The future of teacher talk: culturally sensitive, impactful & IT smart
Lead Research Organisation:
University of Cambridge
Department Name: Faculty of Education
Abstract
Teaching quality is the most important school factor driving student performance (Chetty et al., 2014). Some classroom interventions (e.g., teacher feedback, direct instruction) have proven influential for student achievement (Hattie, 2009). However, the effects of such interventions tend not to generalise to teacher practice after trials/training end (Riley-Tilliman & Eckert, 2001; Dickinson, 2011). On the other hand, the predictive effects of naturally occurring (without intervention) teacher talk on student outcomes have seldom been studied (see Howe & Abedin, 2012; Mercer & Dawes, 2014; Teo, 2016). This is surprising, given that the study of 'natural teacher talk effects' could inform a new generation of close-to-practice research and teacher interventions with higher chances of generating lasting impacts on teacher practice and student outcomes. To this end, through publications and other impact-generating outputs, this fellowship aims to explore the effects that varied types of teacher talk naturally occurring in classrooms have on student self-regulation and learning dispositions across cultures.
It builds on PhD research that explored the effects of talk used by teachers to regulate students' thinking, encourage self-improvement, understanding, and performance, manage students' sense of competence, and promote collaboration. During the fellowship, the effects that most of these types of teacher talk had for students' self-regulation levels, strategies, and learning dispositions in England and Chile, will be published in two academic articles.
Furthermore, the project will explore the possibility of automating the identification of these types of teacher talk through teaching analytics, a new research technique used to analyse teachers' voices for words, voice waves, and turn taking patterns. Similar automations have been successful in identifying teacher-student dialogic exchanges (Blanchard et al., 2015) and authentic questions (Kelly et al., 2018). It might, therefore, be a promising, efficient alternative to the current slow, challenging and costly human analysis required for identification of classroom teacher talk. If trials prove successful, this project could have a lasting impact on the amount/expense of teacher talk research and teachers' ability to self-monitor their own talk. Results of this methodological enquiry will be published in a third academic article.
The effect of teacher talk across cultures will be disseminated among academics at BERA, AERA, & ORACY conferences. A main 30-40min. fellowship talk will be delivered for practitioners at education festivals and in teacher development courses/workshops. A simple tool kit to help practitioners self-monitor their own teacher talk will be uploaded to the Cambridge T-SEDA website (already agreed). A series of short blogs and podcasts targeted at practitioners will be uploaded to Te@cher Toolkit (UK website with 8.5 million visits - already agreed) and Chartered College of Teaching - CCT - website (already agreed). Two summary blog entries will be published on the LSE Blog aimed at policy makers (100,000 readers a week - already agreed) and BERA blog for academics (800,000 reads in 2018). The main talk of the fellowship will also be tailored and delivered for policy makers/advisors such as DfE, Ofsted, NCTL, EEF, the National Centre for Social Research (NatCen, already agreed).
Furthermore, a new UK-wide network for teacher talk for student learning and development will be initiated. A one-day workshop will be organised to launch the network. UK-based researchers, teachers, and professionals from the key organizations mentioned above will be invited. The work presented at this launch event will be published in a journal Special Issue about teacher talk. Finally, a new three-year research project will be developed throughout the year to gain funds to continue researching 'the future of teacher talk'.
It builds on PhD research that explored the effects of talk used by teachers to regulate students' thinking, encourage self-improvement, understanding, and performance, manage students' sense of competence, and promote collaboration. During the fellowship, the effects that most of these types of teacher talk had for students' self-regulation levels, strategies, and learning dispositions in England and Chile, will be published in two academic articles.
Furthermore, the project will explore the possibility of automating the identification of these types of teacher talk through teaching analytics, a new research technique used to analyse teachers' voices for words, voice waves, and turn taking patterns. Similar automations have been successful in identifying teacher-student dialogic exchanges (Blanchard et al., 2015) and authentic questions (Kelly et al., 2018). It might, therefore, be a promising, efficient alternative to the current slow, challenging and costly human analysis required for identification of classroom teacher talk. If trials prove successful, this project could have a lasting impact on the amount/expense of teacher talk research and teachers' ability to self-monitor their own talk. Results of this methodological enquiry will be published in a third academic article.
The effect of teacher talk across cultures will be disseminated among academics at BERA, AERA, & ORACY conferences. A main 30-40min. fellowship talk will be delivered for practitioners at education festivals and in teacher development courses/workshops. A simple tool kit to help practitioners self-monitor their own teacher talk will be uploaded to the Cambridge T-SEDA website (already agreed). A series of short blogs and podcasts targeted at practitioners will be uploaded to Te@cher Toolkit (UK website with 8.5 million visits - already agreed) and Chartered College of Teaching - CCT - website (already agreed). Two summary blog entries will be published on the LSE Blog aimed at policy makers (100,000 readers a week - already agreed) and BERA blog for academics (800,000 reads in 2018). The main talk of the fellowship will also be tailored and delivered for policy makers/advisors such as DfE, Ofsted, NCTL, EEF, the National Centre for Social Research (NatCen, already agreed).
Furthermore, a new UK-wide network for teacher talk for student learning and development will be initiated. A one-day workshop will be organised to launch the network. UK-based researchers, teachers, and professionals from the key organizations mentioned above will be invited. The work presented at this launch event will be published in a journal Special Issue about teacher talk. Finally, a new three-year research project will be developed throughout the year to gain funds to continue researching 'the future of teacher talk'.
Publications
Torres P
(2021)
A systematic review of physical-digital play technology and developmentally relevant child behaviour
in International Journal of Child-Computer Interaction
| Description | There have been four lines of enquiry carried out so far within the project. Two of these lines have focused in exploring the relationship between, on the one hand, teacher talk relevant for students' learning motivation, and on the other hand, the extent to which students exercise self-regulation behaviours. Among the self-regulation behaviours studied have been students' persistence, help-seeking behaviours, and cognitive strategy use. The type of teacher talk explored have included teacher feedback including teacher talk communicating positive (praise and high expectations) or negative (criticism, disappointment and low expectations) reactions to students' work or ideas, and teacher talk communicating the value of student improvement, understanding or performance in the classroom. All these analyses have been carried out in the frame of a study enquiring about the role of culture in this type of relationship. Here I summarise the findings of two different studies carried out for the project so far: Effects of mastery-performance oriented teaching on self-regulation and achievement across cultures This study explored the predictive effect of mastery oriented teacher talk (valuing students' improvement/understanding) and performance oriented teacher talk (valuing students' good performance) for cognitive strategy use, persistence, and task accuracy of their 8- to 9-year-old students. Eight teachers and 49 students located in eight schools/classrooms from Chile and England participated in the study. To measure teacher talk, functional (socio-cultural) discourse analysis was applied to each teacher utterance found in a sample of 3 hours of Literacy lessons per classroom. Students' self-regulatory behaviours, such as cognitive strategy use and persistence, and students' task accuracy were measured by means of behavioural analysis using challenging tasks. OLS regressions were used to model the predictive effects of mastery and performance oriented teacher talk on students' cognitive strategy use, persistence and task accuracy. Results show universality in the way that mastery and performance oriented teacher talk predicted strategy use, but cultural-specificity in how they predicted persistence and task accuracy. In particular, strategy use was positively associated to performance oriented talk and negatively associated to mastery oriented talk across cultures. Persistence and task accuracy were only predicted by teacher talk in England. Specifically, in England, persistence was positively associated to mastery oriented talk and negatively associated to performance oriented talk. Conversely, task accuracy was positively associated to performance oriented talk and negatively associated with mastery oriented talk. This exploration calls to continue studying the universality and culture-specificity of the effects that mastery and performance oriented teacher talk might have on student self-regulation and task achievement. Relationships between encouraging and discouraging teacher talk, children's persistence and help-seeking behaviours across cultures. This study analysed the effects of encouraging and discouraging teacher talk on 9-year-olds' persistence and help-seeking behaviours in Chile and England. To carry out the study, naturally occurring teacher talk was analysed from 4849 interactive teacher turns video-recorded in 8 classrooms. Persistence and help-seeking behaviours of 49 children were studied through behavioural analysis of their engagements in 353 (out of 635) tasks judged as being sufficiently challenging for each individual child (Mtasks= 7.57 per child, SDtasks= 2.23). Multiple regressions showed that more use of teacher talk judged as 'encouraging' had a positive effect on children's persistence in Chile but a negative effect in England, and no effects on help-seeking across countries. In addition, results showed that when teachers made more use of talk judged as 'discouraging', students tended to show lower levels of persistence in England, and lower use of help-seeking behavior in both Chile and England. Overall, the results highlight the importance of culture in understanding the relationship between teacher encouragement and motivational (persistence) as well as social (help-seeking) facets of self-regulation. A third line of enquiry has been the study of the extent to which the analysed teacher talk could be identified automatically. The explorations applied Natural Language Processing over the teacher talk data set to determine if various different types of teacher talk could be identified. Among the types of teacher talk identified were the following mutually exclusive groups of teacher talk: Group 1 (Regulatory talk) = Directive, guiding, and autonomy supportive teacher talk; Group 2 (Justification of ideas talk): Knowledge claims, Justifications of knowledge claims; Group 3 (Motivational talk) = Mastery oriented talk / Performance oriented talk & Encouraging / Discouraging talk. BERT (loss weighted) Natural Language Processing models using both Multilingual (for English sample) and Spanish (for Chilean sample) versions of BERT models show that the both identification of justification of ideas and regulatory talk can be achieved to some degree using NLP, but that motivational talk cannot be identified through the applied models. Justification of ideas talk can be differentiated from unjustified knowledge claims to some reasonable degree, but it is more accurate to just differentiate phrases containing knowledge claim (justified or not) from other talk not including knowledge claims in general. Furthermore, regulatory talk performs to a reasonable although not reliable degree when trying to identify all three types of teacher regulatory talk together versus other non-regulatory talk. It is not possible to identify each specific type of regulatory talk separately. Motivational talk was impossible to identify using the models. The results indicate that the types of talk that use more predictable markers, such as "because" words to mark justifications might help the identification of types of talk, and that other more flexible or less scripted types of talk, such as motivational talk that might rely much more on tones of voices are difficult to model using BERT models. Regulatory talk, which is teacher talk regulating students ideas and activity, is difficult to differentiate in terms of its directiveness which differentiation might rely more on specific on-going aspects of the activity or ideas at hand and are therefore more difficult to predict. Having said that, there is some degree of reliability in the identification of teacher talk regulating students' ideas and activity, suggesting that this cognitive type of talk might have some type of predictable feature to it. Finally, a literature review of comparative studies looking at both teacher talk and parental talk in relation to the development of cognitive, emotional and behavioural self-regulation of children and adolescents is underway. So far two databases (Web of Science; PsycINFO) have been screened and 14 relevant papers analysed. The features of parenting or teaching found to have culture-specific, at times opposing effects, were adult directiveness, guidance, autonomy support, teacher-student closeness and conflict, and strategy training. The features that were found to have culturally invariant (i.e. universal) effects include parental emotion talk and autonomy support. It is noteworthy that these two latter studies compared Chinese and North American cultures, generally considered very different in the literature. In general, teaching and parenting practices tend to have culture-specific effects on children's development of cognitive self-regulation. Furthermore, the results of studies focusing on emotion regulation show that, depending on culture, both autonomy support and maternal directiveness can promote emotion regulation. Similarly, depending on culture, maternal sensitivity might have a positive or a negative effect on emotion regulation. But there are also some practices that could have culturally invariant effects on emotion regulation, such as the positive effect of physical contact in new-borns, and the negative effects of maternal punishment and requesting compliance without reasoning among 10 year olds. Therefore, the research on emotion regulation shows that in new-borns and toddlers parenting can have culture-specific effects. But by the time children reach 10 years of age same parenting practices could also show culturally invariant effects. Finally, in terms of behavioural autonomous compliance, the studies indicate that more controlling parenting leads to lower compliance and higher hyperactivity in 3 to 10 year-old children. Being a close and responsive parent who uses reasons to negotiate with children, and also a parent that guides children in gentle and playful ways seems to lead to higher compliance among 1 to 10 year-olds in most cultures studied. There are exceptions, however, with Palestine being a case in point. Here, what did have a positive effect on child compliance, at least with regards to prohibitions, was when parents and children did things together jointly as a team, engaging therefore in shared regulation. Overall, the results of this ongoing literature review indicate that most teaching and parenting strategies studied have a culture-specific effect on cognitive and emotional facets of children's self-regulation, but a culturally invariant effect on behavioural compliance and hyperactivity. |
| Exploitation Route | The two studies carried out so far for the project will be disseminated in the form of blogs or podcasts for practitioners (teachers) and also mentioned within a public talk to be produced for the project, and offered to practitioners, about the different ways in which teacher talk can be put at work for the development of self-regulation. The NLP study will help advance the use of machine learning techniques to process teacher talk among academics and enterprises in the future. Furthermore, the literature review can inform practice in families and classrooms alike and raise awareness of the importance of culture in the specific teaching and parenting practices that promote self-regulation, working also as a deterrent of inadequate acritical cultural appropriation of educational practices when it comes to develop autonomy among children and young people. |
| Sectors | Education |
| Description | The results from the teacher-talk project are starting to have a real-world impact in primary classroom in Chile. In particular the different types of teacher talk that were part of this project are now being used as part of a new intervention study promoting students' self-regulated learning taking place in 3 primary free schools in Santiago, Chile (British Academy PF21\210122). Teachers of 7-8 year-olds are being trained in how to use guiding talk, motivational talk promoting mastery orientations, as well as talk promoting student collaboration. The intervention is still underway but classroom observations indicate that the main type o teacher-talk having a visible rapid effect in students is that of talk for collaboration. Talk for collaboration, however, has needed to be further developed within the scope of the new intervention study as it was evident that simply creating the space for collaboration and encouraging collaboration is not enough. Students struggled to know how to talk to one another about their learning and therefore teachers needed to teacher students how to interact. They achieve this by including many more scaffolds for students, such as the provision of particular questions to guide students conversations and requiring students to talk to one another within every lesson (for example, asking students to discuss questions with peers and then asking them to report on what their peer had said). Increasing peer talk has generated much more participation from students. Teacher and student interviews/focus groups carried out so far as part of the intervention study have shown that both teachers and students value using more talk for collaboration when learning. It has had a positive effect on participation, student confidence and understanding. Teachers have also used more and more guidance and motivational talk, but the fact that these types of talk are more implicit part of teacher communication makes it more difficult for both teachers and students to report on the effect these types of talk may be having on learning. Nevertheless, strategies such as "ask three before me", which are in line with a more guiding role from teachers (and talk for collaboration), as well as strategies such as teachers' vocalizations of the value they place in students self-improvement, have been popular across participating teachers. Students also have started to notice the change of value system (more autonomy and self-improvement value) in their own classrooms. Student-level final measures of cognitive effects are still to be collected. These measures will provide a good idea of the real effects that changes in teacher talk and related learning activities could have on children within more teacher-centred education cultures such as the Chilean culture. |
| First Year Of Impact | 2022 |
| Sector | Education |
| Impact Types | Policy & public services |
| Description | British Academy Postdoctoral Fellowship |
| Amount | £398,667 (GBP) |
| Funding ID | PF21\210122 |
| Organisation | The British Academy |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 09/2021 |
| End | 08/2024 |
| Description | University of Cambridge COVID-19 Grant Extension Allocation (CoA) |
| Amount | £38,256 (GBP) |
| Funding ID | G108102 JRAG/341 |
| Organisation | University of Cambridge |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 03/2021 |
| End | 07/2021 |
| Title | Dataset on studies about physical-digital play technology and developmentally relevant child behaviour |
| Description | This database is linked to the following academic article: "A systematic review of physical-digital play technology and developmentally relevant child behaviour". This data set includes the categorisation of 635 studies retrieved from academic data bases ACM Library, ERIC, PsycINFO, specialised academic conferences and journals working on the areas of Play + Technology + Child development. The articles were searched in May 2018 and categorised in terms of : 1. Entries not related to play for its own sake (connected to learning or other aim instead) n=313 2. Entries not connected to developmentally relevant behaviours n=309 3. Entries not analysing analogue behaviour in digitally enhanced contexts n=214 4. Entries not studying typically developing 0-12 year olds n=161 5. Entries that were not empirical studies n=96 6. Entries not specifying sample size and/or participants' age* n=24 7. Entries that were extended abstracts or did not provide enough information about study* n=16 8. Entries that were quantitative studies with 10 or less cases* n=5 9. Entries that were not peer reviewed* n=3 10. Entries that were not published in English n=0 *Criterion coded over full text of a subset of 123 pre-selected entries only A second version has been uploaded adding 2 extra journal articles published by 13th of May 2018 and specifying with more precision missing data related to entries that were not pre-selected for the review and which, therefore, were not coded in relation to criteria 6 to 9 indicated above. |
| Type Of Material | Database/Collection of data |
| Year Produced | 2021 |
| Provided To Others? | Yes |
| Impact | The database has been consulted 221 times and downloaded 10 times by others since its publication to date (March 2022). |
| Description | Collaboration with Natural Language Processing analysts |
| Organisation | Amazon.com |
| Department | Amazon UK |
| Country | United Kingdom |
| Sector | Private |
| PI Contribution | We are currently engaged in analysing teacher talk data using Natural Language Processing techniques with professionals that work at Amazon and PhD student from Computer Science, University of Cambridge. The collaboration is trying to progress in the testing of NLP as a tool to identify different types of teacher talk taking place naturally in the classroom. The collaboration is with individuals rather than institutions. |
| Collaborator Contribution | The collaborators have helped analyse data. |
| Impact | We do not have any outputs from this collaboration yet, but we have put together a formal agreement to conduct the collaboration in order to be able to maintain the confidentiality of the analysed data. |
| Start Year | 2021 |
| Description | Collaboration with Natural Language Processing analysts |
| Organisation | University of Cambridge |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | We are currently engaged in analysing teacher talk data using Natural Language Processing techniques with professionals that work at Amazon and PhD student from Computer Science, University of Cambridge. The collaboration is trying to progress in the testing of NLP as a tool to identify different types of teacher talk taking place naturally in the classroom. The collaboration is with individuals rather than institutions. |
| Collaborator Contribution | The collaborators have helped analyse data. |
| Impact | We do not have any outputs from this collaboration yet, but we have put together a formal agreement to conduct the collaboration in order to be able to maintain the confidentiality of the analysed data. |
| Start Year | 2021 |
| Description | Academic presentation EARLI Metacognition SIG 2022 |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Other audiences |
| Results and Impact | This was an academic conference presentation. Here the details of the title and abstract of the presentation Title: Culturally effective teaching/parenting for self-regulation development across the globe: A review Abstract Does the effectiveness of teaching/parenting practices used to promote self-regulation among children and adolescents depend on culture? New evidence shows that teaching practices might have opposing effects on students' achievement, metacognition and executive functions across cultures. However, there is a lack of comprehensive knowledge about such a cultural educational phenomenon and how this might also apply to parenting. To bridge this knowledge gap, a systematised review methodology was carried out to synthesise cross-cultural studies looking at adult promotion of children's and adolescents' self-regulation. 800 titles and abstracts were screened, 14 relevant studies were identified. Results show that most teaching/parenting practices have a culture-specific effect (varying between positive, negative, and null) on children's cognitive self-regulation and emotion regulation. Specifically, for cognitive self-regulation, teacher/parent directiveness, guidance, autonomy support, emotion talk, strategy training, closeness/conflict with children showed culture-specific effects. For emotion regulation the same culture-specific pattern emerged with regards to parental sensitivity, soothing and co-regulation. Nevertheless, the results also show a culturally invariant positive effect of non-controlling guidance, limit setting, body contact, object stimulation and mutual gaze on autonomous behavioural compliance. The review calls for future research to explore the reasons underlying its findings and warns against inadequate cultural appropriations of teaching/parenting practices. The review was started as part of the present award and continued as part of a British Academy Postdoctoral Fellowship (PF21\210122). The presentation was part of a symposium and the discussant of the symposium highlighted the importance of the presentation to inform further research in the topic of more effective teaching across cultures. |
| Year(s) Of Engagement Activity | 2022 |
| Description | Co-convener of Educational Technology Research Network, at Digital Education Futures Initiative (Hughes Hall College, Cambridge) |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Professional Practitioners |
| Results and Impact | In 2021 I was invited by Prof. Rupert Wegerif to led a new Research Network on Educational Technology at Cambridge. The invitation to take this role started thanks to the role that technology started to take in my own work as a product of the ESRC fellowship. As part of the research network we hosted Distinguished Emeritus Prof. Michael Cole from University of California, San Diego and Prof. Roy Pea, from Stanford University to engage with us in a conversation about their research. Other international authors will continue to be invited to engage in conversations with members of this international cambridge-based network in the following years. |
| Year(s) Of Engagement Activity | 2021 |
| Description | Master Class at Nazarbayev Intellectual Schools (NIS) XXI International Research-to-Practice Conference |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | National |
| Primary Audience | Professional Practitioners |
| Results and Impact | The title of the master class was "Culturally effective teaching for the development of self-regulation: What does the evidence tell us?" The master class, delivered in english was translated simultaneously to Russian and Kazakh. About 40 people, mainly teachers and policymakers from Kazakhstan, attended the talk. The master class was hosted by a local Education NGO 'Institute for Childhood OYNA' led by the previous Director of Research for the Nazarbayev International School, Aizhan Ramazanova. As a product of the presentation I met Dr. Glenda Walsh, Head of Early Years Education at Stranmillis University College, who then I invited to be part of a new EARLI Special Interest Group on Play, Learning and Development I funded together with Dr. Valeska Grau and Dr. Antonia Zachariou in 2021 as one of the coordinators (JURE coordinator). As a new EARLI SIG we would then suggest Dr. Walsh as a Key Note speaker for the next EARLI conference to take place in 2023, so a new line of collaboration opened as a product of the session. |
| Year(s) Of Engagement Activity | 2021 |
| URL | https://conf.nis.edu.kz/speakers/pablo-torres/?lang=en |
| Description | Presentation at PEDAL Centre Research Group, University of Cambridge |
| Form Of Engagement Activity | A formal working group, expert panel or dialogue |
| Part Of Official Scheme? | No |
| Geographic Reach | Local |
| Primary Audience | Postgraduate students |
| Results and Impact | I presented my progress with a systematic literature review on adult-child interaction and child self-regulation across cultures as part of the regular PEDAL Centre's fortnight meetings. The title of the presentation was: "Culturally effective teaching for the development of students' self-regulation: What does the evidence tell us?" This was a shortened version of the Master Class I delivered for NIS International Conference in 2021. About 20 PhD students, research assistants, and postdocs from PEDAL Centre attend these meetings. The presentation led to follow up meetings with various PhD students and postdocs interested in the relevance of cultures in the development of self-regulation within the Centre. |
| Year(s) Of Engagement Activity | 2021 |