Analysis of Critical Thinking Path of College Students Under STEAM Course

Abstract

This study examined the differences in critical thinking levels among students with different levels of academic engagement in STEAM courses. In this study, 30 college students were selected as subjects. Before experimenting, they received the academic engagement test and were divided into high, medium, and low groups based on their performance. Then, each group received three STEAM sessions and was asked to complete a topic discussion task. The results show that there are significant differences in the critical thinking level of students with different levels of academic engagement. Specifically, the students with a medium level of academic engagement had the highest critical thinking. Research has shown that the level of academic engagement affects the critical thinking of students in STEAM courses.

Introduction

Critical thinking is one of the twenty-first skills centuries for contemporary college students, ranking as the most sought-after higher-order thinking skills, along with creativity, collaboration, and problem-solving (Lai and Viering, 2012; Vasilyev et al., 2015; Podolsky and Pogozhina, 2017). Critical thinking is defined as purposeful, self-calibrated judgment. This kind of judgment manifests itself in interpretation, analysis, evaluation, inference, and the explanation of the evidence, concept, method, standard, or context on which the judgment depends (Nair and Lynnette Leeseberg, 2013).

The studies indicated that the STEAM course has the potential to improve the critical thinking of students (Allamin et al., 2018; Siregar et al., 2019). The STEAM course is an interdisciplinary and integrated course that combines science, technology, engineering, art, and mathematics (Yakman, 2008; Corbo et al., 2014; Hwang, 2017). In the STEAM course, students can focus on the specific problems rather than being confined to a single subject boundary, and they can practice their thinking from different perspectives and develop cross-border communication in the context of diversified development (Yakman, 2008; Corbo et al., 2014; Hwang, 2017; Hatlevik, 2018). The STEAM course is designed to guide learners to develop problem-solving, critical thinking, and collaboration skills (Tillinghast et al., 2015).

The previous studies results are not consistent on whether STEAM course improves the critical thinking of students. For example, Ridwan et al. (2020) implemented a STEAM course based on a smoke absorber project in a high school chemistry course. Through simple analysis of online discussion texts, the study found that STEAM courses can promote the development of the critical thinking of students (Ridwan et al., 2020). However, other researchers have found a different result. For example, Ho Sha (2019) explored the changes in the critical thinking level of students before and after the implementation of the STEAM course and found that the critical thinking level of students did not improve significantly.

The cultivation of critical thinking in STEAM courses may be influenced by academic engagement. In the study, each student has a different degree of academic engagement (Curry, 1984; Nystrand, 1989; Cancelli, 1993). The research found that academic investment can affect the self-efficacy of students (Yüksel and Alci, 2012; Samareh and Kezri, 2016). Specifically, students with high education have more self-efficacy (Samareh and Kezri, 2016). In addition, the study found that self-efficacy and critical thinking were significantly correlated, specifically, the higher the self-efficacy of students, the higher the level of academic engagement (McKinnon, 2012; Dong, 2016). Then, academic engagement may affect critical thinking in the STEAM curriculum.

This study aimed to examine whether the implementation of STEAM courses had significant differences in the development of critical thinking levels of students at different levels of academic engagement (high, medium, and low). Based on previous research, this study hypothesizes that academic engagement affects the critical thinking of students in STEAM courses. Specifically, the higher the level of academic engagement, the higher is the critical thinking level of students in STEAM courses.

Methods

Participants

The participants in this study were 30 college students (14 male students and 16 female students). The average age of the participants was 20.43 (SD = 0.89). All participants received a small gift worth 30 RMB after the completion of the experiment. The study protocol was approved by the local Academic Committee.

This experiment was an in-subject experiment design, and all subjects received three STEAM courses. The subjects were asked to fill out the questionnaire before receiving the experimental treatment, and according to the results, the subjects were divided into groups with high, medium, and low academic involvement.

Measures

Academic Engagement Scale

The Academic Engagement Scale (Awang-Hashim and Sani, 2008) is used to measure the academic engagement of students in school learning activities. The scale is divided into three dimensions: “behavioral engagement,” “cognitive engagement,” and “emotional engagement.” The scale has a total of 29 items, which consist of eight items of behavioral engagement, 10 items of cognitive engagement, and 11 items of affective engagement. The scale uses Likert-five assessment, with 1 representing “completely inconsistent” and 5 representing “completely consistent.” The higher the score, the higher is the degree of academic engagement. The internal consistency coefficient of the scale was 0.851, among which the behavioral engagement dimension coefficient was 0.782, the cognitive engagement dimension coefficient was 0.835, and the affective engagement dimension coefficient was measured to 0.742.

Critical Thinking Coding Tool

The study adopted the critical thinking coding tool proposed by Murphy (2004) and added the “Null” dimension (encoding topics unrelated to the course and opinions purely expressing emotions) in this study. The critical thinking coding tool is divided into six dimensions: recognition, understanding, analysis, evaluation, creation, and null value, including a total of 25 specific indicators.

Online Collaboration Platform

Each group establishes a group online cooperative learning community through QQ. It is a widely used instant messaging software in China, which can realize voice and text communication, file transfer, and other functions. Before the experiment, the participants were informed that the teacher would post the course tasks that needed to be solved in each class on the QQ group and asked them to discuss according to the task requirements proposed by the teacher, and finally put forward a group plan.

Procedure

Before the formal experiment, each participant was required to complete the Academic Engagement Scale. Then, the experimenter divided the participants into three groups according to the academic engagement questionnaire: high, medium, and low. After that, all participants were given three STEAM sessions. Each course lasts 90 min, and the contents of the three courses are shown in Table 1.

At the end of the three sessions, the online discussion texts were collated for coding and lag sequence analysis (LSA). First, the two researchers coded the code together. They are familiar with coding tools for critical thinking and have discussions before coding. In the coding process, the two researchers separately encoded the text of the online discussion. After the formal coding was completed, 154 texts were randomly checked by the two coders to verify the consistency of the coding results of the two researchers. The results show that the Kappa consistency coefficient is 0.75, which indicates that the data encoding has good reliability and can be used for LSA. Second, GSEQ software is used to analyze the lag sequence. Lagging sequence analysis produces two important tables, namely, behavior conversion frequency table and adjustment residual table. According to the theory of LSA, the Z-score >1.96 indicates that this behavior path is significant (Sjalander et al., 1996). Finally, the behavior transformation diagram is drawn according to the important behavior sequence. The result is research on critical thinking levels.

Results

The Critical Thinking Pathways of Participants With Different Levels of Academic Engagement After the First STEAM Course

According to the coding results, the characteristics of critical thinking pathways of the low, medium, and high academic engagement group students after the first course were analyzed. The specific results are analyzed as follows.

The results showed that, for students with low academic engagement, there are four kinds and five effective single sequences in this stage, such as R2→U5, U5→A4, R1→E1, and C1→C2, as shown in Table 2. For the first time to talk about the single sequence including identification agree with to understand answer questions (R2→U5), understanding to answer the question—analysis explanation view (U52→A4), preliminary identification problem to determine the effectiveness of the current information and value point (R12→E1), project implementation plan to create design new ideas (C12→C2) sequence, respectively 1, 2, 1, 1.

The adjusted residual was calculated by GSEQ software (as shown in Table 3). Among them, there are two significant sequences with residual values >1.96, namely, preliminary identification of problems → judgment of the validity and value points of current information (R1→E1), application of implementation plan → creation of new ideas (C1→C2). In addition, according to the adjusted sequence of significant behavioral residuals, a complete directed path diagram of learning activities is generated, as shown in Figure 1. As can be seen from the visual path diagram discussed in the first discussion, students in the low engagement group initially identified the problem and then judged the validity and value of the information provided. Students will be inspired to put forward new ideas, strategies, and methods based on applying new ideas and plans.

Figure 1

The results showed that for students with medium academic engagement, U5→A4, A4→A5 are generated in this stage, with a total of 25 kinds and 46 effective single sequences, as shown in Table 4. The single sequence generated in the first discussion includes understanding and answering questions → analyzing and explaining viewpoints (U5→A4), analyzing and explaining viewpoints → decomposition of problem viewpoints (A4→A5). There are six and eight sequences, respectively, and there are relatively few process sequences from creating (C).

As can be seen from the visual path diagram of the first discussion, after the students understand and answer the questions, they then analyze and explain the questions, and further decompose the views of the current problems, to create new views for the current problems to be solved (U5→A4→A5→C2). During the discussion, students usually create new ideas (C1→C2) based on creating and implementing new ideas, new conclusions, or new plans. The next step in the classification of evidence is usually to further analyze the similarities and differences between opinions (A2→A3). After the preliminary identification of repeated problems, most students responded to agree with the current specific problems. Then, they will join themselves to solve the current problems of new or new ways of thinking, next, the difference in the behavior way, part of the society, and to detect the current point of view of the consistency and inconsistency. Finally, the students will show their acceptance or refutation of the views and explain why and list the arguments (R1→R2→A1→E3→E6). The other students will first judge the validity of the information sources and the value of the information itself, and then, they further criticize the views themselves, and then put forward their views and find relevant arguments to support the arguments. Finally, we will summarize and state the acceptance or refutation and explain the reasons for the argument (R1→R2→A1→E1→E2→E5→E6). In this process, some students will first test the consistency and inconsistency after criticizing the opinions. Then, directly discuss the acceptance or refutation and explain the reasons for the argument (R1→R2→A1→E1→E2→E3→E6). The common points presented by the visualization path chart are: students will first identify and clarify the existing views, problems, and contradictions, then, explain in-depth, organize the known information, clarify the unknown information, decompose the basic elements of the opinions or problems, and finally clearly criticize and judge the information or opinions. In addition, according to the adjusted sequence of significant behavioral residuals, a complete directed path diagram of learning activities is generated, as shown in Figure 2.

Figure 2

Among them, there were 15 significant sequences with residual values greater than 1.96 (as shown in Table 5).In other words, identify the restatement question → identify the agreement question (R1→R2), identify the agreement opinion → understand and answer the question (R2→U2), identify the agreement question → add new thinking behavior (R2→A1), understand and answer the question → analyze and explain the opinion (U5→A4), add new thinking behavior → judge the validity of information (A1→E1), add new thinking behavior → test 1Causes and inconsistencies (A1→E3), evidence classification → analysis of opinion similarities and differences (A2→A3), analysis and interpretation of opinion → decomposition of problem opinion (A4→A5), decomposition of problem opinion → creation of new opinion (A5→C2), judgment of information validity → critical opinion hypothesis (E1→E2), critical opinion hypothesis → detection of consistency and inconsistencies (E2→E3), critical opinion hypothesis → evidence supporting the argument (E2→E5), detection of consistency and inconsistency → refute acceptance and evidence (E3→E6), evidence supporting the argument → refute acceptance and evidence (E5→E6), evidence supporting the argument → create a new idea (E5→C2), application of implementation plan → create a new idea (C1→C2).

The results showed that for students with high academic engagement, this stage mainly produces 18 kinds of R2→A4 and 21 effective single sequences, as shown in Table 6. The single sequences generated in the first discussion mainly include identification and approval point → analysis and explanation point (A4→A5), and there are four sequences in total.

Among them, there are five significant sequences with a residual value >1.96 (as shown in Table 7). Identifying and reiterating questions → understanding and answering questions (R1→U5), identifying and reiterating questions → analyzing similarities and differences of viewpoints (R1→A3), searching for different evidence → classification and classification of evidence (U3→A2), analyzing and explaining viewpoints → decomposition of viewpoints (A4→A5), and criticizing assumptions → identifying and approving questions (E2→R2). In addition, according to the adjusted sequence of significant behavioral residuals, a complete directed path diagram of learning activities is generated, as shown in Figure 3. As can be seen from the visual path diagram of the first discussion, after the initial identification of the problem, some students in the high engagement group proceeded to evaluate and clarify the problem to explore its essence (R1→U5), while some students proceeded to analyze the similarities and differences of various viewpoints (R1→A3). After looking for different pieces of evidence, students will further classify the collected evidence (U3→A2). After analyzing and explaining ideas, students will further decompose their ideas (A4→A5). Students criticize the opinion, but at the same time, they put forward the part that they agree with (E2→R2).

Figure 3

The results above indicated that, to sum up, in the first discussion, students with high, medium, and low academic engagement showed significant differences in the visualized critical thinking activity path diagrams in the discussion activities (as shown in Figure 4). However, contrary to the expectations, students in both the low and medium engagement groups reached the highest level of critical thinking—creativity; students in the high engagement group only reached the evaluation level. In contrast to the previous hypothesis, there were more significant behavioral sequences in the medium engagement group than in the high engagement group. The significant behavior sequence of critical thinking activities of students in the high and medium engagement groups was more than that in the low engagement group, which was consistent with the hypothesis of the study.

Figure 4

Critical Thinking Pathways of Participants With Different Levels of Academic Engagement After the Second STEAM Course

According to the coding results, the characteristics of critical thinking pathways of students in low, medium, and high academic engagement groups were analyzed. The specific results are analyzed as follows.

For students with low academic engagement, only one effective single sequence U5→A4 is generated at this stage, as shown in Table 8. However, there was no significant behavioral residual (as shown in Table 9).

For students with medium academic engagement, 12 kinds and 18 effective single sequences are generated in this stage, such as U2→A2, R1→U2, R2→R1, A2→C1, as shown in Table 10. The second discussion produced more single sequences mainly for understanding and seeking relevant content → evidence classification and classification (U2→A2). There were four sequences in total and relatively few sequences from the process of creating (C).

Among them, there are five significant sequences with a residual value >1.96 (as shown in Table 11). That is, to identify and reiterate the problem → understand and search for relevant content (R1→U2), understand and search for relevant content → classification and classification of evidence (U2→A2), classification and classification of evidence → application and implementation plan (A2→C1), analysis and explanation of viewpoints → decomposition of problem viewpoints (A4→A5), and supporting arguments → identification and reiterate the problem (E5→R1). According to the adjusted sequence of significant behavioral residuals, a complete directed path diagram of learning activities is generated, as shown in Figure 5. As can be seen from the visual path diagram of the second discussion, students in the input group initially identified the problem, then searched for some contents related to the problem, and then classified these contents. Finally, based on this, they implemented the new plans and ideas according to the requirements of the problem (R1→U2→A2→C1). After analyzing the current opinion, students will usually further break down the current opinion or the problem (A4→A5). After using evidence to support the argument, students will further reiterate the problem to be solved (E5→R1).