Professor Dunn is a chemistry teacher with 120 students in three sections of her class. Her students are all reading the same materials, working on the same projects, and yet, after the midterm, she sees that the results across the courses are different. Why is one section having a problem when the other two are not?
Then she receives several emails from her college’s learning analytics software concerning several students in the underperforming class. The emails were generated by a computer program that looks at what is happening in the school’s online interface with student data from other parts of the campus such as Admissions and Records, EOPS or the athletics department and uses that to recommend specific action or interventions to help the students. For example, one email recommends a time-management learning unit and tutorials, while for another student, the program recommends tutoring and that the students contact their academic advisor. Why the different strategies?
Professor Dunn’s classes are part of a learning analytics project that can help her students. What the learning analytics program “knew” that Professor Dunn did not know was that one student is fresh out of high school with little direction but is on the basketball team, while the other student is returning to college after being in the workplace for a few years and is a single parent of two young children.
The learning analytics program keeps track of what her students are doing and why. It is a computer program that analyses data for correlations that most instructors don’t have the time or access to the necessary information to perform for each student. The software tracks who her students are, how they did in the past, where they are now, how much time they spend with the materials she posts online and how they do on her tests. The program will help her predict what kind of help her students will need to succeed.
Again, the concept of analytics is nothing new; researchers and academics have been writing about analytics since the 1970s and learning analytics as we now think of it since at least 2004. This is a concept not just for businesses but for many traditional colleges as well. For instance, Purdue University uses a system called “Course Signals” which can be seen at www.itap.purdue.edu/learning/tools/signals/. Rio Salado College in Arizona is using learning analytics as well with its “PACE” program.
Are there problems with learning analytics? Why do some consider it controversial? Some consider any kind of analytics to be an invasion of privacy. Others think that because businesses use analytics to sell T-shirts on Facebook that the only application for analytics is business. A few have written that the teachers should be making the decisions and not a computer.
A lot of care is taken with learning analytics software to encrypt student identities. The information that the program accesses is already being gathered by the schools and other agencies; learning analytics just uses that information to help faculty make decisions about how to help their students. And no computer program is perfect; no computer program will remove the need for critical thinking skills on the part of the administrators, teachers, or students.
The business sector has been using analytics for many years to make decisions. They have been very successful at it. If Amazon, Facebook and other online services can provide a custom commercial environment, why can’t education create a customized learning environment for the students? This is not a trivialization or even a commercialization of education. Business is not the only field that uses analytics. The field of epidemiology, the study of how disease spreads, has been on the forefront of analytics for years. They specifically track the relationship between human behavior and disease.
Analytics is basically making decisions using data. We do that now. The real question is, how do we get out of the way of our data and let it tell its story? William James said that “the art of being wise is the art of knowing what to overlook.” Human beings get hung up on details; computers do not. Humans also have the unfortunate habit of combining perfectly usable data with our opinions, misconceptions and prejudices, everything that is otherwise known as “common sense.” Machines are not making decisions for us. We are using them to gather relevant data and to show how the data interrelates for the benefit of our students.
The goal of learning analytics is to enable teachers and schools to identify educational opportunities for each student’s level of need and ability. More than just a buzz word, learning analytics promises to harness the power of advances in data mining, interpretation and modeling to tailor education to individual students more effectively.
- 7 Things You Should Know About First-Generation Learning Analytics (hollymccracken.wordpress.com)
- On the reductionism of analytics in education (annezelenka.com)
- 1st International Workshop on Learning Analytics and Linked Data (LALD2012) (linkededucation.wordpress.com)