Faculty Interest Group “Using Data to Support Teaching and Learning”
Notes from Meeting 2 (Nov 5, 2018 11:30am -12:30pm)
During this meeting we continued our discussion on data collection, analysis, interpretation and the importance of data-informed decisions for students, faculty, and institution. We discussed the article of Nadasen and List, “Using Community College Prior Academic Performance to Predict Re-enrollment at a Four-Year Online University” and the following items.
- How data analysis helped in estimating the impact that first semester grade has on persistence and graduation?
This study examines student learning characteristics and behavior at a community college of 8200 students who transferred to an online a four-year college and their predictability in “persistence” – meaning re-enrollment in the second semester of a four-year college. This early persistence is considered a strong predictor of graduation and is used as a dependent variable. Data was dis-aggregated into: 1) learners characteristics (age, gender, race, and marital status), 2) course taking features, 3) course efficiency (credits earned/credits attempted), and 4) first-term GPA.
- Using Data for Detection: We discussed possible ways that Faculty and Administrators can use to detect student failure before it occurs as well as how to learn from successes of other programs.
Items that were not discussed yet and could in the future are:
– How can we apply to education the use of Big Data in private sector?
o Can we use similar techniques to have early detection of student potential failure similarly to fraud analysis/prevention system that banking system uses?
- Fraud Analysis/Prevention. Data is collected and analyzed. Estimates have provided a predictive model. Use of prediction to prescribe an action: Transactions that are flagged can then be investigated further before payments are made. Banks use big data to flag potential fraud in credit card use.
o Can we find complement courses or optimal sequence of courses depending on what students have taken before and their performance and satisfaction similarly to “market basket analysis”?
- Shopping Cart and “Market Basket” Analysis – Many companies realized that data can show retailers shopping patterns of consumers in their industry. If Amazon knows that customers who purchase a specific item also tend to purchase another item, then it can present that second item when the first is bought. — Could such advising be automated or use artificial intelligence (AI) advising?
– Review the figure to understand how different end users of data can work together.
Attachments
This entry is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license.
