Satisfaction with Research-Based Learning and Academic Performance by Big Data Analysis ^†

Abstract

Data science analysis was used to investigate the association between Research-Based Learning (RBL) and student’s academic performance. The research professors certified in RBL participated in this research. The participants used an advanced platform with over 500 resources. Tutors guided students in designing, implementing, and evaluating the RBL. We employed a quantitative and quasi-experimental design without a control group, correlational analysis, and ex post facto elements. A satisfaction survey was administered after the capacitation. In total, 342 teachers were trained and received accreditation. The analyses incorporated non-conventional data science methods such as correlation analysis, principal component analysis, and unsupervised learning utilizing the Density-Based Spatial Clustering Application with Noise algorithm. The professors’ satisfaction with RBL was related to student’s academic performance, which was found in clustering analysis.

1. Introduction

The COVID-19 pandemic significantly changed university teaching and learning approaches. Videoconferences became the norm, the demand for information and communication technology (ICT) surged, and evaluation systems were redesigned. This transformation also increased the workload for teachers and students, highlighting deficiencies in ICT proficiency and resulting in feelings of isolation and disorientation [1]. The pandemic has exposed a substantial gap in educational institutions regarding infrastructure and teaching and research staff training. However, institutions have responded swiftly and creatively to these challenges. Educators have enhanced their training and shown greater interest in resources and strategies to improve teaching [2].

Inquiry-based learning emerges as a powerful teaching strategy, adaptable to various disciplines and educational objectives. College instructors employ a range of classifications or vocabularies to delineate investigative learning assignments from basic to advanced levels of achievement. This versatility underscores the effectiveness and adaptability of inquiry-based learning, instilling confidence in its application across diverse educational settings [3].

At the educational institution where this research was conducted, Challenge-Based Research [4,5] was promoted to accelerate solutions to societal issues. This methodology combines innovation with pedagogical practices. Additionally, faculty members integrate their teaching with research via Research-Based Learning (RBL), addressing the need for meaningful learning experiences and providing engaging teaching formats that captivate students’ interest [6]. This professional development initiative is implemented through a completely digital and utterly online model featuring a portal rich with over 500 high-quality instructional materials.

The courses or topics considered for RBL are taught in the different schools of the institution, including the School of Humanities and Education, School of Engineering and Sciences, School of Medicine and Health Sciences, School of Business, School of Architecture and Design, and School of Social Sciences and Government at a university in Mexico. The university’s high school system is integrated into this framework. Moreover, faculty members of different undergraduate and high school programs actively engage in RBL initiatives and are trained through an accredited RBL workshop.

In this study, we observed the satisfaction of professors who have undergone RBL accreditation. Taking the initiative to assess this process, we calculated and analyzed the correlation between the contentment levels of RBL-accredited educators and student’s academic progress. Then, the impact of RBL on teacher satisfaction and student learning outcomes was investigated with unsupervised learning.

2. Inquiry-Based Learning and Academic Achievement

It has been observed that researchers and professionals do not collaborate and lack synchronization [7]. This disconnect might dissipate student’s engagement in experiential learning through research. RBL approaches can potentially foster a favorable transformation. Educators undergo professional growth by participating in research and developmental endeavors. Research initiatives directly impact teaching quality, bolster student motivation and engagement, and shape the academic atmosphere concerning faculty research endeavors. Furthermore, when educational institutions prioritize research, students actively participate and cultivate interactions, perseverance, intellectual curiosity, critical thinking abilities, and other skills [8].

The study of learning through student accomplishment is becoming increasingly indispensable. Students actively engage in learning, employ meaningful tactics, attain enhanced academic outcomes, and showcase competencies and engagement in their assignments and academic endeavors. Several universities recognize the benefits and have integrated RBL into their educational programs and embedded it within the syllabus [9]. This approach offers an opportunity for myriad students to engage in research endeavors, especially in technology, science, engineering, and mathematics (CTIM), leading to a rich learning experience [10]. Moreover, the RBL methodology fosters cognitive and affective-emotional growth for research in alternative fields, like the social sciences [6].

In RBL, four distinct modes of student involvement in inquiry are distinguished. Students express enthusiasm for the research results or process, while others adopt active or passive roles. These modes are combined in four modalities to establish the link between teaching and research [9].

• Research-Guided Learning: students acquire knowledge from the findings of investigations.
• Research-led: students learn about research outcomes with the teacher determining the communication strategy.
• Research-oriented: students derive learning from the research process in this mode.
• Research-based: students assume the role of researchers, engaging directly in the research process.

Reference [11] defines RBL as a procedure where students regulate themselves, assisted by their research inquiries. Through the progression of research projects, students engage in cognitive, behavioral, and affective activities that demonstrate the advantages of conducting RBL [12]. In RBL [3], (1) scientific investigation is utilized in the teaching–learning process, (2) teaching centers on the learner, (3) learning is stimulated by raised inquiries and questions, (4) an instructor assumes the role of the facilitator, learning results from the construction of knowledge that promotes cognition and metacognition, and (5) self-direction is encouraged. Students who engage in RBL report enhanced learning experiences and increased academic attainments. Once students become enthralled with a research project that they can conduct independently, they never revert to their previous methods [12], but transform.

An understanding of learning through the RBL approach is enabled in an interpretative process that leads to understanding reality [13]. Learning is manifested through the capacity to debate and even bolster practical experiences. Practical learning entails solving problems, honing skills through encounters with real-world issues, and suggesting viable solutions that can be implemented with expert guidance. Learning comprehension is observed when the student assumes an active role. Engaging in a science project transcends the mere completion of another assignment for the student [12]. It represents an encounter involving self-discovery, wherein they take on, often for the first time, the responsibility for their self-learning. This demands intellectual maturity in conducting research and assuming accountability for the process. Moreover, as previously discussed, integrating teaching with research entails advantages for the institution.

Several studies notice the significance of RBL in influencing students’ academic achievements and enhancing their learning outcomes (

Table 1. Similar studies to this research.

Reference	Description	Method	Data Analysis	Findings
[14]	Assess the influence of research-oriented learning on students’ academic outcomes and evaluate their contentment with the experience.	The study utilized longitudinal data to determine if RBL improved learning, academic performance, and overall satisfaction.	Linear and logistic regression analysis	A positive correlation exists between academic performance and RBL. The authors concluded that research-based learning can enhance students’ engagement and knowledge acquisition.
[15]	It is a method for designing a research-oriented course. It enables faculty across various fields to involve students in innovative manners.	Qualitative	Qualitative	Research-based learning is a potent pedagogical approach for developing self-directed students. Successful completion of Inquiry 1SS3 correlates with learners achieving passing grades, maintaining positions on the dean’s honor roll, and persisting in their university studies.
[16]	To assess the impact of training digital stories through RBL on the inclination of prospective educators to embrace technology instruction	Experimental study utilizing a convergent parallel design.	Shapiro–Wilk test and t-test.	The study suggests that preparing digital stories through an inquiry-based learning approach significantly reduces forthcoming teachers’ resistance toward technology-based teaching and performing research.
[17]	The paper summarizes the creation and implementation of four integrated laboratory sessions at the School of Medicine, King Saud University.	It is based on promoting inquiry-based learning and motivating students to engage in collaborative tasks within small groups.	The chi-squared test, Mann–Whitney U, and Cronbach’s alpha.	The research also demonstrated that the revised design promoted self-guided learning, knowledge integration, and apprenticeship among students.

). It is crucial to carry out further research to identify the factors that enhance the learning patterns and research activities of both teachers and students. This increases their performance, skill development, and overall learning. It is beneficial to receive feedback from some of their teachers [8].

3. Method

After completing the training program, a satisfaction survey was administered to a non-random convenience sample of professors to evaluate their contentment level. A focus group examined the survey to ensure its quality, assessing the content and structure of each question. We employed a quantitative, quasi-experimental design without a control group and ex post facto aspects. Correlational analysis, principal components, and the Density-Based Spatial Clustering Application with Noise (DBSCAN) algorithm for clustering were utilized for analysis.

3.1. Intervention

Educators were certified through the RBL workshop using the digital plus model (100% online format) in response to the confinement restrictions imposed by the COVID-19 pandemic. The state-of-the-art platform provided instructors with more than 500 resources, including complimentary courses, library tutorials, templates, readings, websites, and research hubs. The professors devised an RBL strategy based on the approaches outlined in Ref. [9]. Mentors provided feedback to teachers who utilized the platform featuring RBL resources as a tool. They conceived, formulated, executed, and evaluated an RBL strategy for their designated classes within one semester. The educators were assessed using rubrics and student evidence. The students acquired and honed competencies through research and presentation of their findings. At the culmination of the intervention, a participant satisfaction survey was administered, and the educators received acknowledgment in their profiles. The trained educators facilitated their students’ learning through research endeavors.

3.2. Sample

The sample consisted of 342 educators, with 41% being female and 59% male. Of the instructors, 18% are affiliated with SNI, the national research system in Mexico. In terms of age, 32% were aged between 25 and 40, 31% between 41 and 50, 28% between 51 and 60, and 8% were over 60. Further, 42% taught in the School of Engineering and Sciences, followed by the Business School at 18%; the School of Social Sciences and Government at 10%; the School of Humanities and Education at 9%; the School of Architecture and Design at 7%; and the remainder included preparatory schools (high schools), the School of Medicine, and administrative sectors.

3.3. Data Analysis

The responses were coded for descriptive analysis, and dummy variables were created for the categorical variables for the school and campus. Subsequently, correlation analysis was conducted to explore the relationships between these variables. To delve into the structure of the dataset, dimensionality reduction was conducted using principal component analysis (PCA) [18]. This statistical technique was used to extract essential information from the dataset by creating new features named principal components, a new collection of orthogonal variables [19]. Dimensionality is successfully reduced when a small number of components explain a significant portion of the total variance, enabling the identification of patterns, trends, and outliers [20].

The Python library sklearn was used in PCA. Notably, the two main components elucidated 74% of the data’s variability, with the first accounting for 61% and the second for 13%. Consequently, these components were plotted to visualize potential patterns or clusters. Irregular clustering patterns were observed, prompting further investigation through unsupervised analysis of a non-convex clustering structure. The DBSCAN algorithm was employed to identify clusters as high-density regions separated by low-density areas [21].

The DBSCAN procedure was introduced in Ref. [22]. This method operates with just two parameters: the neighborhood size (eps) and the minimum number of points required to constitute a cluster (minPts). DBSCAN identifies clusters by seeking regions where points are densely packed, with sparse spaces between them [23]. While the algorithm employs a Euclidean algorithm to compute similarity, alternative metrics can also be utilized. DBSCAN offers benefits, including the ability to detect groups of any shape, straightforward parameter interpretation, and efficient performance even with large datasets [22].

4. Results and Discussions

4.1. Descriptive Analysis

In total, 95% of the respondents expressed satisfaction with implementing and utilizing RBL, rating themselves as either satisfied or extremely satisfied. Additionally, 5% scored themselves as somewhat satisfied, indicating the significance of using RBL. When RBL was implemented, teachers trained in RBL under the institution’s educational model reported that the following six activities were frequently employed by students. First, efficient search strategies, planning, and specialized database management accounted for 41%. Second, the analysis of scientific articles represented 37%. Theoretical framework elaboration, elaboration/execution of projects, teamwork, and bibliographical analysis accounted for 22–24%.

In total, 98% of the teachers believed that the Challenge-Based Learning (CBL Model) must be linked with RBL to achieve greater rigor, quality, performance in learning, and student aptitude development. Further, 92% believed that the RBL strategy is necessary for the educational model of the institution where the research was developed, particularly within the CBL strategy. Teachers also recommended that the learning units designed for the educational model must integrate CBL-RBL.

The professors used RBL in classes of varying sizes, from 2 to 245 students, with 30 students being the most common (8%). Regarding academic performance, 56% of the respondents stated that their grades improved when RBL was used, 34% reported that an increase did not apply because of how the courses are usually evaluated, and 10% reported no improvement.

Table 2. Percentage growth with RBL implementation.

Increase in Academic Performance		Percentage
Yes	1–5%	12%
	6–10%	17%
	11–20%	14%
	>20%	13%
NA		34%
No		10%

summarizes the percentage increase distribution in academic performance when RBL was implemented. Scopus was the most used database among teachers, with 43.7% implementing it in RBL. Web of Science and Elsevier followed, with 28 and 27.6% usage.

4.2. Correlation Analysis

The assessment of the RBL certification course and instructors’ contentment in executing and employing the RBL framework showed a correlation coefficient of 0.38. The rise in academic performance percentage associated with RBL utilization and the satisfaction level in implementing and utilizing the RBL framework showed a correlation coefficient of 0.35. Instructors’ satisfaction with the RBL model was positively correlated with evaluating the RBL certification and students’ academic achievement. Therefore, prioritizing instructor contentment with pedagogical approaches is crucial, as it directly impacts student outcomes.

4.3. Unsupervised Learning

In the PCA, the clustering of the two principal components was irregular. Consequently, the DBSCAN algorithm was employed, utilizing the Scikit-Cluster function, with Euclidean distance utilized for the distance matrix. The determination of the neighborhood size and cluster count was based on the plot of the two principal components, with eps established at 0.0003 and minPts set at five. Figure 1 illustrates the outcomes of the clustering analysis.

The DBSCAN algorithm was used to effectively group the irregular clusters.

Table 3. Clustering statistics.

Group		1	2	3	4	5
Cluster Size		11%	45%	17%	14%	13%
Female (1)		42%	43%	32%	30%	46%
How satisfied are you with implementing and using the RBL model?	Very good	60%	50%	70%	77%	72%
	Good	37%	38%	28%	23%	26%
	Regular	3%	11%	2%		2%
Would you consider it essential to integrate the CBL framework with the RBL approach to enhance the rigor, quality, and effectiveness of learning, as well as the development of student competencies?	Extremely satisfied	47%	21%	36%	64%	70%
	Very satisfied	53%	72%	64%	36%	30%
	Somewhat satisfied		7%
Did the academic performance increase with the RBL accreditation?	Yes	100%	5%	100%	100%	100%
	No		23%
	NA		72%
How much was the percentage increase?	NA	0%	100%	0%	0%	0%
	(1–5%)	100%	0%	0%	0%	0%
	(6–10%)	0%	0%	100%	0%	0%
	(11–20%)	0%	0%	0%	100%	0%
	>20%	0%	0%	0%	0%	100%

presents several noteworthy details for each group. In Figure 1, the red cluster (Cluster 1) represents teachers who have witnessed a 1–5% increase in their students’ academic performance. The largest group, identified as the blue cluster (Cluster 2), exhibits no notable increase in academic performance. Cluster 3 (green) pertains to teachers who have observed a performance increase between 6 and 10%, while Cluster 4 (purple) denotes an increase between 11 and 20%. Lastly, Cluster 5 (pink) comprises cases where the increase exceeded 20%.

No significant distinctions were detected among the clusters based on age groups, gender, SNI membership, campus, or school. Clusters 4 and 5 exhibited the most notable enhancements in academic performance, which indicated higher levels of satisfaction with the implementation of the RBL model with rates of 77 and 72%, compared to Cluster 2, which reported only 50% being extremely satisfied with RBL implementation. Conversely, Cluster 5 demonstrated the most substantial increase in academic performance and displayed predominant utilization of databases for RBL implementation (80%).

5. Conclusions

In this research, the positive impact of implementing RBL on academic performance was determined. It reaffirmed the effectiveness of the unsupervised learning algorithm DBSCAN in identifying irregular clustering patterns within educational contexts, underscored by its reliability and applicability in educational research. The results highlighted the role of teacher satisfaction in RBL implementation, which directly influenced students’ academic outcomes. RBL enabled a wide range of successful activities, including strategic planning, efficient database management, article analysis, theoretical framework development, research project execution, bibliographic analysis, and collaborative teamwork. This versatility of RBL underscores its potential to enhance various aspects of education. Thus, fostering teacher satisfaction with RBL implementation is essential in fostering students’ academic success.