A network approach to teachers’ interactional management in whole-class discussions

ABSTRACT

Dialogic education advocates varied forms of interactional management by teachers. In the context of whole-class discussions, teachers are encouraged to both prompt student involvement through direct mediation, and “step back” to give students a greater sense of agency and responsibility for turn-taking when appropiate. However, it is methodologically challenging to capture the relation between teacher–student and student–student interactions through dialogue in whole-class contexts. One perspective that is especially designed for exploring complex relations in social networks but not as commonly used in dialogic education is social network analysis (SNA). Drawing on 35 whole-class discussion episodes derived from lower secondary classrooms in Norway, this study uses SNA to explore interactional patterns. Methodologically, the study provides novel examples of how SNA can be used both alone and in conjunction with the indicators of interactional productivity to describe the relation between teacher–student and student–student interactions in whole-class discussions. Empirically, a key finding is that an increase in students responding directly to each other in a student-to-student fashion is associated with a reduction in the distribution of productive interactions through dialogue. Theoretically, the study directs attention to the complexities and tensions inherent in key models and concepts in dialogic education.

KEYWORDS:

• Dialogic education
• the role of the teacher
• social network analysis
• centrality measures
• interactional structure
• interactional productivity
• thinking together

Acknowledgments

I would like to thank two anonymous reviewers for their invaluable feedback on an earlier draft of this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. ‘Nodes’, ‘actors’ and ‘vertices’ are sometimes used interchangeably in the SNA literature. Similarly, ‘edges’, ‘lines’ and ‘connections’ are often used interchangeably. However, there may be contexts in which clearer distinctions should be made. For example, it is possible to describe networks where both the artefacts and people have roles. In these cases, it would probably be more meaningful to label the artefact a node and the human an actor.

2. Undirected edges are reciprocal (actor A influences actor B at the same time as actor B influences actor A); directed edges have a one-way effect (actor A influences actor B, but actor B does not influence actor A at the same time). Binary edges ignore any quantitative difference between the existing edges (i.e., they either do or do not exist), while weighted edges can take into account such differences (i.e., the number of times a particular edge is present) (Scott 2017).

3. A whole-class discussion episode started when the teacher invited the students to join a whole-class discussion on a specific topic. The episode ended when there was a change in activity type (e.g., from whole-class to group or individual work) or topic.

4. Episode length varied from 5.38 to 29.58 minutes (M = 14.74, SD = 7.05). Especially regarding distributed productivity, one can expect a natural decline in frequency of new students entering the activity as the activity gets longer. This was confirmed with a correlational analysis. The relationship between episode duration and distributed productivity was statistically significant and negative (r = −.45**). The relationship between episode duration and general productivity was not statistically significant but negative (r = −.28). Therefore, it was necessary to control for the length of episodes when exploring the relationship between SNA measures and the indicators of interactional productivity.

5. There were linear relationships between all the variables, as assessed by scatterplots and partial regression plots. There were three univariate outliers in the data, as assessed by an inspection of a boxplot for values greater than 1.5 box lengths from the edge of the box. None of these were extreme points, and after closer investigation, I decided not to exclude these outliers in the analysis.

Additional information

Funding

This work was funded by the Research Council of Norway [FINNUT/Project No: 254761].

Notes on contributors

Jo Inge Frøytlog

Jo Inge J. Frøytlog is currently associate professor of Educational Science at University of South-Eastern Norway. His work revolves around dialogic education and the use of digital technology for pedagogical purposes.