lessons were clearly, though not strongly related to reflecting on and changing one s own teaching (r s ranging between.46 and.70). An interesting finding is that teaching experience was related to the use of other-viewing for changing one s own teaching (r =.50). Apparently, the more experienced teachers in the project were especially willing to benefit from viewing their colleagues work. Finally, the usefulness of giving and receiving feedback were related both with each other and with reflection on one s own teaching (r s ranging between.37 and.73). Interestingly, the usefulness of receiving feedback as experienced by the participants was related to the opportunity they had to choose freely

28 with whom they would cooperate in a pair (r =.34). Together, these relationships confirm the findings in table 14 about the perceived learning value of peer coaching activities using video.

29 Figure 3. Relationships between background variables, characteristics of the learning environment, learning activities and perceived learning value Number of lessons observed.80 Number of lessons discussed Number of lessons filmed.40 Use of self-viewing for change of teaching.46.47 Use of self-viewing for reflection Teaching experience.50 Use of other-viewing for change of teaching.75 Use of other-viewing for reflection Free choice of colleague.34 Use of receiving feedback.73.49 Use of giving feedback.37 Influence of feedback on reflection

30 Conclusions and discussion In this study, learning outcomes were found in four cohorts totalling 45 experienced secondary teachers, who participated in a reciprocal peer coaching project in which digital video was used as a medium for professional development. Three project goals were pursued: increasing variety in classroom activity formats, increasing opportunities for differentiation and promoting self-directed learning. Most results were attained for the first goal, which lay closest to the participants existing teaching practice. As a strategy for professional development, combining reciprocal peer coaching with digital video representations of participants own work in classrooms appeared to have considerable potential for promoting teacher learning. The participating teachers reported having made changes both in their professional action and their professional thought. This is a noteworthy result in the face of gloomy assessments of the possibilities for promoting teacher learning that can be found in some of the professional development literature. At the same time, this study also shows that certain changes in teacher behavior, notably those involved in the second and third project goals, can be hard to achieve. This finding can be attributed to the conditions under which teachers regularly work, especially the fact that most teaching takes place at fixed times in fixed places in fixed groups. Such a context generates cultural routines and role expectations which can only change gradually. This is not impossible, but it takes time. Indications supporting this assessment arose in the third and fourth project years, when the participating teachers were beginning to come to grips with the challenge of increasing opportunities for differentiation by exploring strategies and techniques for using blended learning environments. Not only learning outcomes were explored in this study our first research question but also in what ways the teachers in the project learned the second research question. An important finding is that learning results were achieved intentionally as well as unintentionally. Besides intentional learning leading to success or failure, also unintentional processes involving serendipity or standstill played a considerable part in teacher learning. These different development patterns were seen to occur in different degrees for different teaching skills. In addition, relations between thought and action in teacher development manifested themselves as complex and dynamic. In this study, indications were found that giving as well as receiving feedback using video can promote teacher learning. Also, the evaluations of the plenary meetings and pair work conducted by the participants pointed out the importance of a learning environment in which teachers can interact personally on the basis of mutual trust. This appears to be a necessary condition, if video recordings of one s teaching and collegial feedback are to have a productive impact on teacher learning. The value of the three principles guiding the project volunteering, egalitarianism and ownership and voice were clearly borne out. The combination of reciprocal peer coaching and visualization of teacher work received a particularly high rating from the participants. From a review of 21 empirical studies of teacher learning with digital video we had derived the expectation that this combination is capable of generating changes in teachers action in the classroom. This study yielded indications in the same direction. We think that reciprocal peer coaching with video can be a powerful approach to teacher professional development, because it is a strategy that makes visual representation functional for social meaning-making. Our literature review also indicates, though, that such forms of teacher learning require time to unfold.

31 From the findings reported here we conclude that integrating digital video in reciprocal peer coaching can be a promising contribution towards the professional development of teachers (cf. Allen & LeBlanc, 2005, Martin-Kniep, 2004 and Humpert & Dann, 2000). The research reported in this paper has two limitations which need to be addressed. In the first place, it is based solely on self-report data on a group level. Although self-report data are not per se without value (cf. Ericsson & Simon, 1980), we do intend to supplement this study with analyses of the video data we collected. This should enable us to reconstruct on an individual level how teacher learning unfolded over time. An interesting issue to pursue in this way follows from the indications found for different development patterns (see table 9). These indications as well as observations and conversations with the participating teachers suggest to us that studying video in a collegial context may induce behavior change in particular ways. We surmise that not only can teacher thought precede and direct teacher action which is a conventional notion about how human behavior comes about but just as well the reverse sequence could drive development. More than once during plenary discussions of video recordings of colleagues lessons, teachers exclaimed: Oh, that s a good idea! I ll try that one tomorrow. Soon after the meeting they would just do so, even without fully knowing or understanding exactly why the particular classroom activity they had seen in a colleague s lesson would work for pupils. One implication of the above findings and observations is that spontaneity and playfulness have a valuable role to play in teacher professional development (cf. Tochon, 1999). Another implication is that more attention should be paid to informal processes in teacher learning (cf. Zwart, 2007), both in research and in professional development. To explore such aspects we intend to analyze the video data we collected qualitatively and on a microscopic level. Another idea for future research is to conduct in-depth retrospective interviews with a strategic selection from the teachers involved in this project in order to trace effects of their participation on a longer term.

32 References Ackland, R. (1991). A Review of the Peer Coaching Literature. Journal of Staff Development (12)1, 22-27. Allen, D.W., LeBlanc, A.C. (2005). Collaborative Peer Coaching that Improves Instruction. The 2+2 Performance Appraisal Model. Thousand Oaks: Corwin Press. Brouwer, C.N. (2007). Imaging Teacher Learning. A Literature Review on the Use of Digital Video for Preservice Teacher Education and Teacher Professionalization. Nijmegen: ILS Graduate School of Education [unpublished]. Brouwer, C.N., Korthagen, F.A.J. (2005). Can Teacher Education Make A Difference? American Educational Research Journal 42(1), 153-224. Chan, P.Y.K., Harris, R.C. (2005). Video Ethnography and Teachers Cognitive Activities. In: Brophy, J., Pinnegar, S. (eds.). Learning from Research on Teaching: Perspective, Methodology and Representation. Advances in Research on Teaching Volume 11. Amsterdam: Elsevier JAI. Ericsson, K.A., Simon, H.A. (1980). Verbal Reports as Data. Psychological Review (87)3, 215-251. Clarke, D.J., Hollingsworth, H. (2002). Elaborating a Model of Teacher Professional Growth. Teaching and Teacher Education 18, 947-967. Fortkamp, J.T.J., Berg, E. van den (without year). Leraren in beeld. Literatuurstudie naar de mogelijkheden voor coöperatief leren van beroepsvaardigheden met behulp van digitale video in een hypermedia omgeving [Teachers in view. Literature review about the possibilities for cooperative learning of professional competencies with the help of digital video in a hypermedia environment]. Eindhoven/Enschede: Fontys/Universiteit Twente. Humpert, W., Dann, H.-D. (2001). KTM Kompakt. Basistraining zur Störungsreduktion und Gewaltprävention [Basic Training for Reducing Disruption and Preventing Aggression]. Bern: Verlag Hans Huber. Korthagen, F.A.J. Kessels, J., Lagerwerf, B., & Wubbels, T. (2001). Linking Practice and Theory. The Pedagogy of Realistic Teacher Education. Mahwah, New Jersey: Lawrence Erlbaum Associates. Martin-Kniep, G.O. (2004). Developing Learning Communities through Teacher Expertise. Thousand Oaks: Corwin Press. Scheepens, R. (2004). Rapportage mentor meeloopdagen (MML). Nijmegen: Urban Gymnasium Nijmegen. Tochon, F.V. (1999). Video Study Groups for Education, Professional Development, and Change. Madison, WI: Atwood Publishing. Zwart, R. (2005). Teacher Learning in a Context of Reciprocal Peer Coaching. Nijmegen / The Hague: ILS Graduate School of Education, Radboud University Nijmegen / Netherlands Organisation for Scientific Research (NWO).