2.1. Educational and Psychological Foundations

This study is rooted in educational and psychological theories that emphasize learner-centered approaches and social interaction as critical components of learning. The first important theory is Vygotsky’s (1978) sociocultural theory, which highlights the importance of collaboration and social interaction in cognitive development, specifically through scaffolding provided by peers and instructors. This theory aligns with the study’s focus on how background characteristics, such as gender, class level, or personality type, may influence participation and collaboration in BL environments. Similarly, the second important theory, Bandura’s (1986) social constructivist perspective, highlights the value of observation, reflection, and interaction in shaping individual learning experiences. Together, these foundational theories emphasize how discussion-based and collaborative tasks create opportunities for meaningful interaction and scaffolded learning experiences.

Additionally, Kolb and Kolb’s Experiential Learning Cycle (2017) provides further insight into how learners construct knowledge through active engagement. The model outlines a four-step process, including Concrete Experience, Reflective Observation, Abstract Conceptualization, and Active Experimentation, which emphasizes the dynamic process of learning through doing and reflecting (See Figure 1). While the model offers a foundation for designing effective learning tasks, its application in this study focuses on how individual differences in background characteristics shape the ways learners engage with and benefit from the learning process as a whole. For example, learners with different personality types may approach reflective activities, such as self- and peer evaluations, in unique ways, potentially influencing their overall learning outcomes.

[Figure 1]

Experiential Learning Cycle (Kolb & Kolb, 2017, p. 11)

In sum, these theoretical foundations highlight the significance of understanding how diverse learner characteristics interact with educational practices. They also offer a framework for interpreting the findings of this study, particularly in terms of how gender, class level, and personality type might impact learner achievement, self-assessment, peer assessment, collaboration, and satisfaction.

2.2. BL and Instructional Activities in Relation to Background Characteristics

BL instructional approaches have become increasingly significant in the context of EFL, particularly in promoting student engagement and enhancing language skills. Research has shown that combining face-to-face and online learning components in a carefully structured approach can provide students with more active, flexible, personalized, and beneficial learning experiences (Bae & Lee, 2021; Irvin & Im, 2022; Lee, 2017; Klemsen & Seong, 2012; Riwayatiningsih & Sulistyani, 2020). This method, which blends the use of digital platforms alongside traditional classroom instruction, allows students to engage with content in ways that address their learning preferences and individual needs. In the context of English language courses, BL creates opportunities for students to practice skills in a variety of settings, encouraging meaningful communication and fostering collaborative learning (Garrison & Kanuka, 2004; Graham, 2006).

One commonly used activity utilized by many General Education professors as part of a BL instructional approach is a DB. DBs are often selected because they enable students to engage in online dialogue, collaborate on assignments, share opinions with one another, establish feedback mechanisms, and allow learners to reflect on learning outcomes. Moreover, DBs have been shown to promote critical thinking, peer interaction, and self- directed learning, which are all crucial for language acquisition and improving academic achievement (Kang & Park, 2014; Alharbi, 2015). However, while much of the existing literature related to DB use in language courses found increases in academic achievement and motivation (Im, 2022; AlJeraisy et al., 2015), increased learner engagement and more peer interactions (Dalelio, 2013), increases in collaboration (Bakar et al., 2013), greater satisfaction (Lee, 2017; Dalelio, 2013), and a preference towards audio instead of video submissions (Hew & Cheung, 2012), only a small number of studies have explored the influence of background characteristics on these outcomes, and even fewer have considered them in combination.

Finally, student-created videos, a form of Learner- Generated Digital Media, can be a valuable tool to encourage more interactive, self-reflective, and collaborative learning environments that meet diverse student needs (Reyna & Meier, 2018). Previous research suggests that student-created videos can enhance language learning by promoting creativity, engagement, and language proficiency (Edwards & Lane, 2021; Lee & McLoughlin, 2007). Furthermore, prior research has indicated that these videos have led to increased motivation and active participation (Lee & McLoughlin, 2007), facilitated collaboration through peer feedback and shared work (Seong & Kim, 2023; Hung et al., 2004), and enhanced speaking skills while fostering digital literacy and self-confidence (Im, 2021; Edwards & Lane, 2021; Innes, 2020). Thus, student- created videos have been shown to have several benefits as stand-alone activities. However, Irvin (2024) attempted to expand on the use of student-created videos in a General English course by incorporating them into a 4-skills DB BL approach. This novel approach offered learners opportunities to develop speaking, listening, reading, and writing skills simultaneously. Despite their efforts, few other studies have examined how background characteristics, such as gender, class level, and personality type, might affect students’ learning outcomes. This gap is significant, as such characteristics may play a crucial role in shaping student experiences.

In sum, the integration of student-created videos, DBs, and BL in this study presents a unique opportunity to explore how these approaches intersect with learner characteristics. Specifically, this study seeks to examine how gender, class level, and personality type might impact learner achievement, self-assessment, peer assessment, collaboration, and satisfaction.

2.3. Background Characteristics and Their Impact on BL Courses

3.1. Research Design

This study was conducted as part of a larger research project that examined the potential main effects and interaction effects of adopting a 4-skills DB BL approach to enhance Korean EFL learners’ outcomes, with a particular focus on the use of a 4-skills DB incorporating student-created videos (Irvin, 2024). While the current research shares many similarities with the previous project, it specifically focuses on quantitative analysis of the main effects of learners’ background characteristics (i.e., gender, class level, and personality type) in relation to their learning outcomes across the three different DB types (i.e., 4-skills, 2-skills, and No DB).

To examine these effects, this study was conducted during the Spring 2023 semester (i.e., March to June) at a private university in Korea, in a first-year EFL General English course (i.e., College English 1 or CE1) emphasizing 4-skills language acquisition. The course followed a BL approach, integrating both face-to-face instruction and online DB activities, with three different DB groups: a 4-skills DB incorporating student-created videos with typed posts, a 2-skills DB focused on the reading and writing of typed posts, and a No DB group that relied solely on in-class activities. Convenience sampling was used to select participants from six course sections taught by the researcher, who also served as the instructor. While this dual role raised ethical considerations, it provided a deeper understanding of students’ experiences through direct observation (Griffith, 1998). At this university, the CE1 course is divided into three levels (i.e., low, middle, and high), with student placement determined by the university’s administration. However, due to an uneven distribution of course sections, the middle- and low-level groups were merged into a single middle-low level. Each DB type was then randomly assigned one high-level and one middle-low-level section, with random assignment based on balancing the groups according to the number of participants rather than individual background characteristics. Despite efforts to maintain participation across groups, the final sample showed a 48% retention rate after excluding students who opted out or failed to complete required assignments, surveys, or evaluations. Thus, the study design builds on prior BL research while offering insights into how individual differences influence learning in this context.

3.2. Participants

Building on the research design, this study focused on students randomly assigned across six CE1 sections. As previously mentioned, CE1 is a first-year General English course required for graduation. While this course is primarily targeted at freshmen, there are occasions when second- to fourth-year students must take it to fulfill their graduation requirements. Additionally, students take a TOEIC-style entrance exam before their first semester at the university to determine their class levels. Those scoring below 400 are placed in the low-level class, those scoring between 400 and 700 are placed in the middle- level class, and those scoring above 700 are placed in the high-level class. Students who do not take the exam are generally placed in the middle-level class by default. At the start of the semester, a total of 168 students were enrolled across the six CE1 sections taught by the researcher, with each section randomly assigned a specific DB type. By the semester’s end, 81 students took part in the study, representing a 52% attrition rate. The average age of the participants was 21.1 years in Korean age, and a full breakdown of the participants’ background characteristics can be found in Table 1.

<Table 1>

Background Characteristics of Study Participants

A closer examination of Table 1 reveals the demographic breakdown for the entire sample and across the three DB types. The final sample included 81 participants, with a higher proportion of females (n = 50), many more students in the high-level sections (n = 52), and a majority of self-identified introverts (n = 58), as determined by self-reported MBTI results. A further look at Table 1 provides the breakdown of demographic characteristics for each of the three DB types. First, the 4-skills DB group consisted of 29 students, with a majority of females (n = 21), a considerably higher number of high-level students (n = 22), and most participants identifying as introverted (n = 24). Second, the 2-skills DB group included 27 students, with a slightly larger proportion of females (n = 15), slightly more middle-low-level students (n = 14), and a majority of self-reported introverts (n = 20). Third, the No DB group consisted of 25 students, with slightly more females (n = 14), double the number of high-level students (n = 17), and slightly more self- reported introverted participants (n = 14).

Overall, the complete breakdown of Table 1 highlights the demographic characteristics across the three DB types. Several factors, however, influenced the final distribution of participants. Firstly, the low sign-up rate for certain sections led to an uneven distribution, with one section starting with 33 while another began with only 20. Next, the gender balance within specific DB groups was influenced by the majors of the students, with certain majors having a higher proportion of either females or males, leading to an imbalance across the groups. Lastly, a higher-than-expected attrition rate, particularly in the middle-low-level sections, became more pronounced as the semester progressed. Each week, the researcher observed that more students failed to complete their required assignments, surveys, or evaluations, leading to their eventual exclusion from the study. This resulted in a reduction in all six sections; however, the final participant list revealed a significantly higher attrition rate in the middle-low-level sections compared to the high-level groups, most notably in the 4-skills and No DB groups.

3.3. Procedures

Following the participant details and course structure described in Section 3.2, this section outlines the procedural steps taken throughout the study. As previously stated, CE1 is a first-year General English course with sections that have three distinct class levels. Each level follows a standardized curriculum established by the department, which includes common primary objectives and a shared textbook series. However, instructors have the autonomy to adjust their course delivery to fit their teaching style and the needs of their students. As such, a BL approach was employed in this study, combining in-class instruction with online DB activities. The DB types were chosen because they provided students with opportunities to practice and build their English skills outside of the classroom. In particular, the 4-skills DB offered extra chances for students to practice speaking and listening through student-created videos. To compare experiences across different learning environments, a No DB group was also included in the study. The DB types were selected to examine how different course designs influenced learning outcomes, with a specific focus on participants’ background characteristics. This design aimed to explore how these characteristics shaped student experiences across the three DB types.

To ensure that this study met these objectives, a pilot study was conducted in Fall 2022 to refine the measurement tools and develop a procedural timeline. Insights gained from this pilot study ensured consistency in how tasks were assigned and assessed across DB types, as well as the overall course structure. Table 2 presents a detailed procedural timeline, outlining the sequence of activities and tasks completed throughout the semester.

<Table 2>

Procedural Timeline for the Study

The timeline presented in Table 2 shows that throughout the 15-week course, all students completed the same tasks across the three DB types and the No DB group. This ensured consistency while accommodating the different structures of each DB type. The final goal of the course was for students to create a 2-5-minute demonstrative speech video, which required them to record a video, self-evaluate their work, and evaluate at least two peer videos. To support this, the course began with initial homework assignments aimed at familiarizing students with the school’s Learning Management System (LMS) and providing them with opportunities to learn how to create digital media. After these preliminary tasks, students participated in four distinct DB activities, each focused on a specific topic (e.g., Likes/Dislikes or Music - Favorite Song). These tasks were customized to the DB type, allowing each group to complete the activities in their corresponding modality. Throughout the course, adjustments were made only to fit the structure of the DB type, ensuring that the course objectives and English language use remained consistent across all groups.

The course structure presented in Table 2 was further supported by specific tasks designed to promote students’ skill development. The first phase began with participants completing a background survey, followed by Homework 1, which required students to create an introduction video, allowing them to practice the skills of video creation and submission. Homework 2 asked students to create a 2-5-minute video about their most precious item, serving as a pre-test to assess presentation skills. Additionally, students were asked to complete a self-evaluation of their Homework 2 videos to familiarize themselves with the evaluation rubric. After these initial assignments, students participated in two discussion activities during the first half of the semester, focusing on simple topics like preferences and favorite songs. These activities helped students become familiar with the different modalities, whether online discussions for the DB groups or graphic organizers for the No DB group. Following DB/Discussion 1, each group created summary videos, allowing them to further practice self- and peer evaluations. Furthermore, following Discussion 2, the No DB group was asked to complete a summary paragraph to align with the expectations of the DB groups. The second half of the course expanded on these activities, with each group participating in more complex tasks, such as debates and preparing topics for their final project. The 4-skills DB group combined video, text-based responses, and peer evaluations to strengthen speaking and listening skills, while the 2-skills DB group focused on reading and writing through text discussions and in-class activities. The No DB group participated in in-person discussions and written assignments, enhancing speaking and writing abilities. All students completed the final project, which involved creating a demonstrative speech video, along with a self-evaluation and evaluations of at least two peers. Lastly, the semester concluded with an end-of-semester survey, allowing students to reflect on their experiences.

Finally, it is important to acknowledge the dual role of the researcher in this study. The author served not only as the researcher but also as the instructor for all six CE1 sections. While this role could introduce biases in data collection or influence student participation, Griffith (1998) argues that the ”insider” role can provide valuable insights into students’ experiences that an external researcher might miss. This dual role allowed for closer observations, enhancing the depth and validity of the findings, as the researcher was able to directly observe student interactions and adjust instructional strategies as needed.

3.4. Measurement Tools

This study used three measurement tools to assess the study’s five dependent variables: academic achievement, self- and peer evaluation, collaboration, and satisfaction. The first tool was the Public Speaking Competence Rubric (PSCR), developed by Schreiber et al. (2012), which contains 11 items to assess learners’ speech presentations. For this study, the original PSCR was adapted to include 10 of the original items to assess students’ content, verbal skills, and nonverbal skills. Each participant was required to submit two 2- to 5-minute videos, one at the start of the semester (i.e., the pre-test) and one at the end of the semester (i.e., the post-test). Furthermore, three reviewers, including the researcher, were recruited to evaluate the videos. Reliability for the pre- and post-test videos was confirmed with Cronbach’s Alpha values of 0.759 and 0.848, respectively, and inter-rater reliability showed coefficients of 0.853 for the pre-test and 0.870 for the post-test, indicating high reliability.

The second tool was the Tertiary-level English Oral Presentation Scale (TEOPS), proposed by Aryadoust (2015). This 18-item rubric allows students to assess both their own and their peers’ presentations, focusing on content, organization, verbal skills, and nonverbal communication. Participants completed self- and peer evaluations twice during the semester: once after the first DB/Discussion and again after the final exam video. Reliability tests showed that the Cronbach’s Alpha for the second self-evaluation was slightly higher (α = 0.876) than the first self-evaluation (α = 0.859), with both considered very reliable. However, the opposite was true for peer evaluation. The Cronbach’s Alpha for the first evaluation was slightly higher (α = 0.879) than the second evaluation (α = 0.850), with both considered very reliable, confirming the robustness of the TEOPS as an evaluation tool.

The third measurement tool was the end-of-semester survey, adapted from Hsieh (2015), which originally included 34 quantitative questions regarding collaboration, satisfaction, and technology use. For this study, the end-of-semester survey was revised to focus on collaboration, satisfaction, and participants’ experiences based on the specific DB type to which they were randomly assigned. The final version contained 23 quantitative items: 9 on collaboration and 14 on satisfaction, using a 4-point Likert scale ranging from ”Strongly Agree” to ”Strongly Disagree.” Reliability for these sections was high, with Cronbach’s Alpha values of 0.881 for collaboration and 0.844 for satisfaction. Additionally, seven descriptive items were added to the end-of-semester survey to complement the existing questionnaire. These items included five Likert-scale questions assessing participants’ self-reported English proficiency (reading, writing, speaking, listening, and overall), as well as two preferential questions about their favorite and least favorite assignments. These additions were made to provide greater context regarding students’ experiences and perceptions of participating in the study across the three DB types.

3.5. Data Collection and Analysis

Data collection for this study began with the administration of a background survey. Anonymity was a major ethical consideration, and to ensure this, all participant data were securely stored in a locked room and coded with random numbers (e.g., s1, s2…s81) to protect identities. Hard copies of the PSCR and TEOPS rubrics were provided for training, and virtual copies were distributed via the university’s LMS. Pre- and post-test videos were organized by participant codes and uploaded to a cloud service for evaluation. Three evaluators, including the researcher, all with more than ten years of experience teaching General English courses, used the PSCR to evaluate students’ academic achievement, and the TEOPS was used by learners to rate themselves and their peers. Self- and peer evaluations along with the end-of-semester survey were collected using Google Forms to maintain consistency across participants. Data were subsequently compiled into a master Excel sheet for analysis, with only the researcher and advisor granted access to sensitive data.

Data analysis for this study utilized factorial analysis of variance (ANOVA) to examine the main effects of the independent variables (i.e., gender, class level, and personality type) on five dependent variables: academic achievement, self-evaluation, peer evaluation, collaboration, and satisfaction. ANOVA tests were performed using Statistical Package for the Social Sciences (SPSS) version 27 to analyze the main effects of each background characteristic (Myers et al., 2013). Secondary analysis was conducted on selected questions from the end-of-semester survey related to each background characteristic that were used to provide context for the quantitative findings. Additionally, as a project ‘insider’ (Griffith, 1998), the researcher drew on instructor observations to contribute to the interpretation of certain quantitative data findings.

4.1. Findings Related to Gender

The first background characteristic considered in this study was gender. A factorial ANOVA test was conducted to determine whether significant differences existed between females and males in terms of academic achievement, self-evaluations, peer evaluations, collaboration, and satisfaction. The results for gender are presented in Table 3.

<Table 3>

Factorial ANOVA for Gender and the Five Dependent Variables

A closer look at Table 3 reveals no significant difference between females and males in terms of self-evaluations, peer evaluations, and satisfaction. However, a significant gender difference was found for academic achievement and collaboration. For academic achievement, F(2, 78) = 5.72, р = 0.019, partial п² = 0.071, with females (M = 8.85) outperforming males (M = 8.19). In collaboration, F(2, 78) = 2.12, р = 0.025, partial п² = 0.065, females (M = 3.45) reported higher collaboration levels than males (M = 3.20).

The quantitative results for gender showed mixed results, indicating a lack of a clear consensus and aligning with previous research. On one hand, three of the five dependent variables in this study had no significant differences, which are consistent with prior studies such as Cha et al. (2022) and Machado (2011), both of which also reported no significant gender differences. On the other hand, the two significant findings related to academic achievement and collaboration are in line with other studies suggesting that females often demonstrate better learning styles and higher engagement levels, leading to superior performance (Lee et al., 2003; Suh, 2010; Bostock & Lizhi, 2005; Tsai et al., 2015). To further understand these results, analysis from the end-of- semester survey offered additional context.

One notable observation from the end-of-semester survey was the absence of comments explicitly mentioning gender differences. This supports Valenziano’s (2008) suggestion that, while gender may influence learners’ interaction patterns, students often do not recognize or acknowledge these differences. This lack of recognition might reflect a diminishing impact of gender stereotypes inside Korean general education classrooms (Lee et al., 2003; Tsai et al., 2015). Nevertheless, future research could explore frameworks (Reyna et al., 2017) that focus on leveraging strengths and addressing challenges specific to each gender group.

A second notable observation from the end-of-semester survey was the discrepancy in self-reported skill improvements. While Table 3 shows that females statistically outperformed males in academic achievement, responses to self-reported Likert-scale questions related to perceived skill improvements revealed that male participants self-reported higher gains in every category, as illustrated in Figure 2. This discrepancy might stem from the unequal gender distribution among participants. In this study, females represented approximately 24% more than males, with the 4-skills DB group consisting of around 73% females. Alternatively, this result could suggest that female learners may undervalue their contributions and strengths in collaborative and academic settings.

[Figure 2]

Self-Reported Skill Gains at the End of the CE1 Course

Finally, observations from this study highlighted the relationship between gender representation and academic majors. As noted in Section 3, many class sections were organized by college, department, or major. This study found that majors such as nursing tend to have a higher percentage of female students, while natural science majors were generally more balanced or had a higher percentage of male students. As noted by Krentler and Willis-Flurry (2005), student characteristics, including gender and major, significantly influence the adoption of ICT tools and class participation. This study extends their findings, suggesting that gender demographics also influence collaboration levels. Notably, female students were found to be significantly more collaborative than males, underscoring the need to consider gender composition when designing BL courses.

4.2. Findings Related to Class Level

The second background characteristic considered for this study was class level. A factorial ANOVA test was conducted to determine whether significant differences existed between high-level and middle-low-level learners in terms of academic achievement, self-evaluations, peer evaluations, collaboration, and satisfaction. The factorial analysis results for class level are presented in Table 4.

<Table 4>

Factorial ANOVA for Class Level and the Five Dependent Variables

A closer examination of Table 4 reveals no significant differences between high-level and middle-low-level learners in terms of collaboration and satisfaction. However, significant class-level differences were observed for academic achievement, self-evaluations, and peer evaluation. For academic achievement, F(2, 78) = 25.71, р = 0.000, partial п² = 0.255, with high-level learners (M = 9.02) significantly outperforming middle-low-level learners (M = 7.85). For self-evaluations, F(2, 78) = 7.10, р = 0.009, partial п² = 0.240, high-level learners (M = 3.96) scored slightly higher than middle-low-level learners (M = 3.70). In peer evaluations, F(2, 78) = 23.69, р = 0.000, partial п² = 0.240, high-level learners (M = 3.84) received higher ratings than middle-low-level learners (M = 3.35).

One notable observation from the instructor was related to the attrition/retention rate. As outlined in the methods section, the original design for this study aimed for a minimum of 50 participants per group, and the numbers were on track during the first week. However, as the semester progressed, participant retention dropped sharply. Of the initial 168 participants, 91 were high-level learners and 77 were middle-low-level learners. By the end of the study, the final sample consisted of 81 participants, with the high-level group retaining 19% more students than the middle-low-level group (i.e., 57% retention for the high-level group compared to 38% for the middle- low-level group). Notably, the middle-low-level section in the 4-skills DB group had a 74% attrition rate, primarily due to absenteeism and failure to complete tasks. This finding highlights the greater commitment and persistence displayed by high-level learners compared to their middle-low-level counterparts. Furthermore, this finding aligns with research by Rodriguez-Juarez and Oxbrow (2010), who asserted that students with higher proficiency levels are often more motivated and dedicated. Thus, the differences in retention rates between high-level and middle-low-level learners support the idea that high-level learners tend to be more academically engaged, resulting in more consistent participation and task completion.

A second key observation from the end-of-semester survey focused on learner motivation, specifically related to task engagement and perceived usefulness. Previous research on class level and DBs suggests that learner preferences, especially regarding DB topics, are more important for high-level learners (Bang & Park, 2020; Yang, 2021). However, this study’s results slightly contradicted that assertion. When asked to select their favorite and least favorite DB activities, 23 out of 52 high-level students (i.e., 44%) selected DB/Discussion 2 as their favorite, while 11 out of 29 middle-low-level learners (i.e., 38%) chose the same activity. Similarly, when asked to select their least favorite DB task, both groups selected DB/Discussion 4 (i.e., 32% of high-level students and 41% of middle-low-level learners). These results align with the findings of Irvin and Im (2022), who suggested that learners are more motivated by DB tasks they find enjoyable. This study also highlights that both high-level and middle-low-level learners value engaging and interesting DB topics, underscoring the universal importance of task design in fostering learner motivation, regardless of proficiency level.

4.3. Findings Related to Personality Type

The final background characteristic analyzed in this study was participants’ personality type. Specifically, learners’ personality types were classified as extroversion or introversion based on self-reported MBTI data. A factorial ANOVA was conducted to assess whether differences between extroverts and introverts existed across the five dependent variables: academic achievement, self- evaluation, peer evaluation, collaboration, and satisfaction. The results are summarized in Table 5.

<Table 5>

Factorial ANOVA for Personality Type and the Five Dependent Variables

A closer look at Table 5 revealed no statistically significant differences between extroverts and introverts for any of the dependent variables. These findings contradict prior research suggesting that extroverts might benefit more from active and collaborative approaches, such as BL environments that emphasize speaking and discussion (Chew & Ng, 2016; Harland, 2005; Lee & Lee, 2006; Mirhosseini & Abousaedi, 2023). Given the lack of significant findings, further analysis of the end- of-semester survey or instructor observations is not necessary.

However, the lack of significant differences may be partly due to the limited scope of the MBTI data. The small sample size (i.e., 58 extroverts and 23 introverts) likely influenced the results. As noted by Zhou (2015), small sample sizes are common in DB research and can inflate or obscure findings, especially when less rigorous methods are employed (Reyna & Meier, 2018). Additionally, the study’s focus on the extroversion-introversion preference pair alone may have overlooked other aspects of personality that could influence learning outcomes. Therefore, a more comprehensive analysis, with a larger number of participants per group or the inclusion of additional preference pairs, could provide a clearer understanding of personality’s impact on academic achievement, self-evaluations, peer evaluations, collaboration, and satisfaction.