Mining Sequential Patterns in Classroom Discourse: Insights from

Visualization-Supported Primary Instruction

Fan Chen, Pengjin Wang, Deliang Wang, Wei Jia and Gaowei Chen

Faculty of Education, The University of Hong Kong, Hong Kong, China

Keywords: Classroom Discourse, Academically Productive Talk, Visualization-Supported Dialogic Instruction,

Sequential Sequence Patters, Educational Data Mining.

Abstract: In primary education, effective dialogic strategies employed by teachers play a crucial role in stimulating

student engagement in classroom discussions. Despite this, a gap exists in practice due to teachers’ reliance

on subjective assessment of their questioning strategies, which can impact students’ engagement in classroom

discourse. This study introduces a classroom dialogue analyser designed for primary school teachers to bridge

this gap. The analyser processes classroom videotapes to produce visualization-based reports on dialogue and

student engagement over three months. This facilitates teachers’ self-reflection and refining dialogue

strategies within the 20-student classroom setting. Data mining techniques were utilized to evaluate shifts in

teachers’ questioning strategies, students’ participation in classroom dialogues, and the occurrence of frequent

teacher-student interaction sequences. Results indicate an increase in teachers’ use of talk moves and student

participation in discussion. Furthermore, by combining data on teachers’ and students’ dialogue engagement,

several high-frequency dialogue sequences were identified. Such sequences included instances where students

responded to teachers’ requests to “say more” and expressed their agreement following teachers’ revoicing of

their opinions. Within these sequences, consistently employed talk moves facilitate classroom dialogue

between teachers and students. Identifying these high-frequent dialogue sequences discovered that teachers’

conscious use of talk moves benefits students’ engagement in classroom dialogue.

INTRODUCTION

Developing problem-solving abilities and logical

thinking skills in young students is crucial but

challenging within regular primary classrooms. Many

early childhood students lack the necessary skills to

overcome learning obstacles and maintain focus,

relying heavily on their teachers’ verbal support to

guide their thinking processes (van der Graaf et al.,

2019). Research by Nystrand and Gamoran (1991)

highlights the significance of heuristic questions and

collective discussions in enhancing students’ learning

performance. Therefore, it becomes crucial for

teachers to promote active participation (Mercer &

Littleton, 2007), utilizing various types of questions

to stimulate thinking (van der Wilt et al., 2022).

Teachers need advanced skills to facilitate

interaction and collaboration, enabling meaningful

dialogue and deepening discussions (van der Veen et

al., 2017). Productive classroom dialogue focuses on

core issues and encourages critical thinking about

potential solutions (Resnick et al., 2010), guided by

thought-provoking questions from teachers.

Nevertheless, developing effective questioning

strategies and teacher-student talk patterns can be

challenging for teachers (Khong et al., 2019),

especially considering the demands of managing

tasks in primary classrooms with numerous students

and limited time (Lehesvuori et al., 2019). In regular

classrooms, teachers may rely on familiar

communication strategies without engaging in deep

reflection on their practices (Pehmer et al., 2015),

leading to negative consequences for students’

thinking and working skills and unsatisfactory

learning outcomes.

1.1 Promoting Productive Classroom

Discourse

Researchers have explored various conversation

guides and frameworks to promote productive

classroom discourse and facilitate effective talk

moves between teachers and students. These

Chen, F., Wang, P., Wang, D., Jia, W. and Chen, G.

Mining Sequential Patterns in Classroom Discourse: Insights from Visualization-Supported Primary Instruction.

DOI: 10.5220/0012613300003693

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 16th International Conference on Computer Supported Education (CSEDU 2024) - Volume 2, pages 339-348

ISBN: 978-989-758-697-2; ISSN: 2184-5026

339

frameworks include academically productive talk

(APT) (Michaels & O'Connor, 2015; Resnick et al.,

2010), dialogic teaching (Alexander, 2008), and

exploratory talk (Mercer & Littleton, 2007). These

approaches share the goal of fostering productive

classroom talk to enhance classroom interaction and

academic achievement. Previous studies have shown

a strong relationship between teacher-student

dialogic interaction and students’ authorship,

communication skills, and academic performance.

For instance, Forman et al. (2017) found that teachers’

use of discursive moves supported students’ scientific

argumentation and transformed the teacher’s role from

mentor to partner. Cheng et al. (2022) discovered that

when teachers provided opportunities for students to

express their voices, students assumed the role of

authors, and their authorship was further enhanced

through engaging in rich dialogic discourse. van der

Veen et al. (2017) uncovered the significant effect of

productive classroom talk and metacommunication on

early students’ oral communicative competence.

Additionally, Amodia-Bidakowska et al. (2023)

compared classroom dialogues in different curriculum

contexts and revealed that students’ elaboration of

ideas was associated with improvements in reading,

spelling, punctuation, and grammar skills.

Considering the impact of productive classroom

talk on students’ classroom participation and

academic performance, numerous studies have

explored practical dialogue strategies to assist

teachers in effectively guiding student participation

(Chen et al., 2020; Michaels & O'Connor, 2015).

Academically productive talk, as a structured

conversation approach, aims to equip all classroom

participants with the skills to engage in academically

productive talk. It emphasizes accountability to the

learning community, accepted standards of

reasoning, and knowledge. Strategies for APT enable

teachers to employ accessible and practical discourse

strategies to facilitate open and extended classroom

dialogue (Mercer, 2002; Michaels et al., 2007).

However, an urgent problem remains: how can

we better support primary ICT teachers in employing

more APT and understanding changes in classroom

talk? The significance of addressing this issue lies in

the high dependence of early childhood primary

students on teachers’ guidance, as their thinking and

working skills are developed through engaging in

classroom discourse, ultimately benefiting their long-

term academic performance. Previous research has

explored strategies such as video-based professional

development programs (Hennessy et al., 2018),

visualization-supported instruction reflections (Chen

& Chan, 2022), and instructional intervention

scaffolding supported by analytics technology (Aslan

et al., 2019). There is a consensus regarding the need

to support productive classroom talk, and empirical

evidence shows that professional development

programs and visualization tools aimed at improving

classroom talk moves have a positive impact on

classroom interaction and academic performance

(Chen et al., 2020; Pehmer et al., 2015).

1.2 Sequential Pattern Mining of

Educational Data

Sequential pattern mining is a method used for

temporal analysis and detecting transitions in learning

processes, including the passage of time and the order

in time (Molenaar & Wise, 2022). The passage of

time provides insights into what happened, how long

it occurred, and the sequential order of events. The

order in time examines consecutive events within a

period to understand the learning process (Zhang &

Paquette, 2023). While time series analysis is

commonly employed to understand learners’ online

behavioural patterns, sequential analysis has also

been applied to analyse learners’ dialogue and

classroom discourse sequences.

In the online learning context, researchers have

utilized lag sequential analysis and clustering

techniques to uncover behavioural patterns and group

interactions among online learners (Hou et al., 2010;

Perera et al., 2009). Moreover, sequential analysis has

been used to reveal differences in inquiry learning

processes and the impact of technological support on

group inquiry transition patterns (Lämsä et al., 2020).

Wong et al. (2019) employed a sequential pattern

mining algorithm to explore differences in students’

behavioural patterns with and without regulated

prompts.

While sequence analysis has primarily been

applied in the online field, there are also studies that

have analysed learners’ dialogue sequences. For

instance, Yang et al. (2022) identified meaningful

participation patterns in online discussions from a

temporal dimension, while Ricca et al. (2019)

examined collaborative discourse to uncover

temporal patterns of group dynamics. Some studies

have also explored classroom dialogue from a

temporal perspective, investigating sequential

patterns of classroom discourse across different

subjects (Furtak et al., 2018; Song et al., 2022).

These previous studies demonstrate the value of

sequential analysis in understanding the relationship

between learning events over time, particularly in the

context of classroom dialogue. It not only reveals the

participation of individual teachers and students in

CSEDU 2024 - 16th International Conference on Computer Supported Education

340

classroom discourse but also facilitates the joint

analysis of conversations between teachers and

students, shedding light on the types of questions that

guide talk moves in the classroom (Amodia-

Bidakowska et al., 2023; Song et al., 2022).

This study aims to 1) provide teachers with

visualization-supported reports to support their

reflection on classroom discourse and 2) employs

sequence pattern mining to reveal changes in

classroom dialogue during visualization-supported

instruction. The analysis firstly examines the

questioning strategies employed by teachers and how

students respond during dialogic instruction in

visualization-supported primary classrooms.

Additionally, it investigates the patterns of discourse

moves between teachers and students and examines

how these patterns evolve over time in visualization-

supported primary classrooms. To address these

research questions, the study will analyse the

classroom dialogue participation of teachers and

students separately, as well as compare the joint

classroom dialogue interaction between teachers and

students to reveal changes in classroom dialogue

during three-month intervention period.

METHODS

2.1 Context

The study was conducted in second-grade ICT classes

at a local primary school in Hong Kong, China. The

participating teacher had over ten years of experience

in teaching ICT but had not previously used APT or

visualization-supported tools in teaching. The class

consisted of 20 students, with an average age of 8

years, including an equal number of boys and girls.

Informed consent forms were obtained from the

parents of the participating children and the school.

The classroom instruction spanned a period of

four months and aimed to develop students’ logical

thinking and programming skills through visual

programming using Dash robots. Before the teaching

began, the teacher was introduced to the core

concepts of APT and communicative strategies, with

examples provided for better understanding. The

basics of visualization analytics technology were also

explained to ensure the teacher’s comprehension of

the visualization reports. Throughout the experiment,

the teacher received visualization-based reports from

the classroom dialogue analyser via email. Biweekly

workshops were conducted to assist the teacher in

analysing and understanding the content of the

visualization-based reports, with targeted

explanations on the use of APT to improve classroom

teaching practices.

The classroom instruction data were used to

analyse the sequence of classroom dialogue and

focused on two inquiry topics: Dash for dancing

(topic 1) and Dash for food delivery (topic 2). Dash

for dancing involved students using programming to

enable Dash robots to follow music and perform

simple dance steps, representing their initial attempts

at controlling Dash. Dash for food delivery required

students to use visual programming to control Dash

robots in delivering food and announcing the dish’s

name, representing their second attempt at controlling

Dash. Each topic comprised four videos, and the

teaching structures were similar. Initially, the teacher

explained the content and requirements of the topic to

the whole class, guiding them through progressively

deepening class discussions to formulate visual

programming plans. In the subsequent two classes,

students worked in groups to explore and practice

their programming plans. The teacher allocated time

during the cooperative process for students to report

on their progress and discuss any challenges they

encountered, with the teacher asking questions to

stimulate ideas and guide students’ thinking. In the

final class, students presented their outcomes in

groups, and the teacher guided them in summarizing

and reflecting on their learning process and

achievements. Thus, these two complete classroom

videos can to some extent reflect the changes in

teacher-student interaction throughout the classroom

since the teacher’s involvement in the experiment.

2.2 Classroom Discourse Analyser:

A Visualization-Supported Tool

The Classroom Dialogue Analyzer (CDA), depicted

in Figure 1, was used in this study to support teachers

in reflecting on and improving their classroom

dialogue through visualization-based analysis (Chen

et al., 2015). CDA offers a comprehensive overview

of teaching videos, transcribed text, and visualization

graphs for each class, allowing teachers to easily

understand the frequency of their talk, use of talk

moves, and turn-taking dynamics throughout the

entire class. After completing classroom sessions,

teachers can log into the system to access the

visualization analysis of their classroom dialogue.

This platform has been previously utilized as a

reflective tool in teachers’ professional development

programs, resulting in positive effects on teachers’

talk move usage and students’ academic performance

(Chen, 2020; Chen & Chan, 2022).

Mining Sequential Patterns in Classroom Discourse: Insights from Visualization-Supported Primary Instruction

341

Figure 1: Classroom discourse analyser.

In this study, we generated visualization-based

reports for teachers, corresponding to the platform’s

analysis. These reports, as shown in Figure 2,

provided insights into the classroom dialogue

between teachers and students, as well as students’

overall participation in the class. By comparing these

analytics results, we aimed to identify changes in

teachers’ talk and students’ engagement. The analysis

reports were then emailed to the teachers. Throughout

the experiments, biweekly workshops were

conducted, utilizing visualization-based analysis to

support teachers in reviewing and reflecting on their

classroom practices. This analysis focused on

examining talk move usage, students’ responses, and

the level of student engagement throughout the entire

class. Based on these insights, teachers could enhance

their instructional strategies in future teaching

practices. These ongoing changes and progress were

videotaped in this study.

2.3 Data Collection and Analysis

The data collection for this experiment involved eight

classroom teaching videos, with an average duration

of 32.17 minutes per video. These videos provided

comprehensive coverage of the classroom teaching

situations and encompassed two complete learning

topics conducted over three months. Each topic had

an average duration of 128.69 minutes, allowing for a

thorough exploration of the instructional content and

student engagement within the specified timeframe.

The analysis of the video data proceeded as follows:

First, we transcribed the classroom conversations

using CDA and manually checked the accuracy of the

Figure 2: Visualization-based report.

transcribed text. The teacher’s classroom dialogue

was coded using the APT framework, which is a core

dialogue strategy for facilitating productive

classroom talk, based on the study by Michaels and

O'Connor (2015). For coding student participation in

classroom dialogue, we referred to the study by

Pimentel and McNeill (2013). Each turn in the

dialogue, which consisted of one or more utterances

by an individual, was assigned a code. For example,

if a teacher used a talk strategy to elicit students’

reasoning, it was coded as a turn. A dialogue

sequence represents a chain of exchanges containing

multiple turns, reflecting the topic-focused discussion

between teachers and students (Jin et al., 2016). The

specific coding framework and examples can be

found in the Appendix. The first author coded all

classroom conversations based on turns, while the

fourth author coded 50% of the data. The coding

consistency coefficient, Cohen’s kappa value, was

0.71. In cases of coding inconsistencies, the two

coders engaged in negotiations until a consensus was

reached.

After completing the coding process, we

employed the prefixSpan algorithm, a highly efficient

sequential sequence mining algorithm, to extract

classroom dialogue sequences (Jian et al., 2001). This

CSEDU 2024 - 16th International Conference on Computer Supported Education

342

algorithm is commonly used in educational settings

for behavioural sequence analysis (Bermudez et al.,

2020), personalized recommendation systems (Salehi

et al., 2014), and identifying problem-solving patterns

(Liu & Israel, 2022). We calculated and identified

sequences with a support value greater than 9, along

with their corresponding confidence levels. The

support of a sequence (X → Y) represents the number

of sequences containing items from X followed by

items from Y in the sequence set. For example, if the

sequence {SAM} appears without intervals before the

items from {ELA} for three times, the support of the

sequence SAM → ELA would be three. The

confidence of each sequence (X → Y) reveals the

support of the sequence divided by the number of

sequences containing items from X. It can be

interpreted as the conditional probability P(Y|X). For

instance, if the confidence of the sequence SAM →

ELA is 100%, it means that every time teachers use

the talk strategy “say more,” they always receive a

response from students involving “elaboration.”

Once the analysis was completed, we compared

the frequently used questioning strategies by teachers,

the answering strategies of students, and the dialogue

interactions between teachers and students across the

two topics. These findings are presented in the

findings section of this study.

FINDINGS

3.1

Teachers’ Use

of APT

The analysis focused on examining the use of APT

employed by the teacher during classroom teaching,

as guided by Michaels and O'Connor (2015). The

objective was to compare the teacher’s usage of APT

in the two topics and identify any developmental

changes in their ability to foster productive classroom

discourse with students over three months.

Upon analysing the teacher’s questioning

strategies in the two topics, a notable disparity

emerged: teachers in the second topic employed a

significantly higher number of talk moves compared

to those in the first topic. Specifically, topic 1

comprised 87 question sequence sets, while topic 2

encompassed 103 sequence sets. To gain a more

comprehensive understanding, we conducted a

detailed examination of the teacher’s use of each APT

tool in both topics. This analysis aimed to facilitate a

thorough exploration of the similarities and

differences in the implementation of APT between

the two topics, as depicted in Figure 3.

Figure 3: Teachers use the frequency of APT within two

topics.

Overall, for both topics, the teacher most

frequently employed questioning strategies that

guided students to explain their opinions, using talk

tools like “say more” and “revoice.” Strategies aimed

at guiding students to reason, listen, and engage with

others’ opinions were less frequently used, with slight

differences observed between the two topics. To

further investigate the variations in questioning

strategies between the two topics, a comparative

analysis was conducted.

The analysis revealed that in topic 2, the teacher

used the strategies of “say more,” “press for

reasoning,” “challenge,” “agree/disagree,” and “add

on” more frequently compared to topic 1.

Additionally, a decline in the use of “revoice” and

“restate” was observed in topic 2. It is worth noting

that the questioning tool “explain with others” was

not used in either topic.

3.2 Students’ Opportunities to Talk

Analysing students’ responses to classroom

conversations yields valuable insights into their

participation and the influence of teachers’ use of

APT. Overall, we observed a significant increase in

students’ engagement in class discussions over three

months of visualization-supported dialogic

instruction. In the first topic, we identified 86

sequence sets, while in the second topic, there were

104 sequence sets. This indicates that teachers’

increased utilization of APT positively influences

students’ involvement in classroom discussions.

Furthermore, we conducted a detailed analysis

and comparison of students’ responses during

discussions. The results, presented in Figure 4,

revealed interesting patterns. In both topics, the

majority of students’ responses focused on explaining

their thoughts without providing reasoning.

Mining Sequential Patterns in Classroom Discourse: Insights from Visualization-Supported Primary Instruction

343

Figure 4: Students’ frequency of talking within two topics.

Following that, students often offered reasoning

to support their ideas or expressed agreement or

disagreement with specific statements. A small

number of students’ responses deviated from directly

addressing the teacher’s questions and instead

involved raising their own questions to the teacher

and their peers.

When comparing the students’ responses between

the two topics, several notable differences became

apparent. In the second topic, students tended to

provide elaboration on their opinions without

accompanying reasoning, while there was less

emphasis on expressing agreement or disagreement

with the teacher’s questions. However, students’

responses to elaboration with reasoning and querying

remained consistent across both topics.

3.3 Sequential Patterns of Discourse

Among Teachers and Students

The joint analysis of teachers’ use of APT and

students’ responses provides valuable insights into

the impact of visualization-supported dialogic

instruction on classroom interaction. In this study, we

employed a sequential pattern mining algorithm to

examine the interaction patterns of classroom

dialogue supported by APT, thereby revealing the

effectiveness of this dialogue scaffolding on

classroom interaction.

Our findings indicate that three months of

visualization-based support resulted in more diverse

classroom dialogic patterns. Specifically, in the first

topic, we discovered 1,161 sequence sets, while in the

second topic, the sequence set expanded dramatically

to 29,614. This substantial increase in the second

topic demonstrates that with the support of

visualization, the classroom dialogue interaction

patterns guided by teachers assumes an exceptionally

rich form.

Table 1: Sequential patterns of teacher-student discourse

within topic 1.

Order Discourse sequence

Support Confidence

1 2SAM → 1ELA 23 79.31%

2 1ELA → 2REV 14 87.5%

3 2SAM → 1AGD 12 41.38%

2REV → 1AGD

11 64.71%

1ELA → 2REV →

1AGD

10 87.5%,

71.43%

……

Total 1,161

Notes.

1. The number of each item in the discourse sequence

indicates the speaker role (teacher: 2; student: 1).

2.. The first confidence refers to the confidence level of the

first discourse sequence P(2REV|1ELA), while the second

confidence pertains to the confidence level of the second

discourse sequence P(1AGD|1ELA → 2REV).

To further explore the patterns of interaction, we

present Table 1 and Table 2, which illustrate the

frequent sequential discourse patterns among

teachers and students. In the first topic, the classroom

activity involved collaborative work with a Dash

robot for a dancing task, serving as the students’

initial collaborative assignment. The analysis of

dialogue sequences in Table 1 revealed that direct

communication was the most common sequence, with

teachers prompting students to “say more” and

students elaborating on their opinions. This sequence

had a support of 23 and a confidence of 79.31%,

indicating a high likelihood of students providing

their thoughts without reasoning when prompted to

elaborate. The second most frequent sequence

involved the teacher restating children’s answers after

they expressed their opinions, with a support of 14

and a confidence of 87.5%. Another notable sequence

was when students expressed their agreement or

disagreement after the teacher elaborated on a

judgment and invited further input, with a confidence

level of 41.38%. Additionally, a pattern emerged

where students would express their agreement or

disagreement after the teacher revoiced their

expressions and students had finished elaborating on

their opinions.

CSEDU 2024 - 16th International Conference on Computer Supported Education

344

Table 2: Sequential patterns of teacher-student discourse

within topic 2.

Orde

Discourse sequence Support Confidence

2SAM → 1ELA

29 87.88%

2 2SAM → 1ELA →

1ELA

19 87.88%,

65.52%

2SAM → 1AGD

11 35.48%

4 1ELA →2SAM 11 35.48%

5 1ELA → 2SAM →

1ELA

11 35.48%,

100%

6 2SAM → 1ELA →

2SAM

10 87.88%,

34.48%

7 2SAM → 1ELA →

2SAM → 1ELA

10 87.88%,

34.48%,

100%

8 2SAM → 1ELA →

2REV

10 87.88%,

34.48%

9 2SAM → 2SAM →

1ELA

10 35.48%,

90.91%

1ELA → 2REV

10 30.30%

11 2ADD → 1ELA 10 100%

……

Total

29,614

In the second topic, the classroom interaction of

dialogue displayed even greater diversity, as

presented in Table 2. The analysis identified eleven

high-frequency talk sequences between teachers and

students. The highest-ranked sequence resembled the

one observed in topic 1, where after the teacher

invited students to express their opinions, the students

responded accordingly. This sequence had a support

of 29 and a transition probability of 87.88%. The

second-ranked sequence was closely correlated with

the first one, with one student explaining their point

of view and another spontaneously adding their

perspective. This sequence had a support of 19, and

the probability of the second student adding another

explanation was 65.52%. The third high-frequency

sequence involved the teacher inviting students to

explain their opinions and then eliciting their

agreement or disagreement, with a support of 11 and

a confidence level of 35.48%. Further analysis of the

4th to 7th high-frequency sequences revealed that

when students explained their opinions and the

teacher asked them to continue explaining, there was

a 100% probability of obtaining a clearer explanation.

This indicates that teachers who frequently utilized

the “say more” strategy were more effective in

facilitating talk moves. Additionally, the 8th and 10th

high-frequency sequences highlighted the teachers’

frequent use of the questioning strategy “revoice

students’ opinion.” In topic 2, we also found that

teachers asked students to add others’ opinions.

Notably, when teachers invited students to add

others’ opinions, there was a 100% chance of

receiving a more detailed explanation from students.

Overall, our analysis of the sequential discourse

patterns reveals the impact of APT on classroom

dialogue interaction. The utilization of APT, coupled

with visualization-based support, not only leads to

more diverse conversation patterns but also enhances

the effectiveness of teachers in facilitating productive

dialogue among students.

DISCUSSION AND

CONCLUSION

After analysing the conversation data from eight

second-grade primary school classes, we observed an

increase in the use of APT by teachers, specifically

in inviting students to express their opinions, provide

reasoning, and support their peers’ ideas. However, in

topic 2, we noticed a lower frequency of teachers

inviting students to restate their peers’ answers.

Students showed improvement in expressing their

ideas with more content, although reasoning was still

limited. Furthermore, the need for simple agreement

or disagreement responses decreased in topic 2. These

findings align with previous research indicating that

increased use of talk moves by teachers facilitates

students’ engagement in meaningful dialogue and

critical thinking, thereby enhancing their oral

language competences (van der Veen et al., 2017),

and promoting higher mathematics achievement

(Chen et al., 2020).

Figure 5: The common high frequent sequential patterns of

dialogue within two topics.

Note. In topic 1, confidence is represented by the colour

green, while in topic 2, confidence is represented by the

colour blue.

Mining Sequential Patterns in Classroom Discourse: Insights from Visualization-Supported Primary Instruction

345

Our analysis of frequent dialogue sequences

revealed both shared and unique patterns in the two

topics, as shown in Figure 5 and Figure 6.

Consistently inviting students to elaborate further

resulted in more detailed explanations without

reasoning. This finding was particularly evident in

topic 2, indicating the effectiveness of persistent

questioning strategies in facilitating ongoing student

engagement and expanding classroom discussions

(Orsolini & Pontecorvo, 1992). Revoicing, where

teachers rephrase and present students’ responses,

played a crucial role in facilitating productive

classroom discourse and clarifying students’ thoughts

(Michaels & O'Connor, 2015). After revoicing,

students often responded with agreement or

disagreement, creating a negotiation space among

children (Mercer & Littleton, 2007).

Figure 6: The high-frequent sequential patters of dialogue

in topic 1 and topic 2, respectively.

The study demonstrates that dialogic teaching

with visual support can optimize dialogue strategies

and improve students’ participation in classroom

discourse. Teachers’ increased utilization of talk

moves led to more dialogue interactions and a trend

towards more productive classroom discourse.

However, due to limitations in experiment duration

and available videos, advanced talk strategies were

not fully explored. Future experiments with extended

duration are needed to validate these findings.

In summary, this study highlights the

importance of timing in classroom dialogue and the

potential of dialogic teaching with visual support to

enhance dialogue strategies and student participation.

APT by teachers effectively guided students’

thinking, while strategies for facilitating student self-

expression were also employed. The use of the

revoice strategy was prominent, given the young age

of the participants (Mercer & Littleton, 2007;

Orsolini & Pontecorvo, 1992). As the experiment

progressed, the “add on” strategy emerged as a

frequently employed questioning technique.

Conducting larger-scale and longer-term experiments

will provide more comprehensive insights into

effective talk moves in educational settings and

promote productive classroom discourse.

ACKNOWLEDGEMENT

This work was supported by Hong Kong Research

Grants Council, University Grants Committee (Grant

No.: 17605221) and by the Innovation and

Technology Commission of the Government of the

HKSAR (Grant No.: ITB/FBL/7026/20/P).

REFERENCES

Alexander, R. (2008). Towards dialogic teaching: Rethinking

classroom talk. Dorchester Publishing Company.

Amodia-Bidakowska, A., Hennessy, S., & Warwick, P.

(2023). Disciplinary dialogues: Exploring the association

between classroom dialogue and learning outcomes

within and between subjects in English primary schools.

Learning, Culture and Social Interaction, 43.

https://doi.org/10.1016/j.lcsi.2023.100742

Aslan, S., Alyuz, N., Tanriover, C., Mete, S. E., Okur, E.,

D'Mello, S. K., & Esme, A. A. (2019). Investigating the

Impact of a Real-time, Multimodal Student Engagement

Analytics Technology in Authentic Classrooms

Conference on Human Factors in Computing Systems,

Glasgow, Scotland Uk. https://doi.org/10.1145/3290

605.3300534

Bermudez, R. S., Sison, A. M., & Medina, R. P. (2020).

Extended PrefixSpan for Efficient Sequential Pattern

Mining in a Game-based Learning Environment Asia

Pacific Information Technology Conference, Bali Island,

Indonesia. https://doi.org/10.1145/3379310.3381044

Chen, G. (2020). A visual learning analytics (VLA) approach

to video-based teacher professional development: Impact

on teachers’ beliefs, self-efficacy, and classroom talk

practice. Computers & Education, 144. https://doi.org/

10.1016/j.compedu.2019.103670

Chen, G., & Chan, C. K. K. (2022). Visualization- and

analytics-supported video-based professional develop-

ment for promoting mathematics classroom discourse.

Learning, Culture and Social Interaction, 33.

https://doi.org/10.1016/j.lcsi.2022.100609

Chen, G., Chan, C. K. K., Chan, K. K. H., Clarke, S. N., &

Resnick, L. B. (2020). Efficacy of video-based teacher

professional development for increasing classroom

discourse and student learning. Journal of the Learning

Sciences, 29(4-5), 642-680. https://doi.org/10.1080/10

508406.2020.1783269

Chen, G., Clarke, S., & Resnick, L. (2015). Classroom

discourse analyzer (CDA): A discourse analytic tool for

teachers. Technology, Instruction, Cognition and

Learning, 10(2), 85-105.

Cheng, W. K., Ng, O.-L., & Ni, Y. (2022). Enhancing

Student Authorship and Broadening Personal Latitude in

the Mathematics Classroom with Rich Dialogic

CSEDU 2024 - 16th International Conference on Computer Supported Education

346

Discourse. ECNU Review of Education. https://doi.org/

10.1177/20965311221142887

Forman, E. A., Ramirez-DelToro, V., Brown, L., &

Passmore, C. (2017). Discursive strategies that foster an

epistemic community for argument in a biology

classroom. Learning and Instruction, 48, 32-39.

https://doi.org/10.1016/j.learninstruc.2016.08.005

Furtak, E. M., Bakeman, R., & Buell, J. Y. (2018).

Developing knowledge-in-action with a learning

progression: Sequential analysis of teachers' questions

and responses to student ideas. Teaching and Teacher

Education, 76, 267-282. https://doi.org/10.1016/j.tate

.2018.06.001

Hennessy, S., Dragovic, T., & Warwick, P. (2018). A

research-informed, school-based professional

development workshop programme to promote dialogic

teaching with interactive technologies. Professional

Development in Education, 44(2), 145-168.

https://doi.org/10.1080/19415257.2016.1258653

Hou, H. T., Chang, K. E., & Sung, Y. T. (2010). Applying

lag sequential analysis to detect visual behavioural

patterns of online learning activities. British Journal of

Educational Technology, 41(2). https://doi.org/10.

1111/j.1467-8535.2009.00935.x

Jian, P., Jiawei, H., Mortazavi-Asl, B., Pinto, H., Qiming, C.,

Dayal, U., & Mei-Chun, H. (2001, 2-6 April 2001).

PrefixSpan: mining sequential patterns efficiently by

prefix-projected pattern growth. International

Conference on Data Engineering, Heidelberg, Germany.

Jin, H., Wei, X., Duan, P. R., Guo, Y. Y., & Wang, W. X.

(2016). Promoting cognitive and social aspects of inquiry

through classroom discourse. International Journal of

Science Education, 38(2), 319-343. https://doi.org/

10.1080/09500693.2016.1154998

Khong, T. D. H., Saito, E., & Gillies, R. M. (2019). Key

issues in productive classroom talk and interventions.

Educational Review, 71(3), 334-349. https://doi.org/

10.1080/00131911.2017.1410105

Lämsä, J., Hämäläinen, R., Koskinen, P., Viiri, J., &

Mannonen, J. (2020). The potential of temporal analysis:

Combining log data and lag sequential analysis to

investigate temporal differences between scaffolded and

non-scaffolded group inquiry-based learning processes.

Computers & Education, 143. https://doi.org/10.1016

/j.compedu.2019.103674

Lehesvuori, S., Hähkiöniemi, M., Viiri, J., Nieminen, P.,

Jokiranta, K., & Hiltunen, J. (2019). Teacher

orchestration of classroom interaction in science:

exploring dialogic and authoritative passages in whole-

class discussions. International Journal of Science

Education, 41(17), 2557-2578. https://doi.org/10.1080/0

9500693.2019.1689586

Liu, T., & Israel, M. (2022). Uncovering students’ problem-

solving processes in game-based learning environments.

Computers & Education, 182, 104462. https://doi.org/

10.1016/j.compedu.2022.104462

Mercer, N. (2002). Words and minds: How we use language

to think together (1 ed.). Routledge. https://doi.org/1

0.4324/9780203464984

Mercer, N., & Littleton, K. (2007). Dialogue and the

development of children's thinking: A sociocultural

approach. Routledge. https://doi.org/10.4324/9780

203946657

Michaels, S., & O'Connor, C. (2015). Conceptualizing talk

moves as tools: Professional development approaches for

academically productive discussions. In Socializing

Intelligence Through Academic Talk and Dialogue (pp.

347-361). https://doi.org/10.3102/978-0-935302-43-

1_27

Michaels, S., O’Connor, C., & Resnick, L. B. (2007).

Deliberative Discourse Idealized and Realized:

Accountable Talk in the Classroom and in Civic Life.

Studies in Philosophy and Education, 27(4), 283-297.

https://doi.org/10.1007/s11217-007-9071-1

Molenaar, I., & Wise, A. F. (2022). Temporal Aspects of

Learning Analytics - Grounding Analyses in Concepts of

Time. In The Handbook of Learning Analytics.

https://doi.org/10.18608/hla22.007

Nystrand, M., & Gamoran, A. (1991). Instructional

Discourse, Student Engagement, and Literature

Achievement. Research in the Teaching of English,

25(3), 261-290.

Orsolini, M., & Pontecorvo, C. (1992). Children's Talk in

Classroom Discussions. Cognition and Instruction, 9(2),

113-136. https://doi.org/10.1207/s1532690xci0902_2

Pehmer, A.-K., Gröschner, A., & Seidel, T. (2015). How

teacher professional development regarding classroom

dialogue affects students' higher-order learning.

Teaching and Teacher Education, 47, 108-119.

https://doi.org/10.1016/j.tate.2014.12.007

Perera, D., Kay, J., Koprinska, I., Yacef, K., & Zaïane, O. R.

(2009). Clustering and Sequential Pattern Mining of

Online Collaborative Learning Data. IEEE Transactions

on Knowledge and Data Engineering, 21(6), 759-772.

https://doi.org/10.1109/TKDE.2008.138

Pimentel, D. S., & McNeill, K. L. (2013). Conducting Talk

in Secondary Science Classrooms: Investigating

Instructional Moves and Teachers’ Beliefs. Science

Education, 97(3), 367-394. https://doi.org/10.1

002/sce.21061

Resnick, L. B., Michaels, S., & O'Connor, M. C. (2010). How

(Well-Structured) Talk Builds the Mind. In D. D. Preiss

& R. J. Sternberg (Eds.), Innovations in educational

psychology: Perspectives on learning, teaching, and

human development (pp. 163–194). Springer Publishing

Company.

Ricca, B. P., Bowers, N., & Jordan, M. E. (2019). Seeking

Emergence Through Temporal Analysis of

Collaborative-Group Discourse: A Complex-Systems

Approach. The Journal of Experimental Education,

88(3), 431-447. https://doi.org/10.1080/0022097

3.2019.1628691

Salehi, M., Nakhai Kamalabadi, I., & Ghaznavi Ghoushchi,

M. B. (2014). Personalized recommendation of learning

material using sequential pattern mining and attribute

based collaborative filtering. Education and Information

Technologies, 19(4), 713-735. https://doi.

org/10.1007/s10639-012-9245-5

Mining Sequential Patterns in Classroom Discourse: Insights from Visualization-Supported Primary Instruction

347

Song, Y., Cheng, B., Zhu, J., & Hu, X. Y. (2022). Exploring

the collective process of classroom dialogue using

sequential pattern mining technique. International

Journal of Educational Research, 115, 14.

https://doi.org/10.1016/j.ijer.2022.102050

van der Graaf, J., van de Sande, E., Gijsel, M., & Segers, E.

(2019). A combined approach to strengthen children’s

scientific thinking: direct instruction on scientific

reasoning and training of teacher’s verbal support.

International Journal of Science Education, 41(9), 1119-

1138. https://doi.org/10.1080/09500693.2019.1594442

van der Veen, C., de Mey, L., van Kruistum, C., & van Oers,

B. (2017). The effect of productive classroom talk and

metacommunication on young children’s oral

communicative competence and subject matter

knowledge: An intervention study in early childhood

education. Learning and Instruction, 48, 14-22.

https://doi.org/10.1016/j.learninstruc.2016.06.001

van der Wilt, F., van der Veen, C., & Michaels, S. (2022).

The relation between the questions teachers ask and

children’s language competence. The Journal of

Educational Research, 115(1), 64-74. https://doi.org

/10.1080/00220671.2022.2029806

Wong, J., Khalil, M., Baars, M., de Koning, B. B., & Paas, F.

(2019). Exploring sequences of learner activities in

relation to self-regulated learning in a massive open

online course. Computers & Education, 140.

https://doi.org/10.1016/j.compedu.2019.103595

Yang, B., Tang, H., Hao, L., & Rose, J. R. (2022). Untangling

chaos in discussion forums: A temporal analysis of topic-

relevant forum posts in MOOCs. Computers &

Education, 178. https://doi.org/10.1 016/j.compedu.20

21.104402

Zhang, Y., & Paquette, L. (2023). Sequential pattern mining

in educational data: the application context, potential,

strengths, and limitations. In A. Peña-Ayala (Ed.),

Educational Data Science: Essentials, Approaches, and

Tendencies. Springer.

APPENDIX

Coding scheme for teachers’ questions to the classroom

discourse.

Code