Title: Why? A Case for Learning Analytics

Author Name: Stephanie McCaster

1. Introduction

Research shows that nearly 2.5 quintillion bytes of data is created each day. Big data has an enormous impact on our everyday lives and careers in surmountable ways. While analytics is not a new concept; it is currently in its infant stage as it relates to education. We have become keen to our search engines and subscription based services like Amazon, Spotify, and Hulu examine selection patterns to make eerily accurate future recommendations. All of the aforementioned platforms operations are largely based on business intelligence or analytics. Operating systems that combine business analytics, data mining and visualization using data synthesizing tools and infrastructure to help drive data driven decisions. Learning analytics aims to perform in a similar manner but under an educational setting. Defined as “the measurement, collection, analysis and reporting of data about learners and their contexts, for the purpose of understanding and optimizing learning and the environment in which it occurs.” (Long & Siemens, 2011) Learning analytics has several goals:  we make data driven decisions about education using quantitative methods drawn from big data sets. We use  data to better understand and inform teaching and learning for better outcomes for the student, instructor, and institution. 

Puget suggested we try to reflect on Learning Analytics as four separate categories or maturity levels. By doing so we gain a greater understanding and respect for the methodology.  (1) Descriptive analytics: describes learning behaviors seeking to uncover important data points. Answering, What does the learning data tell us is happening? (2) Diagnostic analytics: decodes learning results and trends. Answers the question, What does the learning data tell us about why a thing is happening? (3) Predictive analytics: predicts what will happen next, seeks to anticipate outcomes, What does the learning data tell us that is going to happen? (4) Prescriptive analytics: suggests specific solutions. Answers the question, What does the learning data tell us should be done?

In an education setting the designer should first set out to pinpoint the data that is needed (descriptive analytics) Patterns in student behavior that may correlate to positive or negative outcomes are examined while class is in session or immediately after (diagnostic analytics). Behavior patterns are then examined and evaluated to predict and/or validate outcomes for the next session (predictive analytics). Should specific patterns of student behavior be exhibited during the next session, instructors or administrator can help students course correct by recommending tailored suggestion  (prescriptive analytics)Dietz et al. (2018)  (p. 105)

2. Overview of the Case

Takeaways from case studies help us understand the impact of Learning Analytics has in the field of education.  One of the earliest documented learning analytics case studies is the Course Signals project at Purdue University. Course Signal is a student success system that permits instructors to offer feedback to learners based on predictive models. The algorithm consisted of four elements: (1) Performance: based on points earned in the course (2) Effort: level of involvement on the VLE compared to classmates (3) Prior academic perform: standardized test scores and GPA. (4) Student characteristics: age, social economic background and credit load. Each element is weighed based on an agreed upon level of importance filtered through the algorithm which then provides a traffic signal color. Red signifies a student has a high probability of not completing the course successfully. Yellow indicates the student is in the trouble zone. Green indicates a high probability of success. Clear-cut and concise feedback on the behalf of the instructor along with Signal appeared to have a greater impact. Signal deployment enhanced student achievement and retention resulting in a 14% reduction in D and F students.

Lessons from Case studies help us understand the impact Learning Analytics can have when implementing for student success.  One of the earliest documented learning analytics case studies is the Signals project at Purdue University. The algorithm consisted of four elements: (1) Performance: based on points earned in the course (2) Effort: level on involvement on the VLE compared to classmates (3) Prior academic perform: standardized test scores and GPA. (4) Student characteristics: age, social economic background and credit load. Each element is weighed based on an agreed upon level of importance filtered through the algorithm which then provides a traffic signal color. Red signifies a student has a high probability of not completing the course successfully. Yellow indicates the student is in the trouble zone. Green indicates a high probability of success. Clear-cut and concise feedback on the behalf of the instructor along with Signal appeared to have a greater impact. Signal deployment enhanced student achievement and retention resulting in a 14% reduction in D and F students.

3. Solutions Implemented

Learning analytics’ primary benefit is that the information allows for actionable insights where instructors are equipped with real time metrics related to key performance indicators in  student habits and are able to provide in the now resources to improve their learning experience. Resources prescribed for challenged students in the form of additional learning materials, peer to peer tutoring or time management training. 

4. Outcomes

Each element is weighed based on an agreed upon level of importance filtered through the algorithm which then provides a traffic signal color. Red signifies a student has a high probability of not completing the course successfully. Yellow indicates the student is in the trouble zone. Green indicates a high probability of success. Clear-cut and concise feedback on the behalf of the instructor along with Signal appeared to have a greater impact. Signal deployment enhanced student achievement and retention resulting in a 14% reduction in D and F students.

5. Implications

There are many benefits of using analytics as a tool for steering performance however many question if the benefits outweigh the challenges. The primary concern when it comes to challenges is privacy. The foremost challenge relates to privacy concerns. From the beginning stages of collecting data through shared data to further research efforts, each step of the process requires conscientious and careful consideration of learners’ privacy.

References

Reiser, R. A., & Dempsey, J. V. (2018). Trends and issues in instructional design and technology. New York, NY: Pearson Education.

https://www.researchgate.net/figure/Course-Signals-at-Purdue-showing-green-safe-and-yellow-lights-borderline-safe_fig4_336374701

Arnold, K. E. 2010. Signals: Applying academic analytics. EDUCAUSE Quarterly, 33, 1. www.educause.edu/library/EQM10110

https://er.educause.edu/articles/2011/9/penetrating-the-fog-analytics-in-learning-and-education

https://er.educause.edu/podcasts/educause-exchange/the-ethical-issues-around-learning-analytics

[Back to Home]