Mentoring conversations in teacher education

1305© Unisa Press ISSN 1011-3487 SAJHE 27(5)2013 pp 1305–1323

Mentoring conversations in the professional preparation of teachers

H. H. TillemaDepartment of EducationUniversity of LeidenLeiden, Netherlandse-mail: [email protected]

G. J. van der WesthuizenDepartment of Educational PsychologyUniversity of JohannesburgJohannesburg, South Africa e-mail: [email protected]

AbstractThis article reports on a study that focused on the ways in which the quality of teacher education may be enhanced by mentoring, specifically conversational strategies used by lecturer mentors and the expected and actual impact on student teachers’ learning. The notion of knowledge productivity in mentoring conversation was highlighted to emphasise the importance of mentoring in the professional preparation of teachers. Using a comparative case design, 12 conversations between a student teacher and his/her mentor were video-recorded and analysed with regard to mentors’ conversational moves to help students attain learning goals. This was compared with student teachers’ perceived knowledge productivity as measured in terms of stated intentions to change practices. An instrument was developed to code the mentor’s conversational moves. The findings of the study suggest that: (1) the mentor’s approach during conversation differed, signifying how different strategies relate to the attainment of learning goals; (2) conversational moves did not significantly influence the student teacher’s perceived knowledge productivity. We noted two dominant moves: a scaffolding and prescriptive one, and an exploring one; and (3) student teachers who have a closer relationship involving regular interaction with a mentor, benefited in terms of higher knowledge productivity. Although the findings indicate an overall positive effect of mentors’ conversational moves on student teachers’ learning outcomes, almost 60 per cent of the conversational talk was non-learning goals related, as opposed to relational talk. No direct relation was found between specific mentor conversational moves and perceived knowledge productivity.

Keywords: mentoring, professional preparation, teacher education, learning outcomes, knowledge productivity, conversational moves

INTRODUCTION

Mentoring plays an important part in the professional education of a student teacher.

1306

H. H. Tillema and G. J. van der Westhuizen

This involves the collaboration of a more experienced teacher with a novice teacher who provides ‘systematic and sustained assistance’ to the learner (Huling-Austin 1990). In teacher education, professional learning is supported and facilitated through mentoring (Loughran 2003; Orland and Yinon 2005; Tomlinson, Hobson and Malderez 2010).

The need for research into mentoring processes is being recognised in South Africa, with studies on the development of mentors through reflection (Geber and Nyanjom 2009); how mentors and mentees understand mentoring and their training needs (Mukeredzi, Ndamba and Weda 2009; Schulze 2010); the importance of constructivist approaches in mentoring (Greyling and Du Toit 2008); and the role of mentoring in teacher education (Maphosa, Shumba and Shumba 2007). Van Louw and Waghid (2008, 217) call for critical perspectives on mentoring and the conceptualisations which favour the ‘democratic participation of the mentee’ in South Africa. This study is supportive of this call, and attempts to clarify the dynamics and forms of mentoring conversations.

To a large extent, the student teacher’s professional knowledge is developed and framed within the conversation with a mentor (Edwards 1995). The mentor’s approach during the mentoring conversation therefore may influence the learning outcomes profoundly. In a mentoring conversation a mentor can use different approaches to help the student teacher in his/her learning process (Huling-Austin 1990; Smithey and Evertson 1995). Analysis of mentoring conversations shows that a mentor predominantly determines the format and topics of the conversation, its start and finish (Strong and Baron 2004). In the literature several ingredients for successful mentor conversational approaches have been outlined. According to Daloz (1986), support and challenge are key ingredients. Franke and Dahlgren (1996) point out the benefits of a reflective approach to mentoring. Edwards (2004) stresses the importance of relational and interpersonal skills in conversation. Garvey (2011) acknowledges a focus on meaning making and relevancy of conversation. In their review, Hennissen, Crasborn, Brouwer, Korthagen and Bergen (2008) construct an explicit framework to categorise different approaches (called styles) which mentors may use in conversations. They distinguish between directive and non-directive approaches. A directive approach is characterised as informative, critical, instructive, corrective and advising. Its constituting conversational moves are: assessing, appraising, instructing, confirming, expressing one’s own opinion, offering strategies, and giving feedback. An opposite non-directive approach is defined as reflective, cooperative, guiding and eliciting. The corresponding moves in the non-directive style are: asking questions, guiding to developing alternatives, reacting empathetically, summarising and listening actively.

These conversational moves serve the essential purpose of ‘mentoring’, that is, they ‘systematically and sustainably assist’ the learning and expertise development of the mentee. In line with Ericsson’s (2002) theory on developing expertise, a mentor may accelerate the learning process by giving feedback and knowing what aspects of the performance are ‘ready’ to be improved at a next level of proficiency (Ericsson


1307

2007). Ericsson’s work states that deliberate practice leads to enhanced improvement in performance. A ‘mentored’ deliberate practice in essence builds representations of desired performance goals; knowledge on how to execute the performance; and provides monitoring of performance. This interactive process is depicted in Figure 1.

Figure 1: Model of deliberate practice by Ericsson (2002)

This representation of a deliberate practice can be taken to gauge actual mentoring conversations in order to establish the speech moves the mentor utilises to scaffold and support the learner in the attainment of high(er) levels of proficiency. In our view, the purpose and function of mentoring can be depicted as ‘climbing mount improbable’ (to paraphrase Dawkins 1996) in that the ‘skilled mentor’ (Crasborn and Hennissen 2009) brings the mentee to a level of attainment previously believed to be hard or difficult to reach. This view is brought forward in a slight rearrangement of the model on deliberate practice and shown in Figure 2.

1308


Figure 2: Climbing mount improbable; relating three mental representations

This metaphor stands for the idea that a seemingly complex goal becomes achievable by way of many, gradual, and supportive steps that point out the relevant paths to pursue that which was most often previously unseen. This metaphor may be of help to interpret mentoring conversations as vehicles of deliberate practice.

The purpose of a mentoring conversation is to help bridge the gap between the student teacher’s prior beliefs, unfamiliar theoretical knowledge, and still unattained states of proficiency, and to guide the student through the necessary or requisite knowledge required for action (Edwards 2011). Moves in mentoring conversations can stay at the level of exploring (move 3 in Figure 2), that is, talking about personal tacit beliefs as they relate to the existing knowledge base to be learned as a student, or be a monitoring and supportive move (move 2 in Figure 2) to scaffold learning, that is, starting from the student’s position (in beliefs or performance) and aligning it with a learning goal perspective, or alternatively, (move 1 in Figure 2) deliberately guiding the student toward the to be attained end result, that is, providing directed feedback on relevant knowledge functional to the goal performance. In essence this set of moves resembles an assessment for learning orientation as Sadler (1995) has put forward: (1) knowing where you are; (2) deciding where to go; and (3) specifying the steps to get there.

It is especially the case in teacher education, that the mentors’ position and role is to raise the level of proficiency of their students and conversation is their main vehicle. We were interested in learning how mentors select the conversational moves to ‘climb mount improbable’, that is, to attain learning goals. Is a mentor aware of the risk of guiding the student teacher on a path that is too steep? Or alternatively, select moves to reach a certain level of attainment too early? Or to stay at length at the low road (3) of exploring one’s positions without any new learning occurring? To reach the desired goal performance, that is, the summit of ‘mount improbable’, the mentor


1309

may need to take from time to time a ‘high road’ in conversation, that is, to push forward in the right (goal) direction as is typical for mentoring in the professions (Garvey 2011) as it is for sustaining what Ericsson (2002) calls deliberate practice (Strong and Baron 2004).

Our position is that this framework is helpful in detecting and interpreting mentoring approaches in conversations. For instance, in one scenario a mentor who intends to help the student teacher to ‘monitor his/her performance’ may scaffold and guide the student towards the end goals established by means of persistent reflective questions about the student teacher’s performance compared to the desired goal, is in our view combining moves 1 and 2 (Figure 2) in the conversation. This ‘high road’ approach, or ‘challenging approach’ as typified by Daloz (1986), can be compared with a ‘reflective approach’ mentioned by Franke and Dahlgen (1996) and also be related to the non-directive approach as described by Hennissen et al. (2008), which involves the mentor staying on the ‘low road’ and building acquaintance and comfort by means of moves consisting of discussing and eliciting comments.

LEARNING AS OUTCOME OF CONVERSATION

Mentoring in the professions (Edwards 2011), as is the case in teacher education, is directed toward the attainment of (high) levels of proficiency by mentees. In teacher education, mentoring is aimed at supporting and facilitating the professional development of student teachers. New insights into the professional development of teachers (Tillema, Van der Westhuizen and Van der Merwe, in press) point to the interactional and collaborative nature of teacher knowledge which is built through shared understandings and gradual approximations in performance. In the current study, knowledge attainment for the profession as an outcome of conversation in mentoring was studied from the perspective of knowledge productivity (Tillema and Van der Westhuizen 2006). Knowledge productivity is the creation of conceptual artefacts to improve professional practice (Bereiter 2002). Conceptual artefacts (i.e. tools useful for professional practice) are the outcomes of shared understandings and (often) are collaborative approximations of practice that can be argued about and shared among professionals (Tillema and Orland Barak 2006). These artefacts become productive (i.e. tangible and useful) through conversations and include, for instance, plans, protocols and action schemes (Tillema 2005). Knowledge productivity may then be conceptualised as a measure of such ‘learning’ outcomes (see Bereiter 2002). Challenging (or ‘climbing’) conversations (Farr-Darling 2001) can stimulate knowledge productivity (Baxter Magolda 2004), thereby leading to learning outcomes that evidence themselves in conceptual artefacts. The notion of knowledge productivity was used in the current study to appraise outcomes of conversations, specified the following three evaluative (perceptive) criteria: • Raising problem understanding: this criterion relates to an increased awareness,

better understanding and insights gained as a result of the collaborative exchange, namely, conversation. The most important question under this

1310


criterion is: Is the dialogue related to the practice of the student and does the student acknowledge the issues spoken about as relevant?

• Shifting perspective: this criterion relates to a conceptual change in the beliefs of the student by listening to the viewpoints of the mentor. The most important question of this criterion is: Does the student find the ideas, brought forward during conversation, relevant and useful for practice?

• Commitment to apply: this criterion relates to how the student is involved and interested in the conversations. Social exchange and interaction with the mentor is regarded as important for subsequent follow-up of advice and recommendations made. The most important question here is: Is the student interested in actively following up recommendations? (Tillema 2005).

The central question in the study then was: To what extent do the mentor’s moves in conversation relate to the perceived learning outcomes of the student teacher? More specifically:• To what extent does the mentor’s selection of three different moves during

conversation relate to perceived ‘understanding’, ‘perspective shift’ and ‘commitment to apply’? Conceptually speaking: Is taking a ‘high road’ approach in the mentoring conversation leading to higher perceived learning outcomes?

• As a competing perspective: To what extent do the student’s (prior experience based) expectations with (the mentor approach to) conversations influence the student teacher’s learning outcomes?

• Conceptually: Do established relations in mentoring have an impact on the choice of conversational moves?

METHODS

RespondentsTwelve dyads of student teachers and their mentors participated in the study. Eight student teachers were enrolled in teacher education for secondary education and four attended teacher education for primary education. The students were between 18 and 28 years old and taking courses in their first to their fourth year of education.

Relationships in the dyads varied in closeness, that is, the length or duration of the relationship between a mentor and a student. Four of the 12 mentors were the regular mentors of the student teachers both working together in teaching practice classes. Six mentors were involved as supervising teacher educators. They visited the students at their internship-schools and met regularly for mentoring conversations. Two mentors were working as mentor coordinators. This means they regularly visited, observed, and evaluated student teachers at different sites. The 12 mentors differed in their experience and position as a mentor, with an average of 6.5 years,


1311

and used in the analysis as distinguishing factor.

Design of the studyA comparative case design (Linn 1998) was used in the study to explore the nature of interaction in conversations between mentors and student teachers in different school settings. In a comparative case design it is possible to explore framed contexts in a qualitative and quantitative way (Druckman 2005). The framing in terms of the selection of settings was varied by means of the ‘closeness’ variable, that is, the personal mentoring relationship established between the stakeholders for an extended period of time. The moderating variable was the mentor’s moves in conversation, determined by analysis of propositions in the transcribed mentoring conversation using content analysis methods (Bovar and Kieras 1985). As outcome variables student expectations with conversation as a learning event were measured using a questionnaire as well as in depth interviewing based on the Memorable Event method (Tillema 2005). To determine the learning outcomes of mentoring, the questionnaire on perceived knowledge productivity was used (see Table 1 for an overview of instruments used).

Table 1: Concepts, variables, instruments

Concept Variable Instrument Conjecture

Mentor’s approach

Mentor’s moves

Content analysis coding on prescriptive, scaffolding and exploring propositions by mentor

Prescriptive and scaffolding propositions are related to high road approach and exploring propositions are related to low road approach

Mentoring relationship

Mentoring expectations

Adjusted Ideal Mentoring Scale (IMS)

High expectation is related to positive relationship

Perceived learning impact

Memorable events open interview

High experienced effects are related to positive relationship

Learning outcomes

Knowledge productivity

Questionnaire on perceived knowledge productivity on understanding, perspective shift and commitment

High perceived knowledge productivity is related to high perceived learning outcomes

ProcedureThe selected 12 cases consisted of both a mentor and a student teacher interacting in a mentoring relationship. They were invited by mail to join the study and indicated their willingness to participate. Beforehand they received a short introduction on the nature of the study and its procedure. After both student teacher and mentor gave consent to participate, an appointment was made for videotaping their upcoming mentoring conversation. Before the mentoring conversation, the student was asked to fill out the questionnaire on mentoring expectations. When the regularly scheduled mentoring conversation took place, the researcher visited the site (most often the

1312


internship schools) and gave a short repetition of the procedure and answered possible questions. With the camera installed, the researcher left the room and waited outside during the conversation room so as not to interfere with the process. After the conversation had ended, the researcher administered the questionnaire on perceived knowledge productivity and took the memorable events interview.

Measuring instruments

Student teacher’s mentoring expectationsThe student teacher’s expectation instrument records the way in which a student teacher values a mentoring conversation as contributing to his/her learning. For this purpose a questionnaire was developed based on Rose’s (2000) Ideal Mentoring Scale (IMS). The IMS measures the abilities a student appreciates most in a mentoring relationship. The three sub-scales evaluating the student’s appreciation with the mentor are: integrity, guidance and relationship. Rose’s original questionnaire was adjusted to appraise the student’s current expectations before the conversation with the mentor took place. Therefore, the opening question of the IMS was changed from ‘My ideal mentor would ...’ to ‘What I would like in this conversation with my mentor is ...’. The items of the original IMS were not changed. The adjusted instrument was used to measure the student’s expectations of the mentor.

Before the mentoring conversation the student teacher filled out the questionnaire that consisted of 34 statements on a five-point Likert scale ranging from ‘not true at all’ to ‘very true’.• Integrity consisted of 14 items l (e.g. ‘What I see in my mentor is that he/she

values me as a person’).• Guidance consisted of 10 items (e.g. ‘What I see in my mentor is that he/she

helps me plan a timetable for my research’).• Relationship consisted of 10 items (e.g. ‘What I see in my mentor is that he/she

helps me realise my life vision’).• The internal consistency for these items in three categories was measured by

means of Cronbach’s alpha: for integrity r = 0.87, for guidance r = 0.75 and for relationship r = 0.78.

Memorable events interviewAfter the conversation took place, students received an open interview format with nine evaluative questions pertaining to their satisfaction with the conversation as a learning event. The interview questions asked the student to specify (in writing) the ‘memorable events during conversation as instances of what was said that matters most or was highly relevant to the student on three aspects (with regard to the knowledge productivity of the conversation):• Problem understanding: three questions evaluating whether the student teacher


1313

accepted and learned from the messages expressed in the discussions (e.g. ‘What have you learned and gained from the examples your mentor expressed?’).

• Perspective change: two questions evaluating whether the conversation led to insightful new knowledge (e.g. ‘What the talk you had changed your way of approaching matters in teaching?’).

• Commitment to apply: four items evaluating whether the student teacher took active part in the process (e.g. ‘What kind of consequences would you draw as a result of the mentoring conversation?’).

The student teachers’ answers to each question were coded as positive, negative or neutral. The reliability of this instrument was tested by an inter-rater reliability test. This resulted in an agreement of 88.89 per cent.

Knowledge productivityKnowledge productivity represents the valuation of learning outcomes by the student teacher, that is, did the mentoring support professional practice? This variable is measured with a questionnaire developed by Tillema (2005; Orland Barak andTillema 2006). The questionnaire was administered to the student teacher after the mentoring conversation and consisted of 20 evaluation questions with respect to three categories on a five-point Likert scale (ranging from not true at all to very true).• Problem representation: seven items evaluating whether the student better

understood the topic under discussion and gained insights from the conversation (e.g. ‘I found the problems being discussed authentic and real.’).

• Perspective taking: seven items evaluating the ideas the mentor expressed that contributed to learning (e.g. ‘My thinking changed during the discussion.’).

• Commitment: six items evaluating whether the student teacher was actively involved in the conversation (e.g. ‘I took ideas to practice further.’).

The internal consistency for these items in the three categories was measured by means of Cronbach’s alpha: for problem representation r = 0.71, for perspective taking r = 0.64 and for commitment r = 0.97. To increase the homogeneity of the scale perspective, one item on the scale was deleted (‘I was able to grasp interesting ideas’), thus increasing the value to .71.

Data: Content analysisThe mentor’s moves during conversation were measured with a self-developed coding instrument. The instrument involved a propositional analysis of the transcribed video recording of the conversation. The propositional method in a conversational analysis (Goodwin and Heritage 1995) was chosen to increase rater reliability in scoring the unit of analysis, that is, moves. Moves are speech acts used by the mentor during conversation which, following our conceptual framework, are

1314


categorised as either: (1) prescription: a move containing a reference to the present knowledge base directed toward a performance goal, and included speech acts such as explanation, referencing, guiding, remarking; (2) scaffolding: a move referring to present student performance linking it to a performance goal. Speech acts included : giving hints, providing examples, prompting; or (3) exploration: a move referring to a knowledge base relating it to present student performance, including speech acts such as asking for explication, acknowledgments, invitations. A fourth category contained miscellaneous comments. A guideline was developed for raters to support a reliable scoring (Mazur 2004). Definitions and examples of scoring are:• Prescription: statement in which the mentor tells the student teacher how to act

in a certain situation, how to execute, in order to reach the desired goal (e.g. ‘The best option is sending him/her to his/her seat to reflect.’).

• Scaffold: statement in which the mentor invites the mentee to reflect on classroom behaviour in order to reach the desired goal (e.g. ‘What can you do to prevent this?’).

• Exploration: statement in which the mentor explores the student teacher’s performance in a certain classroom setting (e.g. ‘Were all pupils focused on your instruction?’).

• Other: statement not typically fitting into one of the categories (e.g. ‘I liked your lesson I saw today.’).

The unit of analysis in our conversational analysis is essentially a proposition, that is, a subject–predicate relation (Holsti 1994). In case of unfinished sentences (because of interruptions or pauses) a group of adjacent propositions were used as the unit of analysis. The video recording was transcribed into a meaningful enumeration of units of propositions in order to establish (i.e. score) whether a category had occurred in that particular unit. Only one category was assigned per proposition.

ExampleTo give an example of the coding of mentoring conversations in the study, a part of one mentoring conversation coding is shown step by step:

Step 1: Transcribing the conversationMentor: ‘How could you prevent that for instance? You now say: at the start of the lesson I did not wait for the class to be quiet. You did not check if it was completely clear for the students what your intention was. What your goal for the lesson was, what you expected from the students.’

Step 2: Dividing the conversation in propositions

• How could you prevent that for instance?• You now say: at the start of the lesson I did not wait for the class to be quiet.

gertvdw

Sticky Note

The unit of analysis here is essentially a proposition


1315

• You did not check if it was completely clear to the students what your intention was.

• What your goal for the lesson was, what you expected from the students.

Step 3: Coding the propositionsHow could you prevent that for instance? Scaffolding (question to help the student reflect on

the situation)

You now say: at the start of the lesson I did not wait for the class to be quiet.

Other (citation of the student teacher by the mentor)

You did not check if it was completely clear tothe students what your intention was.

Exploring (exploring the current performance)

What your goal for the lesson was, what you expected from the students.

Exploring (exploring the current performance)

Step 4: Assigning a categoryThe number of specific codes under each category is counted after coding the conversation. The frequency count in each category provides the ‘footprint’ of the conversation. This footprint shows how many propositions in the conversation were prescriptive, scaffolding, exploring or other. In the above example the footprint of this little part of the conversation is: prescriptive: 0, scaffolding: 1, exploring: 2, other: 1.

The reliability of coding by multiple raters was tested. The initial coding agreement on 50 propositions was 46 per cent. Raters then received training; two raters were employed afterwards resulting in an inter-rater reliability of sampled transcripts of k = 0.86

Data inspectionThe scoring of propositions in the content analysis of mentor moves was based on frequencies of three categories to arrive at a ‘footprint’ of each conversation. The categories were: scaffold (n); prescription (n), and exploration (n).

Scores on mentoring expectations were obtained by calculating the mean score on the three questionnaire scales: integrity, guidance and relationship.

Scores on memorable event interviews were obtained by counting the number of positive answers to the nine interview questions. Twelve student teachers answered the scale understanding with a positive instance in 30 of the 36 cases; perspective change was answered positively in 10 of the 24 cases, and on commitment to apply, positive instances were 25 of the 36 answers. In overview, student teachers answered more than half of the questionnaire positively.

The scores on knowledge productivity were obtained by calculating the mean score on the questionnaire scales. The questionnaire consisted of three scales: problem representation, perspective taking and commitment. There were no missing values.

1316


ANALYSIS

To answer the first question on the relation between mentors’ conversational moves and knowledge productivity, the knowledge productivity scale scores were compared for type of ‘footprint’, that is, a combination of categories of conversational moves. We were particularly interested in the effects of a ‘high road approach’ or footprint and a ‘low road’ approach; a high road being dominated by prescription, and/or scaffolding versus a low road being dominated by exploring moves. Taking into account the small number of conversations (n = 12) the Mann-Whitney U-test was used.

To answer the second question on the relation between mentoring expectations and knowledge productivity, two analyses were conducted. Firstly, the scores on knowledge productivity were compared for the high and low expectations students and analysed with a Mann-Whitney U-test. Secondly, the influence of closeness in mentoring relationships on knowledge productivity was analysed. To compare for dyads that were either unfamiliar or familiar in their relationships, the scores were analysed with a Mann-Whitney U-test.

RESULTS

Conversational movesContent analysis of the 12 conversations shows that there was considerable variation in selected moves; grouped in terms of footprints or type of approach, they revealed that three conversations were considered to have a ‘high road’ approach and nine a ‘low road’ approach. Table 2 shows the frequencies for coded categories of all 12 conversations.

Table 2: ‘Footprint’ of all conversations

Conversation Prescriptive Scaffolding Exploring Other High or low road

1 87 64 118 155 High

2 64 8 84 240 Low

3 13 20 38 60 Low

4 13 43 65 122 Low

5 56 19 132 127 Low

6 23 11 11 50 High

7 23 18 89 320 Low

8 10 15 36 112 Low

9 2 5 27 53 Low

10 16 16 39 25 Low

11 47 32 66 54 High

12 27 15 61 46 Low


1317

Mentoring expectationsThe questionnaire on the student teachers’ mentoring expectations contained three scales. The scale integrity had a mean of 4.14 (N = 11, SD = 0.49), the scale guidance had a mean of 3.55 (N = 11, SD = 0.50) and the scale relationship had a mean of 3.27 (N = 11, SD = 0.61). The total mean was 3.71 (N = 11, SD = 0.46). Taking a scale mean of 3.50 to be high on expectations indicated that 7 out of 11 respondents had high expectations.

Knowledge productivityThe knowledge productivity questionnaire contained three scales. The scale problem representation had a mean of 4.35 (N = 12, SD = 0.43), the mean of perspective taking was 3.94 (N = 12, SD = 0.59) and the commitment scale had a mean of 4.23 (N = 11, SD = 0.40). The mean score on all of the scales was 4.16 (N = 12, SD = 0.37).

Conversational moves and knowledge productivityTo answer the first research question, the student teachers’ scores on knowledge productivity were compared under a ‘high road’ approach (n = 3) and mentor used a ‘low road’ approach (n = 9). The median score in the ‘high road’ approach was 3.94 and the median score in the ‘low road’ approach was 4.03.The distributions in the two groups did not differ significantly (Mann-Whitney U-test = 8.00, n = 12, P = 0.31 two-tailed). There was no significant difference in knowledge productivity for students who had a ‘high road’ conversation or a ‘low road’ conversation.

Mentoring expectations and knowledge productivityBased on their expectation score, the student teachers were divided into high and low expectations groups. The knowledge productivity scores were compared for these two groups with a Mann-Whitney U-test. Median score in the high group was 4.37 and median score in the low group was 3.82. The distributions in the two groups differed significantly (Mann-Whitney U-test = 3.00, n = 11, P = 0.04 two-tailed). This result points to student teachers who had high expectations, had higher perceived knowledge productivity.

With respect to closeness in the mentoring relationships, student teachers’ scores on knowledge productivity were compared for a high closeness relationship (n = 6) and low closeness (n = 6). It was expected that students under a high closeness relationship would perceive higher knowledge productivity. For this analysis a Mann-Whitney U-test was executed. The median score in the high closeness group was 4.52 and the median score for low closeness was 3.92. The distributions in the two groups differed significantly (Mann-Whitney U-test = 5.00, n = 12, P = 0.04 two-tailed). This result indicated that student teachers under high closeness, reported higher perceived knowledge productivity.

1318


Both analyses related to mentoring relationship indicated a positive relationship with higher knowledge productivity.

DISCUSSION AND CONCLUSION

The study intended to explore the relation between mentoring conversations and student teachers’ learning, taking into account the students’ relationship with their mentor.

Mentoring relationship and learning outcomesUsing a comparative case design we found support for the influence of student teachers’ mentoring relationship on learning outcomes. The closeness of a mentoring relationship was measured on two outcome variables, namely: student teachers’ expectations, and perceived knowledge productivity of the conversations. When knowledge productivity was compared for student teachers with high and low expectations, our analyses showed a significant difference. Student teachers who are satisfied with their mentors have higher mean perceived knowledge productivity. The same applies when comparing student teachers having a close relationship with their mentors.

Conversational approach and learning outcomesA clear relation between mentor’s moves and student teacher’s learning outcomes was not found. We particularly gauged a ‘high road’ approach versus a ‘low road’ approach by the mentor expecting that prescriptive and scaffolding moves (i.e. a ‘high road’ approach) by the mentor would lead to higher knowledge productivity compared to exploring moves (i.e. a ‘low road’ approach). In fact, the mean knowledge productivity was higher for conversations with a ‘low road’ approach, although no significant differences were found.

In interpreting our findings several reasons can be mentioned why a ‘low road approach’ in mentoring conversation has higher knowledge productivity. A conceptual reason is that prescriptions and scaffolding by the mentor were inadequate, or accepted as stepping stones towards the desired goal. Exploring current performance was considered informative by the students to orient them towards the desired goal. The results in the current study show that exploring current performance had a high frequency of moves as well as miscellaneous moves, indicating that the conversation provides ample time for guiding or prescribing routes, and monitoring performance.

It is also possible that the identified moves are incomplete in giving a full description of the mentor’s intent to use the conversation as a vehicle toward a desired learning outcome. A crucial factor in mentoring that was not included in our selection of moves is the needs of the mentee (Garvey 2011). It can be claimed that student teachers’ learning outcomes are determined by their mentoring needs (Deci and Ryan 2004). In this respect a conversation with low knowledge productivity would not have sufficiently addressed students’ motivational needs. In the study, we


1319

did not cover for students’ different motivational needs, but then again the moves we identified might show a different footprint (a specific combination of three constituting categories). This would constitute an interesting line of study to pursue. In order to satisfy students’ needs, the phase or progression in learning needs to be taken into account (Ormond 2011) since it might have a positive impact on learning outcomes. For example, the needs of a more experienced student teacher may require a different mentoring approach to maximise the learning outcomes.

Another reason for our findings is the sensitivity of our ‘model’, that is, detecting moves in conversations. The instrument we used to measure moves can be improved; not only by training to improve reliability, but also by improving on the content analysis that was used. A propositional analysis converts a conversation as a speech activity into a text, which might lose intent and purpose, as well as interactional cues. The current study favoured a propositional analysis with rigor and control of coding, but may have been at the expense of information and relevancy. In addition, a propositional approach analyses the smallest units possible but in a conversational analysis larger, that is, meaningful units might be a better frame of analysis. In support of this we found that frequent occurrence of sequences of propositions, that is, a scaffolding or a prescriptive proposition is often preceded by several exploring propositions. The coding we used in the study counts the number of propositions in every category rather than their sequence or pattern. It might be of interest to look for patterns, for instance we found that exploring propositions are often introductory for scaffolding or prescription moves.

Another observation with regard to our analysis of moves is the high number of propositions that could not be assigned to one of the three categories based on our model. More than half of the studied conversations had 50 per cent or more ‘other’ propositions. Mena Marcos, Sanchez and Tillema (2010) distinguish between learning oriented moves like prescription or rules and artefacts of scaffolds which were low in frequency of occurrence during mentoring. They found a high number of positive appraisals, namely, comments of reassurance. This might indicate that a considerable amount of time in conversations is needed to cover the ground and provide for an emotional and interactional alliance.

The ‘high road’ moves include giving feedback, providing information and suggesting practical advice, which might only constitute a small (but essential) part of the conversations. Emotional support includes explorative moves which give sympathetic and positive support, attention and empathy.

A mentor’s approach, according to our findings, impacts on students’ setting standards (i.e. expectations) and fruitfulness of discussion (knowledge productivity). Zanting, Verloop and Van Driel (2007) point to the importance of ‘explicating practical knowledge’ in mentoring and argue that (in other words) taking a ‘high road’ approach can be advantageous to student teachers for four reasons, namely: student teachers obtain new information about teaching; they understand the nature of teaching better; they understand their mentor’s mentoring better; and they integrate theory with practice. There are several approaches in conversation to make

1320


knowledge explicit: our study indicates that at least three ‘moves’ are capturing such a conversation. Further research may pursue other possible moves, and need to explore the discursive nature of mentoring interactions (see Tillema, Van der Westhuizen and Van der Merwe 2012; Van der Westhuizen 2012). Further clarity about mentor moves may also be pursued via reflective action research as proposed by Geber and Nyanjom (2006).

ACKNOWLEDGEMENT

We would like to thank Femke Gerretzen (MSc) for her part in the coding and collection of data.

REFERENCES

Baxter Magolda, M. B. 2004. Evolution of a constructivist conceptualization of epistemological reflection. Educational Psychologist 39(1): 31–43.

Bereiter, C. 2002. Education and mind in the knowledge society. Mahwah, NJ: Lawrence Erlbaum.

Bibby, T. 2009. How do children understand themselves as learners? Pedagogy, Culture and Society 17(1): 41–55.

Bornstein, R. F. 1989. Exposure and affect: Overview and meta-analysis of research, 1968–1987. Psychological Bulletin 106(2): 265–289.

Bovair, S. and D. E. Kieras. 1985. A guide to propositional analysis for research on technical prose. In Understanding expository text, ed. B. K. Britton and J. B. Black. Hillsdale, NJ: LEA.

Crasborn, F. J. A. J. 2010. The skilled mentor. Mentor teachers’ use and acquisition of supervisory skills. Dissertation, Technische Universiteit Eindhoven, The Netherlands. Available at: http://alexandria.tue.nl/extra2/675808.pdf

Daloz, L. A. 1986. Effective teaching and mentoring. San Francisco, CA: Jossey-Bass.Dawkins, R. 1996. Climbing mount improbable. New York: W. W. Norton.Druckman, D. 2005. Doing research. Thousand Oaks, CA: Sage.Edwards, A. 1995. Teacher education: partnerships in pedagogy? Teaching and Teacher

Education 11(6): 595–610.Ericsson, K. A. 2002. The path to expert golf performance: Insights from the masters on

how to improve performance by deliberate practice. In Optimising performance in golf, ed. P. R. Thomas, 1–57. Brisbane: Australian Academic Press.

Ericsson, K. A., M. J. Prietula and M. T. Cokely. 2007. The making of an expert. Harvard Business Review 85(7/8): 114–121.

Farr-Darling, L. 2001. When conceptions collide: constructing a community of inquiry for teacher education in British Columbia. Journal of Education for Teaching 27(1): 7–21.

Feldmann, A. 1999. Conversational complexity. Paper presented at the Annual Meeting of the American Educational Research Association, 4–9 April, Montreal.

Franke, A. and L. O. Dahlgren. 1996. Conceptions of mentoring: An empirical study of conceptions of mentoring during the school-based teacher education. Teaching and Teacher Education 12(6): 627–641.


1321

Garvey, B. 2011. A very short fairly interesting and reasonably cheap book about coaching and mentoring. London: Sage.

Garvey, B., P. Stokes and D. Megginson. 2009. Coaching and mentoring, theory and practice. London: Sage.

Geber, H. and J. A. Nyanjom. 2009. Mentor development in higher education in Botswana. How important is reflective practice? South African Journal of Higher Education 23(5): 894–911.

Goodwin, C. and J. Heritage. 1990. Conversational analysis. Annual Review of Anthropology 19(1): 283–307.

Greyling, W. J. and P. H. du Toit. 2008. Pursuing a constructivist approach to mentoring in the higher education sector. South African Journal of Higher Education 22(5): 957–980.

Hargreaves, E. 2010. Knowledge construction and personal relationship: Insights about a UK university mentoring and coaching service. Mentoring & Tutoring: Partnership in Learning 18(2): 107–120.

Hawkey, K. 1998. Mentor pedagogy and student teacher professional development: A study of two mentoring relationships. Teaching and Teacher Education 14(6): 657–670.

Hennissen, P., F. Crasborn, N. Brouwer, F. Korthagen and T. Bergen. 2008. Mapping mentor teacher’s roles in mentoring dialogues. Educational Research Review 3: 168–186.

Holsti, O. R. 1968. Content analysis. In The handbook of social psychology, vol. 2, ed. G. Lindzey and E. Aronson. Reading, MA: Addison-Wesley.

Huberman, M. 1995. Networks that alter teaching, conceptualizations, exchanges and experiments. Teaching & Teachers 1(2): 193–212.

Huling-Austin, L. 1990. Teacher induction programs and internships. In Handbook of research on teacher education, ed. W. R. Houston, M. Haberman and J. Sikula, 39–50. Reston, VA: Association of Teacher Education.

Leone, L., M. Perugini and A. P. Ercolani. 1999. A comparison of three models of attitude-behavior relationships in the studying behavior domain. European Journal of Social Psychology 29(2–3): 161–189.

Loughran, J. 2003. Knowledge construction and learning to teach. Keynote address for the International Association of Teachers and Teaching conference, 26–30 June, Leiden University.

Martin, S. 1996. Support and challenge: Conflicting or complementary aspects of mentoring novice teachers? Teachers and Teaching: Theory and Practice 2: 41–56.

Maphosa, C., J. Shumba and A. Shumba. 2007. Mentorship for students on teaching practice in Zimbabwe: Are student teachers getting a raw deal? South African Journal of Higher Education 21(2): 296–307.

Mazur, J. 2004. Conversation analysis for educational technologists: Theoretical and methodological issues for researching the structures, processes and meaning of on-line talk. In Handbook of research for educational communications and technology, ed. D. Jonassen. New York: MacMillan.

Mukeredzi, T. G., G. T. Ndamba and Z. L. Weda. 2009. Teacher development through distance education: Mentor perceptions of mentoring. South African Journal of Higher Education 23(2): 340–35.

Nespor, J. 1987. The role of beliefs in the practice of teaching. Journal of Curriculum Studies 19(4): 317–328.

1322


Orland-Barak, L. and H. H. Tillema. 2006. Collaborative knowledge construction; the dark side of the moon. Teachers & Teaching 12(1): 1–12.

Orland-Barak, L. and H. Yinon. 2005. Sometimes a novice and sometimes an expert. Mentors’ professional expertise as revealed through their stories of critical incidents. Oxford Review of Education 31(4): 557–578.

Ormond, C. 2011. Tailoring mentoring for new mathematics and science teachers: An exploratory study. Australian Journal of Teacher Education 36(4): 53–72.

Pajares, F. 1992. Teachers’ beliefs and educational research: Cleaning up a messy construct. Review of Educational Research 62(3): 307–332.

Ragins, B. R., J. L. Cotton and J. S. Miller. 2000. Marginal mentoring: The effects of type of mentor, quality of relationship, and program design on work and career attitudes. Academy of Management Journal 43: 1177–1194.

Rodger, F. 2006. The ideal mentor? Professional Learning Today Spring: 29–37.Rolfe, A. 2007. The mentoring conversation. Kincumber, NSW: Synergetic People

Development Pty Limited. Rose, G. L. 2000. What do doctoral students want in a mentor? Development of the ideal

mentor scale. Dissertation Abstracts International 60(12B): 6418. Schulze, S. 2010. Mentees’ views of a structured mentoring programme at Unisa. South

African Journal of Higher Education 24(5): 782–799.Smithey, M. W. and M. W. Evertson. 1995. Tracking the mentoring process: A multi-

method approach. Journal of Personnel Evaluation in Education 9(1): 33–53.Strong, M. and W. Baron. 2004. An analysis of mentoring conversations with beginning

teachers: suggestions and responses. Teaching and Teacher Education 20(1): 47–57.Tillema, H. H. 2005. Collaborative knowledge construction in study teams of professionals.

Human Resource Development International 8(1): 81–99.Tillema, H. H. and G. J. van der Westhuizen. 2006. Knowledge construction in collaborative

enquiry among teachers. Teachers & Teaching 12(1): 51–67.Tillema, H. H., G. J. van der Westhuizen and M. van der Merwe. Under review. Teacher

knowledge building through conversation, submitted to Education as Change.Tomlinson, P. D., A. J. Hobson and A. Malderez. 2010. Mentoring in teacher education.

In International Encyclopedia of Education, ed. P. Peterson, E. Baker and B. McGaw, 749–756. Amsterdam: Elsevier.

Van Louw, T. and Y. Waghid. 2008. A deliberative democratic view of mentorship. South African Journal of Higher Education 22(1): 207–221.

Zajonc, R. B. 1968. Attitudinal effects of mere exposure. Journal of Personality and Social Psychology 9 (Monongraph supplement 2, Part 2).

Zanting, A., N. Verloop, J. D. Vermunt and J. H. van Driel. 1998. Explicating practical knowledge: An extension of mentor teacher’s roles. European Journal of Teacher Education 21(1): 11–28.


1323

App

endi

x

Tabl

e 1:

Sco

res

on a

ll va

riabl

es p

er r

espo

nden

t

Resp

onde

ntM

ento

r’s

appr

oach

Rela

tions

hip

Lear

ning

ou

tcom

es

Con

tent

ana

lysi

sSa

tisfa

ctio

n w

ith

men

tor

Clo

sene

ssK

now

ledg

e pr

oduc

tivity

Pres

crip

tive

Scaf

fold

ing

Expl

orin

gO

ther

Hig

h or

low

ro

adM

Satis

fact

ion

Men

tor’

s so

cial

po

sitio

nC

lose

ness

M

187

(21%

)64

(15%

)11

8 (2

8%)

155

(37%

)H

igh

No

scor

eN

o sc

ore

1H

igh

3.94

264

(16%

)8

(2%

)84

(21%

)24

0 (6

1%)

Low

4.53

Satis

fied

1H

igh

4.67

313

(10%

)20

(15%

)38

(29%

)60

(46%

)Lo

w3.

53Sa

tisfie

d2

Hig

h4.

75

413

(5%

)43

(18%

)65

(27%

)12

2 (5

0%)

Low

3.88

Satis

fied

1H

igh

4.00

556

(17%

)19

(6%

)13

2 (4

0%)

127

(38%

)Lo

w4.

21Sa

tisfie

d2

Hig

h4.

72

623

(24%

)11

(12%

)11

(12%

)50

(53%

)H

igh

3.38

Not

sat

isfie

d3

Low

3.71

723

(5%

)18

(4%

)89

(20%

)32

0 (7

1%)

Low

3.65

Satis

fied

3Lo

w3.

98

810

(6%

)15

(9%

)36

(21%

)11

2 (6

5%)

Low

3.38

Not

sat

isfie

d3

Low

4.13

92

(2%

)5

(6%

)27

(31%

)53

(61%

)Lo

w3.

09N

ot s

atis

fied

3Lo

w3.

86

1016

(17%

)16

(17%

)39

(41%

)25

(26%

)Lo

w4.

09Sa

tisfie

d3

Low

4.

03

1147

(24%

)32

(16%

)66

(33%

)54

(27%

)H

igh

3.94

Satis

fied

1H

igh

4.37

1227

(18%

)15

(10%

)61

(41%

)46

(41%

)Lo

w3.

15N

ot s

atis

fied

3Lo

w3.

78

Mentoring conversations in teacher education

Documents