Encouraging verbal communication in online small-group Maths problem-solving sessions; taking inspiration from individual sessions

Project leader(s): Abi Kirk
Theme: Supporting students
Faculty: STEM
Status: Archived
Dates: August 2021 to October 2023

In OU STEM online tutorials, speech by students is seen to be rare. Evidence of this is provided by a sequence of studies, starting from the early days of online tutorials. However, students do speak in online individual support sessions (ISS). For this reason, this project looked to ISS for ideas on how to encourage speech in group sessions.

ISS were investigated via a survey of tutors, asking what they felt encouraged speech, and via descriptive logs of ISS compiled by tutors. The free-text responses and logs were analysed thematically to discover features that might encourage speech. A key theme was students’ anxiety about exposing their vulnerabilities. It seemed they might overcome this anxiety under certain circumstances: seeing their needs being ascertained and met, talking about familiar material, and being able to work within a framework provided. Another key theme was that students might speak when they saw benefit to them, including through handling material related to assessment. These features were used to devise a small-group Maths problem-solving session; it also incorporated some ideas from sources on group sessions during Covid-19 lockdowns.

The group session design was implemented with students on occasions a year apart. The sessions ran in Adobe Connect, also using MS OneNote at times. Data was gathered in the form of recording transcripts/commentaries, observations, and student feedback. This data was analysed thematically in order to assess the effectiveness of parts of the design, to suggest modifications from the first year to the second, and to assess the effectiveness of these. The problems for the sessions were provided in advance. There were advantage to this, such as lessening the burden of solving problems live while handling technological demands. There were also disadvantages, such as reducing spontaneity and sometimes allowing better prepared students to dominate discussions. The sessions began with a Quiz to identify gaps in students’ knowledge and discover their preferences on what to cover. This successfully guided the tutor’s problem choices. Improvements that emerged were: using a smaller problem set to make the selection easier, and giving brief feedback on both aspects of the Quiz. After the Quiz came the Icebreaker, where the group fed in errors in an existing solution, so as to discuss Maths without exposing vulnerabilities. This was successful at encouraging participation. It was improved by removing an error which had been hard to spot at source and had led to an unintended discussion which could have exposed students too early. The removal of this shortened the Icebreaker, and seemed to make room for more spontaneous discussion within it.

The main body of the session consisted of problem-solving in three group-work styles (a fourth had been discarded at an early stage). Two were done in breakout with plenary feedback, and one completely in plenary. Style A, in breakout, involved one student writing and another telling them what to write. With all pairs but one, there was rich interaction and collaborative problem-solving. However, in the second year, there were occasions where a student wrote part of their own prepared solution without any prior discussion. Key modifications were clear instructions at the start, and the tutor making suitable teaching points where relevant. The main modification to the Feedback section was to move it from the breakouts to the main room, with two whiteboards side-by-side. This facilitated Kirk, A. (2023) Encouraging verbal communication in online small-group Maths problem-solving sessions. comparison between the work of the two pairs. However, the discussion was less lively than those in the first year, perhaps because the problem that year had involved more choices of method. It is recommended to use a problem with method choices, in the hope that this will discourage writing down without discussion, as sometimes happened the second year.

Style B involved a ‘Consequences’ game, where students continued each other’s partial solutions, in a cycle of breakouts. The first year, this did not work well, because students worked at different speeds due to some having trouble writing, and some forgot to stop. The modifications of time to try out writing and clearer instructions were made, and the next year the style worked as planned. Students were generally able to continue solutions without jumps or repetition. Some did not like this style, feeling alone and unsupported, and disliking being the only one solving a given part. It is recommended to encourage sending messages to the tutor to request support, and especially so that the tutor can make sure no one is the only one beginning a part. On the other hand, one student loved being made to think when continuing a method they would not originally have chosen. The second year, the Feedback section was done in the main room, with several whiteboards side-by-side. This facilitated a rich and student-centred discussion, with students deciding to move the discussion from one board to another, as well as downward. The discussion the first year was less student-centred, with the tutor making almost all decisions on what to discuss, but appeared to involve more mathematical depth and teaching points.

Style C took place in plenary, with the tutor providing structure and pointers, and a student solving each part on-screen. This was successful in that the group solved the problem, with each person building on others’ work. More pointers were needed the first year, and were taken up by the students. The interaction between the students was richer the second year. The first year, it became richer toward the end, whereas in the second year this was reversed. Near the end in the second year, a well-prepared student began to dominate the discussion. It is recommended to avoid this by using a problem where the parts split into related subparts that build on each other, but without great leaps between the parts.