Promoting Epistemic Growth with Respect to Sustainable Development Issues through Computer-Supported Argumentation

Abstract

Epistemic growth is a desirable outcome of engaging in argumentation related to sustainable development issues. However, earlier studies have rarely been conducted from the perspective of practical epistemology. This longitudinal study aimed to address this gap and to promote epistemic growth in sustainable development issues via computer-supported argumentation through a practice-based approach, using the Apt-AIR framework. The participants were 96 undergraduate students with various majors. Repeated measures of the frequency and epistemic quality of students’ argumentation comments were taken with respect to six consecutive sustainable development issues to explicate the participants’ epistemic growth. The qualitative data of a specific undergraduate provided procedural evidence confirming a change in the epistemic performance and the epistemic growth curves. The results supported an argumentation-based intervention in education with respect to sustainable development issues and highlighted the possibility that the different aspects of epistemic performance are interrelated.

1. Introduction

Recent years have seen a significant growth in educational efforts (namely “epistemic education”) to foster learners’ epistemological development, i.e., helping learners to understand what knowledge is and how it is generated [1]. Epistemic education aims to improve the adaptability of learners encountering the changing world with opportunities, challenges, and risks [2], which is very important for helping learners to understand and adhere to sustainable development in their future lives. By synthesizing philosophical arguments and empirical psychological research, Barzilai and Chinn argued that epistemic education could be applied to promote learners’ epistemic performance to achieve valuable epistemic ends [3], i.e., how knowledge or a decision arises and changes with respect to sustainable development issues (e.g., why the government restricts the development and application of nuclear power), or to promote learners’ epistemic cognition, that is, cognition about the understanding of the nature of knowledge.

Argumentation, which refers to the process of arguing in an organized or logical way, is oriented to epistemic performance and is considered to be a reliable process that leads to epistemic aims, such as knowledge and justified beliefs [4,5]. Sustainable development issues often appear to be controversial issues related to science, both conceptual and/or procedural [6], and are regarded as the main scenes for argumentation in educational standards [7], e.g., argumentation on the necessity to increase the application of nuclear energy at the cost of sacrificing a certain level of safety. Iordanou et al. argued that epistemic cognition supports argumentation by determining whether known strategies have been executed and vice versa [8]. Moreover, Leung evaluated undergraduates’ epistemic performance in their argumentation on sustainable development issues by using the “Apt-AIR” framework [3] to highlight the need to bridge epistemic understanding and argumentative practice [9]. These theoretical explorations and practical results indicate that argumentation can help students to deepen their epistemic understanding of sustainable development.

Unfortunately, learners have difficulties building high-quality arguments and counterarguments on their positions because they have little experience in face-to-face argumentation [10,11]. Computer-supported collaborative argumentation has proved to be an effective approach for learners to learn how to present explicit arguments in order to reach a consensus in socio-scientific argumentation [10,12,13,14]. Studies have proved that formal epistemology, such as epistemological beliefs, and science could grow through computer-supported socio-scientific argumentation [15,16]. However, very few studies have focused on epistemic growth through the lens of practical epistemology, which should be investigated by scholars with respect to epistemic cognition, using process-based approaches [9,17]. Specifically, what and how different aspects of epistemic performance promote epistemic growth with respect to sustainable development issues require exploration [3].

We designed a collaborative online learning environment to carry out a curriculum consisting of six consecutive sustainable development issues for undergraduates from various majors and collected text-based data from online intragroup argumentation comments. Using content analysis and through tracking the follow-up excerpts, we aimed to answer the research questions below.

Research Question 1: What are the frequency and epistemic quality of the undergraduates’ argumentation comments with respect to the six consecutive sustainable development issues in computer-supported argumentation?

Research Question 2: How do different aspects of argumentative performance lead to the epistemic growth of an undergraduate with respect to sustainable development issues?

2. Theoretical Framework

From integrated psychological and philosophical arguments, Chinn et al. developed a descriptive framework named “AIR” to analyze the pattern of individuals’ epistemic cognition and the pattern’s coherence [4,5]. AIR includes three interrelated components: epistemic aims and values (A), epistemic ideals (I), and reliable processes (R). Epistemic aims and values are the goals set by people in their epistemic behaviors and their value assessments, such as how to acquire knowledge, present explanations, and justify beliefs. Epistemic ideals are the criteria and standards that people use to evaluate epistemic products (e.g., knowledge claims), such as whether they contradict existing theories, are consistent with extensive evidence, or provide convincing proof, etc. Reliable processes are the procedures, strategies, and other practices that help in achieving epistemic aims. For example, natural scientists may consider replication as an important basis to prove the scientific nature of research [18]. Among all three components, epistemic aims and values are in the central position to determine which ideals people will adopt and which processes they trust more [4,5,19]. Chinn et al. suggested that learners’ beliefs about the situations in which argumentation is considered a reliable process lead to epistemic aims related to individuals’ positions on sustainable development issues [19].

The Apt-AIR framework was developed to evaluate apt epistemic performance based on The Virtue Epistemology [20,21,22]. Barzilai and Chinn borrowed ideas from the virtues of reliability to define the term “epistemic competencies” for cognitive perspectives that enable the achievement of knowledge, such as perception, memory, and reasoning [3]. With our focus on epistemic growth with respect to the six sustainable development issues, the current study drew upon the Apt-AIR framework to analyze the argumentative comments from online discussions to evaluate students’ epistemic performance for characterizing their meaningful engagement in computer-supported collaborative argumentation.

Barzilai and Chinn concluded that there are five key aspects of epistemic performance related to epistemic aims, ideals, and reliable processes that are essential for individuals to achieve their epistemic ends [3]: engaging in reliable cognitive processes, adapting epistemic performance to varied contexts, regulating and understanding epistemic performance through epistemic metacognition [23], enjoying the epistemic performance, and collaboratively participating in epistemic performance. In this study, undergraduates in a group aimed to generate solutions (A) through diversified reliable processes (R) (e.g., investigating statistical data, tracking testimonies from multiple sources) for each authentic sustainability problem (e.g., how vaccines work against viruses) embedded in an issue of sustainability referring to the ideals (I) of multiple perspectives (e.g., scientific feasibility, ethical considerations of cloning technology). We developed a situated coding scheme based on Apt-AIR derived from the conceptualization of Barzilai and Chinn to evaluate epistemic growth through (1) the frequency of comments related to the five aspects and (2) the epistemic quality.

3. Methods

3.1. Research Design

We conducted a longitudinal study with embedded mixed methods [24]. We first evaluated the undergraduates’ epistemic growth by examining the frequency and the epistemic quality of their argumentative comments using content analysis. A specific undergraduate’s comments were tracked to interpret epistemic growth in the follow-up qualitative analysis.

3.2. Participants and Procedures

We carried out this study in an 18-week elective curriculum called “Science and Society” for 96 undergraduates from different majors of a local college, including 37 boys and 59 girls. Among the participants, 39 majored in natural science (e.g., biology, physics), 28 in social science (e.g., sociology, economics), 15 in engineering (e.g., chemical industry), and 14 in other disciplines (e.g., mathematics). The participants were assigned into 16 groups of 6 to achieve heterogeneity in terms of major and gender. Before participating in the study, a survey indicated that none of the participants had relevant experience in online argumentation. In addition, the participants from different majors in this intervention did not know each other beforehand, which, to some extent, avoided communication after classes. All participants signed a consent form and volunteered to participate in the study. The participants received two credits for participating in this long-term intervention.

Participants adopted blended collaborative learning in groups covering six consecutive sustainable development issues (e.g., genetically modified organisms and climate change). Each issue lasted three weeks.

Table 1. Exemplar of the learning procedure for the topic of GMOs (italics represent the source of data collection).

Schedule	Offline Tasks, In-Class	Online Tasks, In-Class	Online Tasks, Off-Class
Week 1	Listen to introductory lectures on GMOs, including scientific/social perspectives (30 min)	Answer three questions raised in the learning task “Can we continue developing GM technologies” (60 min)	Upload the answers to the questions and evaluate the answers of other groups (by the end of Week 1)
Week 2	Read and discuss the answers, evaluations, and follow-up responses of two groups selected by the teacher (40 min)	Discuss the defects and faultiness in previous answers. Search public websites for the necessary information to update the answers (50 min)	Upload the updated answers and re-evaluate the answers of other groups (by the end of Week 2)
Week 3	Publicly display the updated answers of the groups (90 min)	Record the follow-up questions about the solutions of other groups (keeping pace with offline speech)	Score the final versions of other groups according to the predetermined criteria (by the end of Week 3)

exhibits an exemplar of the online and offline learning procedure of the fourth issue, “genetically modified organisms (GMOs)”, in detail.

3.3. Learning Environment

The collaborative learning environment was developed using a private studio and embedded in the private platform of the local college. Figure 1 shows a scripted individual interface which was designed to assist the undergraduates’ construction of arguments and counterarguments (for more about collaborative scripts, see Kollar et al. [25]). During collaborative argumentation in online tasks, the undergraduates could choose the “structured mode” to be scripted (or “free mode” without scripts) from the drop-down menu of the different components of arguments (e.g., warrants, rebuttals), attachments (uploaded by undergraduates), and sources (the specific group member) and thus construct high-quality arguments based on Toulmin’s Argumentation Pattern [26,27]. The participants traced arguments via: (1) the information window for the full text of a specific comment, (2) argument maps of comments organized by attachments, (3) a global window for all comments organized by timeline (which could be displayed for the current day, this week, or 3 weeks for each member or the whole group). Each group established and refined their answers based on the intergroup discussion and presented answers to the learning task through the “Class Interface” button. Considering that the intergroup learning procedures were not the focus of the study, we have not shown these in Figure 1.

3.4. Data Collection

The intragroup argumentation yielded 4233 comments collected only from in-class online tasks in the first two weeks of each sustainable development issue, as other comments were created through intergroup argumentation. All the comments were exported in chronological order and coded to form a data series in a text document for content analysis to answer Research Question 1. One undergraduate student was selected for semantic analysis because of their representative epistemic growth curve to answer Research Question 2.

3.5. Coding Scheme and Data Analysis

As social epistemologists have suggested [28], we first analyzed the potential value of the aims, ideals, and reliable processes in the argumentation on sustainable development issues. Secondly, we formulated a coding scheme, i.e., the instrument for semantic analysis, contextually adapted from the Apt-AIR [3] by considering the expected epistemic performance in the computer-supported learning environment. To illustrate this,

Table 2. The coding scheme: operational definitions and example (c.f. Barzilai and Chinn, 2018 [3]).

Aspects	Operational Definition	Example of Coding
Aspect 1: Cognitive engagement in epistemic performance	Students want to answer task-based questions (A). They not only examine the internal structure of testimony but also evaluate whether the information is relevant to prior knowledge or empirical evidence (I). They examine and evaluate the content and the logical structure of the comments (R).	“I think you did not make a convincing argument, because you provided unproved evidence about how GM corns generates.”
Aspect 2: Adapting epistemic performance	Students know that some questions require multiple perspectives (A). They appropriately select criteria from multiple perspectives which match the issue-based context (I). They collect and evaluate information to generate or evaluate comments to fit the situational conditions (R).	“Although we have to reduce carbon emission, we also have to calculate how much we should pay for it.”
Aspect 3: Regulating and understanding epistemic performance	Students understand the meaning of argumentation when answering questions (A). They realize the importance of formulating and referring to the criteria (I). Students know why, when, and how to apply these criteria (R). They monitor the processes of generating and evaluating comments critically and reflectively (R).	“We cannot upload this file, as the author is not an expert in biology that can be trusted in GM related issues.”
Aspect 4: Caring about and enjoying epistemic performance	Students have an evaluative or skeptical stance on comments (A). They value ideals and enjoy the epistemic ends when they fit the ideals (I). Students are self-motivated by intellectual virtues (e.g., curiosity, open-mindedness) to engage in the generation and evaluation of comments even though these processes involve time and effort (R).	“Great! I have found a story relevant to climate change. I think this is going to be an important argument for the causes of climate change.”
Aspect 5: Participating in epistemic performance together with others	Students answer questions collaboratively through comments (A). They collaboratively formulate and evaluate ideals through argumentation (I). They appreciate the processes of collaborative knowledge construction (R).	“We can integrate our searching results for relevant papers written by scientists to solve the second problem about nuclear power.”
Off-task	Students make comments unrelated to the issue.	“I am tired now.”

includes the operational definitions and examples of coding for the five aspects applied to the analysis of epistemic performance regarding sustainable development issues.

Each comment from the intragroup discussion was coded semantically for each aspect of the scheme. That is, if multiple aspects were involved in a comment, all of them were coded for the frequency statistics. Two coders who were familiar with the curriculum and the coding scheme coded each comment. The inter-coder reliability was calculated, and the value of kappa was 0.79, which represented good reliability [29]. Further, as Barzilai and Chinn assumed, the five aspects were not independent factors but instead interwoven with one another and functioned conjointly [3]. We summed the number of aspects involved in each comment to evaluate the epistemic quality. In other words, we believed that a comment with high quality covered more aspects (from 0 to 5). The average quality score of each undergraduate for the six sustainable development issues was calculated. One-way repeated measures ANOVA was used to investigate the epistemic growth.

4. Results

4.1. Characteristics of Epistemic Growth Regarding Sustainable Development Issues

Figure 2 shows the descriptive statistics regarding the performance of argumentation regarding the six issues of sustainable development issues in collaborative learning. From a global view, the results revealed that the undergraduates improved significantly regarding the two dependent variables: the frequency of comments from each aspect of the coding scheme and the epistemic quality of the comments. Regarding the frequency of comments, the curves in Figure 2 (left) show highly significant growth in all five aspects of epistemic performance. There was also a curve indicating the highly significant growth in epistemic quality. However, the non-monotonicity of the curve implied that the frequency and the quality of the comments did not rise steadily but fluctuates (see the decreasing lines in Figure 2). These results also implied that the undergraduates needed more time to experience the processes of argumentation about sustainable development issues to realize significant improvements (note the most significant growth of each variable from Issue 5 to Issue 6).

In light of the results in Figure 2, we investigated the difference in the undergraduates’ growth regarding the variables. A one-factor repeated measure ANOVA with the order of the issue as a repeated factor was conducted to investigate the epistemic growth regarding the two dependent variables discussed above. The results (

Table 3. One-factor repeated measure ANOVA results for the frequency and the epistemic quality of comments.

Source	Issue 1	Issue 2	Issue 3	Issue 4	Issue 5	Issue 6	F Value (p)	Pairwise Comparisons
							Partial η2
Aspect 1	1.95 (0.151)	2.46 (0.155)	1.89 (0.136)	2.96 (0.205)	3.19 (0.202)	4.64 (0.233)	33.413 *** 0.260	6 > 5, 4, 2, 1, 3 ***; 5 > 1, 3 ***; 4 > 1 **; 4 > 3 ***
Aspect 2	1.13 (0.102)	0.74 (0.095)	2.13 (0.147)	2.92 (0.176)	1.62 (0.169)	3.33 (0.226)	45.045 *** 0.322	6 > 3 ***; 6, 4 > 5, 1, 2 ***; 4 > 3 **; 3 > 1, 2 ***; 5 > 2 ***
Aspect 3	1.20 (0.118)	1.69 (0.121)	2.09 (0.126)	2.46 (0.163)	2.30 (0.174)	3.60 (0.220)	30.624 *** 0.244	6 > 4, 5, 3, 2, 1 ***; 4 > 2, 1 ***; 5, 3 > 1 ***; 5 > 2 *
Aspect 4	0.51 (0.074)	0.89 (0.087)	1.37 (0.114)	1.19 (0.101)	1.39 (0.118)	2.08 (0.165)	21.868 *** 0.187	6 > 5, 3 *; 6 > 4, 2, 1 ***; 5 > 2 *; 5, 4, 3 > 1 ***; 3 > 2 **; 2 > 1 *
Aspect 5	1.24 (0.129)	1.63 (0.109)	3.24 (0.204)	3.17 (0.203)	3.20 (0.225)	4.00 (0.255)	32.014 *** 0.252	6 > 5 *; 6, 3, 5, 4 > 2, 1 ***
Epistemic quality	1.00 (0.043)	1.18 (0.045)	1.39 (0.041)	1.41 (0.042)	1.30 (0.045)	1.73 (0.045)	59.541 *** 0.385	6 > 4, 3, 5, 2, 1 ***; 4 > 2, 1 ***; 3 > 2 **; 3, 5 > 1 ***

Note: * p < 0.05, ** p < 0.01, *** p < 0.001.

) showed that the undergraduates improved significantly in terms of the two variables (F-value and p-value) with strong explanatory power (partial η² > 0.14).

Follow-up pairwise comparisons (Bonferroni) indicated that the degree of their epistemic growth varied. Firstly, cognitive engagement (Aspect 1) and collaborative participation (Aspect 5) had the first two highest frequencies in each issue, but emotional participation (Aspect 4) had the lowest, as shown in the descriptive data for Issues 5 and 6 in Figure 2. It was found that although students participated in collaboration, they mainly focused on completing the task themselves and did not fully enjoy the epistemic processes. Secondly, we noticed that adaptability (Aspect 2) and emotional participation (Aspect 4) were aspects developed in the early stage, as shown in Figure 2, where the statistical significance refers to the pairwise comparisons between Issues 1 and 2 shown in Table 3. Oppositely, cognitive engagement (Aspect 1) developed in the late stage (i.e., between Issues 5 and 6). The difference implied that undergraduates could show epistemic transfer and emotional participation in a short time, but it took a long time to form stable epistemic ideals and reliable processes matching the ideals. Thirdly, a downward trend appeared in almost all aspects of the epistemic performance for the fifth sustainable development issue, “food and health”, while an upward trend was seen for the sixth issue (“nuclear power”), as shown in Figure 2, where the statistical significance refers to the pairwise comparisons between (1) Issue 4 and Issue 5 as well as (2) Issue 5 and Issue 6, as shown in Table 3, which indicated the possibility that undergraduates’ epistemic performance varied across different contexts. Fourthly, regulation (Aspect 3) showed an upward trend (this was only significant between Issues 5 and 6, referring to the pairwise comparisons between Issues 5 and 6, as shown in Table 3), indicating that the improvements in meta-epistemic competencies might grow abruptly through the accumulation of enough time. Additionally, the epistemic quality grew in a similar manner to regulation, indicating their potential interrelation.

4.2. A Story of Development: Sally’s Epistemic Growth in the Context of Sustainable Development Issues

To understand the statistical results, we tracked all comments created by an undergraduate named Sally (pseudonym). She was selected because the growth trend of her epistemic quality curve coincided with Figure 2 (right). We selected representative excerpts from her comments about three issues to show how her epistemic growth in argumentation occurred. The number(s) in the bracketed material within the quotes represent specific aspect (s).

4.2.1. Episode 1: “I Have to Do It, but How?”

Excerpt 1: “... Should the first question be answered from the perspective of professional knowledge? [Aspect 1]... I major in biology. I know the nature of the virus, but it seems that it’s not just a biological problem [Aspect 2]... Is this the right answer? We only have two minutes left. Let me fill in the blank of the first question [Aspects 1 and 4]...”

Excerpts of the comments on the topic of Sustainable Development Issue 1 (“virus and vaccines”) were collected. Sally tried to transfer familiar biological knowledge to the context of SSI (Aspect 2); however, she could not engage in the task in cognitive depth because of her vague and unclear epistemic ideals (Aspect 1). She was also unable to invoke epistemic metacognition because there were no criteria for her to evaluate the collaborative argumentation (Aspect 3). Although she was very concerned about whether the task was completed (Aspect 4), this emotional participation was non-epistemic [19]. Furthermore, she did not realize the cognitive existence of the team (Aspect 5).

4.2.2. Episode 2: “I Will Try My Best to Achieve My Success.”

Excerpt 2: “... I am familiar with knowledge of GMOs [Aspect 1]... I can search for and read some research papers with my understanding to solve the first question [Aspects 1, 2, and 3]... Alice, your comment is unreasonable. We promulgate a collective standard just now that we need to pay attention to where the information comes from (Aspect 3, 5)... Because I just read the biology literature in an authoritative journal, I’ll explain it to you [Aspects 1, 3, 4, and 5]...”

The excerpts were derived from the comments regarding Sustainable Development Issue 4 (“GMOs”). Sally realized that her familiar biological academic standards (Aspect 2) seemed to help her achieve the aims (Aspect 1), which made her enjoy (Aspect 4) invoking academic meta-competencies (Aspects 2 and 3) to engage cognitively (Aspect 1). Sally began to notice the cognitive existence of the team (Aspect 5), but she only regarded herself as a knowledge provider and a critic (Aspects 3 and 5). She did not care about the epistemic contribution of her peers. She considered “a collective standard” only as evidence that served her as a critic (Aspect 3).

4.2.3. Episode 3: “I Know How to Contribute to Our Team’s Success.”

Excerpt 3: “… I’m not familiar with this issue. I can check and put forward your solutions for us (Aspect 5)... Carl, do you major in physics? Can you deal with it? (Aspect 1)… We should firstly formulate the evaluation criteria (Aspect 3, 5) … The criteria we formulate in the last issue should be changed slightly, as this issue is different (Aspect 2, 3)… I suggest that each person proposes at least two ideas before integrating (Aspect 2, 3)… Alice, I don’t quite understand what you said, but since it is from the official website of the National Academy of Science, I think it is likely to be credible (Aspect 1)… Your response is logically complete, but I still wonder [Aspect 4]… Why not start voting [Aspect 5]…”

The excerpts were derived from the comments regarding Sustainable Development Issue 6 (“nuclear power”). Sally recognized the importance of the collaborative development of epistemic ideals (Aspects 1 and 5). Although she was not familiar with the domain knowledge of physics, she soon realized that there was a professional peer in the group (Aspects 1 and 5), and she also found her way to contribute epistemically to the group through inspection and suggestion (Aspects 2 and 3). Due to the differences in the contexts, she adjusted her epistemic ideals (Aspect 3) and regulated her peers from multiple perspectives (Aspect 2), which had been summarized and updated in the previous issues. On the basis of these optimized ideals, she invited her peers to jointly implement the processes of collecting evidence from multiple perspectives (Aspects 1 and 5). Even if some comments were unfamiliar, Sally could be skeptical (Aspect 4) and helped her peers achieved the epistemic aims by evaluating the sources and by initiating voting (Aspects 1, 2, and 5).

4.2.4. A Summary: How Did Sally Achieve Epistemic Growth in the Context of Sustainable Development Issues?

In the three episodes of Sally’s story, cognitive engagement (Aspect 1) was the starting point of collaborative argumentation (Aspect 5). She constantly explored the epistemic aims, ideals, and reliable processes via cognitive engagement (Aspect 1) from the individual to the group. Sally initially ignored the cognitive existence of collaboration (Aspect 5). When she began to participate in collaboration (Aspect 5), she viewed collaborative argumentation (Aspect 5) as a window for “output”, which shifted to viewing it as a pyramid for “team-based construction”. Gradually, cognitive engagement (Aspect 1) happened alongside collaborative participation (Aspect 5) at a high frequency. Adaptability (Aspect 2) and emotional participation (Aspect 4) played an important role in these processes. The former helped Sally link and transfer the epistemic performance in a familiar situation with cognitive engagement (Aspect 1) in the current task, and the latter strengthened her motivation for cognitive engagement (Aspect 1) and collaborative participation (Aspect 5). Her epistemic metacognition (Aspect 3) monitored and regulated the processes of cognitive engagement (Aspect 1), emotional participation (Aspect 4), and collaborative participation (Aspect 5) aptly (Aspect 2). Therefore, Sally’s epistemic metacognitive performance (Aspect 3) often occurred together with other aspects. In a comparison of Issue 1 and Issue 4, Sally showed more aspects of epistemic performance in her argumentative comments, especially in terms of meta-competencies (Aspect 3), which highlighted her improvement in epistemic quality. These results based on Sally’s comments were consistent with the results of descriptive statistics and repeated-measures ANOVA based on the five aspects of epistemic performance of all undergraduates in the context of sustainable development issues.

5. Discussion

This study investigated epistemic growth through the use of a blended collaborative learning design with reference to the Apt-AIR framework. The results indicated that the participants meaningfully developed their epistemic performance across the six sustainable development issues. The findings of our study are explored further below.

5.1. Longitudinal Effects of Sustainable Development Issue-Based Argumentation on Epistemic Performance

First, the findings showed significant growth in the frequency of each aspect of epistemic performance, indicating the positive effects of argumentation on the participants’ epistemic growth. In line with prior research [30], our findings also identified the significant contribution of computer-mediated argumentation that was designed to promote epistemic growth within our learning environment in the context of sustainable development issues. Furthermore, we found that adaptability (Aspect 2) might be the aspect that grew the most significantly among the five aspects. Possible explanations for the significant growth in adaptability might be that the prompts of multidiscipline-based feedback enabled the participants to perceive context-based features for creating comments (Aspect 2) [31].

Second, the five aspects of epistemic performance grew synchronously, which was consistent with the hypothesis of Barzilai and Chinn [3]. The qualitative findings of the present study identified the arguments that explained the shift in adherence to the reliable epistemic processes in solving problems related to sustainable development issues (Aspect 1). It would be necessary for more argumentative practice to offer multiple opportunities for learners across diverse contexts (Aspect 2) and enable social interactions (Aspect 5) with sufficient emotional engagement (Aspect 4). Moreover, together with the development of the other four aspects of epistemic performance, learners metacognitively adjusted their epistemic performance to reflect on issue-based problems (Aspect 3). This echoes Leung [9] regarding the importance of experiencing productive motivations and emotions for students to engage in epistemic performance beyond the classroom in a computer-supported context.

5.2. Theoretical Significance: Contribution to an Understanding of Epistemic Growth in the Context of Sustainable Development Issues

As the field of epistemic cognition has developed, scholars have become increasingly interested in identifying and promoting epistemic growth from the perspective of practical epistemology [17]. In this study, the five aspects of undergraduates’ epistemic performance were developed to varying degrees via computer-supported collaborative argumentation. The different aspects showed an overall upward trend in the development curve in the order of the issues, which was consistent with the previous hypothesis of the five aspects having internal correlation, as claimed by Barzilai and Chinn [3]. From Sally’s different performance between the familiar and unfamiliar stages, we can see that the extent of prior content knowledge intensely influenced epistemic performance, and these results were similar to the study of experts’ epistemic transfer [18]. We also noticed that meta-competencies acted on all aspects of epistemic performance through different forms of regulation (e.g., self-regulation, co-regulation, and socially shared regulation; see Järvelä and Hadwin [32]), such as monitoring argumentation via negotiated ideals, and formulating effective strategies to carry out a task via the cognitive division of labor and integration, which is line with the conceptualization of epistemic metacognition proposed by Barzilai and Zohar [23].

5.3. Implications for Epistemic Education in the Context of Sustainable Development Issues

Our findings suggest that students developed not only their epistemic performance but also their epistemic understanding of sustainability through their argumentation regarding sustainable issues. Similar to previous studies, learners’ argumentation ability progressed through issue-based argumentation [13]. Furthermore, we found that the undergraduates deepened their understanding of social epistemic systems regarding sustainable issues through collaborative argumentation. They gradually noticed that epistemic ideals are related to specific institutions (e.g., a specific academic domain), actively explored the differences across epistemic systems (e.g., the different positions of different groups of scientists on global warming), and tried to integrate these differences to draw conclusions based on multiple perspectives. We believe that the undergraduates’ epistemic growth in this study met the aims of epistemic education [1,2]. Obviously, in order for students to develop a deeper understanding of sustainable development, it is necessary to conduct richer issue-based argumentation in schools.

5.4. Contributions to Sustainability

There are three main points that this study contributes to sustainability. First, this study took sustainability as the basic background and allowed undergraduates to argue about sustainable development issues to become more familiar with sustainability. The experimental study was carried out beyond the setting of the sustainable development issues, illustrating that argumentation about sustainable development issues would be effective for improving people’s epistemic performance. Second, the undergraduates in this study significantly promoted their understanding and values in the field of sustainability. They learned some knowledge (e.g., why GM corn can be applied in agriculture) and methods (e.g., how to maintain the pH of the water within an appropriate range to maintain an ecological balance) through argumentation, as well as how to be people that value sustainability highly. Third, we established a framework (i.e., Table 2) to measure people’s understanding of sustainability in an epistemic approach. Traditional instruments have focused on people’s understanding of sustainability by using a Likert-like investigation and quantitative analyses. However, we suggest the approach of quantifying qualitative data to collect argumentative comments and a mixed-methods analysis, which presented finer-grained analytical results and enabled a more accurate analysis of people’s practical processes in argumentation about issues of sustainability [33].

In summary, we believe that this study provides enlightening insights to educators focusing on sustainable education. The people who participate in this study may increase their understanding of the concepts, knowledge, methods, and applications of sustainability. What is more, growth happens at epistemic level, that is, people can understand how their knowledge, methods, and application of sustainability have been generated, and how they may change in future to further recognize the necessity and importance of sustainability.

6. Limitations and Future Directions

Our study was necessarily limited by several factors. Firstly, although we took the heterogeneity of the participants into account as much as possible in the grouping (i.e., heterogeneity in terms of their major and gender), the fact that all participants in this study came from the same university may have imposed certain limitations on the generalizability of this study. Secondly, although the number of aspects included in a comment was used to describe the epistemic quality, the competencies of the undergraduates were not accurately evaluated. We believe that developing a more fine-grained standard for the five aspects (e.g., a scoring rubric for each) may be more helpful. Finally, the internal relationships among the five aspects found in our study need to be explored further. In-depth interviews are needed to investigate how and why undergraduates created some specific comments regarding the sustainable development issues.