Using Big Data to Help Students Graduate

Joshua Lott for The New York Times

     Source Article: Big Data on Campus

   
     Several universities across our country are experimenting with big data mining in order to help students pick a major that best suits them while reducing the average amount of time it takes and increasing the likelihood for that student to graduate. According to the article, only 31% of students at public colleges graduate with their bachelor's degree within four years, and 56% graduate within six years.
     Some researchers at these colleges are currently trying to find a way using big data that can evaluate how well a student may do in a particular course and monitor a student's progress in their degree track. For example, some courses offered at public universities are web-based. Every click is logged by the software used to deliver the course material; these records can tell when a student has logged on, turned in homework, or looked at the syllabus. Adam Lange and his colleagues from Rio Salado College have been able to mine the data recorded by this software and determine by the eighth day of class whether or not a student would make a "C" or better with 70% accuracy! Mr. Lange has developed a software system that can identify certain similarities within the data and determine the characteristics of a poor preforming student. He believes that by identifying these students to a professor, the professor can reach out to these students and prevent them from dropping the class and falling behind on graduating.
     Austin Peay State has developed another algorithm that is similar to the "suggestions" algorithm used by Netflix. Before students register for classes, a "robot adviser" assesses their transcripts and is able to provide suggestions for courses that the students would be likely to succeed in. According to Tristan Denley, provost of Austin Peay State, when people are presented with many options and little information, they find it difficult to make wise choises. He further states what when students take courses that are recommended to them, they tend to do substantially better. By students keeping their GPA above the scholarship cutoff, Mr. Denley believes that it changes their likelihood of graduating.
    At Arizona State University, algorithms are used to determine course content. A system mines the performance data of thousands of students taking an online math course. This system is able to build a profile on each student and provide recommendations on what learning activity they should do next. Some students in the course found this method of material delivery more efficient and easier to understand than the traditional lecture format. The algorithm was much better at being able to tailor the learning material to each individual student- something that would be nearly impossible for a traditional instructor to do on their own.
     Not all universities are looking at just mining performance data- some colleges think that one way to predict if, when, and how a student will graduate is through social media. According to the article, research shows that social ties and academic success are directly related. Blackboard co-founder Matt Pittinsky thinks that if a university is able to model the social network of a college, the data layer could aid universities in reaching out to students that may not be showing evidence of social integration. By helping students find ways in which to socially connect on campus, students will be more likely to use these connections to encourage themselves academically.
     In this article, big data has been proven to be an important aid in helping universities determine at-risk students. By reaching out to these students, the retention rate increases, therefore increasing the probability that a student will graduate and shorten the length of time that a student graduates.