Learning Analytics as a Support Tool for Higher Education

Learning analytics allow us to make insights about the student experience from the data. With this, we can identify new variables that help institutions to better respond to the situations that students experience and make better decisions based on the analysis of the information.

Learning Analytics as a Support Tool for Higher Education
Reading time 8 minutes

What do universities do with all the information they keep on their students? The bachelor’s academic record only includes information related to university admission, professional and graduate studies, and the first years of post-grad. This data is recorded and stored for four to six years on average, but what does all this information tell us? How can we leverage it to help students achieve their vocational training goals? In this article, I share the potential of learning analytics to transform higher education and some examples of real applications and innovative ideas that help us visualize the scope of this area of study. Technological advances such as the computing power to manage large volumes of information (big data), artificial intelligence, machine learning algorithms, and data mining make possible the path educational institutions must follow to test new operations, evaluation, and administration models.

Learning analytics is not a new field of study; its documented research dates from 1979 when The Open University in the United Kingdom analyzed ten years of academic history of its distance students, observing their progress in each course and the different stages of their school life (Ferguson, 2012). According to the Society for Learning Analytics Research (SoLAR), “learning analytics” is the measurement, collection, analysis, and contextual data reporting about learners to understand and optimize learning and its environments. (SoLAR, 2023).

Application of learning analytics in higher education

Below, I share evidence of the usefulness of learning analytics in higher education, which emerges from the report, Research Evidence on the Use of Learning Analytics: Implications for Education Policy, published as joint research among members of the European Union (European Commission, 2018).

Georgia State University in the United States used predictive analytics to monitor student performance, and the data also predicted the dropout potential of some of its most outstanding students due to low-income issues or students who represented the first generation of their families to attend college. With this information, they designed interventions that increased their graduation rates from 32% in 2003 to 54% in 2014 (European Commission, 2018; p. 82).

In 2010, Rio Salado College in Arizona, the United States, implemented a predictive analytics system to monitor the virtual environments of up to 40,000 students. This monitoring included an early warning system for teachers that allowed them to better identify and serve those students at risk of failing a course (European Commission, 2018, p. 77).

In a further application of learning analytics in 2011, Arizona State University carried out a project with Knewton Enterprises to create personalized learning paths for more than 5000 students in online remedial mathematics courses. Adopting this system increased student retention in the remedial mathematics program from 64% to 75% (European Commission, 2018, p. 76).

Learning Analytics at Tecnologico de Monterrey

Research on the potential of learning analytics is paramount at Tecnologico de Monterrey. Various initiatives have been developed to generate artificial intelligence models for different audiences, such as students, teachers, and managers. For example, Hernández-de-Menéndez et al. (2022) described an institutional case where a program supported at-risk students by mining educational performance and behavioral data. They analyzed the correlations of different data such as academic performance (e.g., present and past grades), behaviors (e.g., class attendance), trends (e.g., current and accumulated average), and the experiences of the Academic Improvement Office at Tecnologico de Monterrey (Tec), which implemented a support program for students at risk of dropping out of school, reducing the dropout rate by 15% and the percentage of students who decided to change careers by 20%, and finally, identifying real solutions for 30% of at-risk students.

Another example focusing on learning analytics at Tec began with a previous step: understanding how data is identified, validated, and linked to educational research. To carry this out, initiatives like the Data Hub of the Institute for the Future of Education provide access to institutional data collections to promote educational research and provide a space for researchers to deposit and acquire data from their studies. Thus, the Data Hub issues call for research such as the Call to Address School Desertion, which promoted building a dataset on the topic, and various academic publications (Alvaro-Uribe et al., 2022; Talamás-Carvajal & Ceballos, 2023).

However, what benefits do learning analytics offer a higher education institution? Tecnologico de Monterrey recognized the magnitude of managing educational information over time; their effort resulted in 1) the documentation of the data available for research by information experts, describing its meaning and scope, 2) linking commonly used data such as academic history with other more innovative data, for example, lifelong learning data and degree follow-up surveys, all of which 3) provides a unique space for researchers to obtain curated, validated, and anonymized information, the provenance of data institutionally reviewed.

Notably, to date, the application of learning analytics in Latin America, and the researchers who study them, are a limited number in the field of education (Hilliger et al., 2020), making it challenging to leverage the repertoire of historical information on student’s experiences throughout their academic lives; in most cases the trajectories of their professional lives are unknown. If teachers and administrators knew their students’ entire history, they could develop systems that personalize their learning according to their academic and life experiences, identifying strengths and weaknesses in the development of competencies (as Khan Academy has done for a STEM curriculum), providing recommendations aligned with factors such as personality, gender, study habits, interests, and the observed learning of generations of students over time. Finally, the reflection we need is simple: How does our data-driven strategy allow us to identify the difficulties in acquiring knowledge or skills 20, 10, or 5 years ago that no longer occur?

If we do not have the answer (as the saying about opportunity goes), the second-best time to work on a new educational architecture is now.

The information that we do not see with the naked eye

Considering the amount of data a higher education institution generates over multiple generations, we can understand the tremendous opportunity presented. In 2022 alone, 668,000 students graduated with a bachelor’s degree in Mexico (ANUIES, 2022), each with a data history of approximately four to six years. This sea of information can tell us many stories, for example, 1) those students who graduated and those who did not, 2) the dissatisfaction of students who wanted a more practical experience for work or a more humane service-oriented approach; likewise, we might discover 3) the history of those who did not know they could access some financial support provided by the university or the government,  and others 4) who could receive social support from academic advisors if a sudden drop in their attendance or scholastic averages were identified.

Learning analytics can make it easier for us to make findings from data, identify new variables that could help the institution respond better to different situations that students experience, and make decisions based on information analysis. As leveraging educational data through learning analytics advances, educational institutions can increase their applications each year and possibly adapt structural changes, including data modeling and visualization, machine learning, and the widespread use of artificial intelligence.

Examples of solutions with learning analytics

As in other research fields, learning analytics promotes computational methods such as machine learning to automate artificial intelligence models that help analyze each student’s contextual situation considering different variables. Thus, compared to traditional data analysis, learning analytics research requires understanding the functioning of a data infrastructure (its collection, storage, analysis, visualization, recommendations, and feedback) and an ethical and legal understanding of the care and proper use of data and its implications. Significantly, frameworks such as PERLA (Personalized Learning Analytics) can facilitate a systematic approach to learning analytics (Chatti & Muslim, 2019).

As a reference, we can take the existing applications in different educational institutions worldwide. This information helps us visualize and understand the value and scope of data in our universities.

Feedback via digital boards (dashboards)

In a regular educational context, a student’s grades are the primary form of feedback on performance, but this leaves out valuable information generated over time. Let us imagine using artificial intelligence within a platform to understand their training status. For example, a student could access information for the current semester to see a calendar of assignments, their performance compared to other students, and a list of supports recommended by the institution according to their preferences (e.g., exchanges, scholarships, internship opportunities, job opportunities). Regarding digital dashboards, Roberts et al. (2017) have documented students’ preferences for dashboard choices in facilities that are customized to their needs.

On the other hand, institutions can also benefit from a digital dashboard to identify students at risk of dropping out, disaggregated by categories (faculty, academic program, monitoring status) and levels (international, national, state, local) as well as a record of the actions taken to follow up on these students.

Referral systems for education

Usually known for their application in digital stores or video on demand (streaming), recommendation systems weigh characteristics of user behavior to provide recommendations for different purposes. In an educational scenario, recommendation systems can consider students’ past experiences to suggest reinforcement content such as readings, activities, and data collection of students’ experiences and then integrate them into a data model, analyze, and evaluate them. Suppose the recommendation is appropriate or inappropriate per different hypotheses (e.g., it is appropriate for the academic program but not for the student). In that case, institutions can make the necessary adjustments in the model or request the intervention of a teacher.

Advanced personalized tutors

In another project with the potential to develop learning analytics, Tec de Monterrey developed a conversational agent (chatbot) called TecBot, which was especially useful during the pandemic period (2020 – 2023) to facilitate support for students’ administrative needs. As an additional step due to the growing popularity and accessibility of generative artificial intelligence, the institution will look for ways to use this technology to prepare advanced personalized tutors. Beyond helping to resolve administrative issues, these tutors can accompany teachers and students as assistants in possible activities or personalized practice exercises that reinforce learning with the potential to recognize emotional factors and maintain motivation to learn when there is a flux between the student’s ability and the level of challenge.

Reflection

As briefly described in this article, the possibilities of personalizing learning analytics in higher education are vast. However, it requires involvement and understanding of different information technology and education areas. It is vital to remember that more data is different from better data. For relevant learning analytics, we must consider privacy, ethics, and people’s rights regarding the use of their data. In this sense, integrating learning analytics in higher education opens new ethical and security discussions about using and protecting the data generated for the benefit of the community as a guiding axis, a topic that the Institute for the Future of Education at Tec de Monterrey also explores.

Would you like to know more or collaborate on a learning analytics initiative? Participate in the Living Lab & Data Hub of the Institute for the Future of Education at Tec de Monterrey. Each year, we carry out activities promoting educational data, such as calls for research based on data, seminars open to the general public, and exclusive workshops for the institution’s professors and researchers. For any concerns or questions, contact me by email at the end of the writing, and I will gladly guide you in becoming involved in this fascinating world of learning analytics.


About the author

Gerardo Castañeda Garza (g.castaneda@tec.mx) holds a Ph.D. in Educational Innovation from Tecnologico de Monterrey. He serves as Data Acquisition Coordinator at the Living Lab & Data Hub atthe Institute for the Future of Education at Tecnologico de Monterrey. His line of research revolves around interdisciplinarity in education issues, enjoying a generalist profile that addresses and combines multiple areas of knowledge.

References

Alvarado-Uribe, J., Mejía-Almada, P., Masetto Herrera, A. L., Molontay, R., Hilliger, I., Hegde, V., Montemayor Gallegos, J. E., et al. (2022). Student Dataset from Tecnologico de Monterrey in Mexico to Predict Dropout in Higher Education. Data, 7(9), 119. MDPI AG. Retrieved from http://dx.doi.org/10.3390/data7090119

Asociación Nacional de Universidades e Instituciones de Educación Superior (ANUIES). (2022). Anuarios Estadísticos de Educación Superior (Ciclo Escolar 2021 – 2022). Anuarios Estadísticos de Educación Superior. http://www.anuies.mx/informacion-y-servicios/informacion-estadistica-de-educacion-superior/anuario-estadistico-de-educacion-superior

Bloom, B. S. (1984). The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as One-to-One Tutoring. Educational Researcher, 13(6), 4–16. https://doi.org/10.2307/1175554

Chatti, M. A., & Muslim, A. (2019). The PERLA Framework: Blending Personalization and Learning Analytics. The International Review of Research in Open and Distributed Learning, 20(1). https://doi.org/10.19173/irrodl.v20i1.3936

Ferguson, R. (2012). Learning analytics: Drivers, developments, and challenges. In the International Journal of Technology Enhanced Learning (Vol. 4, Issues 5–6, pp. 304–317). Inderscience Publishers. https://doi.org/10.1504/IJTEL.2012.051816

European Commission, Joint Research Centre, Hillaire, G., Ferguson, R., Rienties, B. (2018). Research evidence on the use of learning analytics – Implications for Education Policy, (R. Vuorikari, editor, J. Castaño Muñoz, editó) Publications Office. https://data.europa.eu/doi/10.2791/955210

Hernández-de-Menéndez, M., Morales-Menendez, R., Escobar, C. A., & Ramírez Mendoza, R. A. (2022). Learning analytics: state of the art. International Journal on Interactive Design and Manufacturing, 16(3), 1209–1230. https://doi.org/10.1007/s12008-022-00930-0

Hilliger, I., Ortiz-Rojas, M., Pesántez-Cabrera, P., Scheihing, E., Tsai, Y. S., Muñoz-Merino, P. J., Broos, T., Whitelock-Wainwright, A., Gašević, D., & Pérez-Sanagustín, M. (2020). Towards learning analytics adoption: A mixed methods study of data-related practices and policies in Latin American universities. British Journal of Educational Technology, 51(4), 915–937. https://doi.org/10.1111/bjet.12933

Roberts, L. D., Howell, J. A., & Seaman, K. (2017). Give Me a Customizable Dashboard: Personalized Learning Analytics Dashboards in Higher Education. Technology, Knowledge and Learning, 22(3), 317–333. https://doi.org/10.1007/s10758-017-9316-1

Society for Learning Analytics Research (SoLAR). (2023). What is Learning Analytics? Society for Learning Analytics Research. https://www.solaresearch.org/about/what-is-learning-analytics/

Talamás-Carvajal, J.A., Ceballos, H.G. A stacking ensemble machine learning method for early identification of students at risk of dropout. Educ Inf Technol (2023). https://doi.org/10.1007/s10639-023-11682-z


Editing

Edited by Rubí Román (rubi.roman@tec.mx) – Editor of the Edu bits articles and the Webinars of the Observatory (“Inspiring learning”) Observatory of the Institute for the Future of Education, Tecnologico de Monterrey.


Review Committee

Ph.D. Leonardo Glasserman. Director of the Master’s program in Educational Entrepreneurship and associate research professor at the School of Humanities and Education of the Tecnológico de Monterrey.

Reviewed: 21 June 2023
Sent: 21 March 2023
Publication date: 27 June 2023

Translation by Daniel Wetta

Gerardo Castañeda Garza

This article from Observatory of the Institute for the Future of Education may be shared under the terms of the license CC BY-NC-SA 4.0