For universities, retention rates are a key indicator of the school’s ability to meet the needs and expectations of its students. Having a low retention reflects poorly on the school and makes it less likely for incoming students to apply. In 2014, Crown College in Minnesota updated their retention strategy using a new approach they called Persistence and Completion combined with data analytics techniques from Jenzabar Technology. According to the article, they compiled and mined data from “first-time, full-time degree-seeking students who entered the college in fall semesters of 2009 – 2014” from which they created a regression model to predict the likelihood of retention for at-risk students. It first ran this model in fall 2015 against incoming full-time freshmen in its School of Arts and Sciences and continued to use the model for the next six semesters. After implementing the program, freshman retention rose from 84 percent in 2015 to 89 percent of freshmen in 2019, and retention of all eligible students rose from 90 to 94 percent for the same years.
At face value, I like this initiative for a number of reasons. Though many schools try not to appear like a business, most operate like one. From a business perspective, an initiative like this can help universities stay competitive or even exceed their competition in retention rates. I also thought that being able to accurately forecast the chance of at-risk students unenrolling enables schools to give those students targeted support rather than either detecting the risk too late or simply waiting for the student to ask for help.
Despite my satisfaction with Crown College’s success story, the article left out many key details about their analytical process. First, they mentioned the school identified nine factors that could predict a student’s risk of unenrolling, but they didn’t say what any of these factors were. I would like to know these factors to gain some insights into what the school was evaluating and how it could yield such accurate predictions. Similarly, the article mentioned that the most successful implementations of this retention plan at other universities has been evaluating students at the beginning of the term. Thus, I wondered how accurate the model is across different grade levels since a freshman would presumably have much less data to factor than a rising senior. Furthermore, since retention is usually a higher concern for underclassmen, this makes we want to know even more what metrics are being used to evaluate risk during the earliest months of a student’s college career. I would venture that it is more difficult to support freshman as their retention rates were lower than that of all eligible students both before and after the Persistence and Completion program.
Overall, I enjoyed this article and was impressed by the use of data to support students in need. Additionally, with some additional research I found that Crown College was able to put its retention rates on par with more selective colleges who average at or below 25% acceptance.
“The Condition of Education - Undergraduate Retention and Graduation Rates - Indicator April (2020).” The Condition of Education, National Center for Education Statistics, Apr. 2020, nces.ed.gov/programs/coe/indicator_ctr.asp.
Zaino, Jennifer. “Case Study: Crown College Uses Predictive Analytics to Retain At-Risk Students.” Dataversity, Dataversity, 4 Feb. 2021, www.dataversity.net/case-study-crown-college-uses-predictive-analytics-to-retain-at-risk-students/.
I agree with you in that universities are better off conducting data research to find reasons why students are either transferring or dropping out of college. This is especially important for universities to understand why certain demographics may not have a high retention rate. These studies can help universities change a series of their aspects in an effort to keep students. I would like to assume that some of the aspects the analytical factors were based off of might have been race, nationality, household income, sex, major/minor, and financial aid. I also agree that universities should try to implement more strategies that further their research into finding out why a student is more likely to drop out. Loyola has a particularly high unenrollment rate. Many students who transfer happen to be students of color and first generation American students. This for example, would be useful for the institution to know and utilize as a way to enhance and prevent this. Loyola and other universities experiencing high dis-enrollment can utilize this data to create more resources that prevent marginalized students or students with other issues from transferring or dropping out. If the data were to show that first-generation students are struggling to stay in university, the university can create programs to help those students feel seen and heard. If a particular major is experiencing high dis-enrollment, perhaps altering the course work would be beneficial. I too am greatly interested in seeing what Crown’s College based their nine factors off of. The article should have stated what these nine factors were based off of so that other universities could benefit from the same research process. These factors may also be picked by a case to case scenario.
ReplyDeleteHi Eric, I really enjoyed the article you discussed and your analysis of it. I think that the college industry is quite vast and could benefit from knowing their retention rate and factors that lead to this. Although I hate to admit it, colleges are really just business and to run a good business you would ideally want all of your students to stay the entire time to ensure the most money from them. Like you mentioned, a low retention rate indicates something within the system is wrong. Whether it is the environment or the teachers, it doesn’t reflect well on the school itself. This makes it very clear as to why colleges would want and how they would use this information.
ReplyDeleteUsing data to create a model predicting likely retention is a great idea on the colleges’ behalf. The article states that the software also helps identify factors that potentially identify students that are more at risk leaving. I think this is especially important because it make the best use of the resources given. Not even student who goes to a school is thinking of transferring, so identifying those who are more likely to leave can help funnel the resources and effort into persuading students who are thinking of leaving to stay.
Much like you, I felt myself wanting to know more about the software and what factors they use to evaluate students. I think it is good that they use nine factors in order to make this distinction, because one or two factors does not really give an accurate picture and could give misleading information. Additionally, I would like to know how timely and accurate this data can be expressed to people in order to get the process moving. While knowing past data can help with future incoming classes, this data is timely for students considering transferring. It would be most helpful to know during the first few weeks a student is at the school, since they are most likely open to either staying or transferring. Trying to persuade someone at the end of the semester when they have already applied to other schools to transfer is wasting effort.
Overall, I thought this article was very interesting and had a lot of potential. I think this software and data collection would be extremely useful. Not only does having a strong retention make the school look better, but it also makes students want to go there more, thus boosting the college’s public image. It was also nice to see the Crown College was being rewarded and their efforts paying off. The college industry is quite big and I can see this being implemented at many schools in the future.
I can relate to this article because I myself am a transfer student. I attended the University of Maryland for three semesters and I did not have the experience that I expected. The idea of gathering information about who will stay at a college and who will transfer is beneficial to the university and the student. Colleges can use this data to tell prospective students what type of person finds the most success and happiness in that environment. This can give students a better idea of how they would fit depending on their personality and academic path they wish to pursue. I can imagine that this was a very difficult task to take on given the large number of varying circumstances and preferences among students.
ReplyDeleteAlthough the article does not mention what the are, I assume that some of factors they looked for were high school, hometown, involvement in the school community, study habits, number of students they know going into the school, finances, and other characteristics of each student. Study habits can be a major indicator based on the difficulty of the school. They could even look into how people with different study habits tend to do in each major. One thing I learned from Maryland and now being at Loyola is that the difficulty of a class has a lot to do with the professors. One section of a course may be much more difficult than another section of the same course because of how the professor runs the class. If the professors in one major tend to be more difficult than other majors, a student’s major and study habits could be major factor on if they stay at that school. Your hometown and background could also play a major role because students may not want to go to a place whose culture is very different from what they are used to. Involvement and entering the school knowing other students has a major impact on how well you fit in because it determines your friend group. If you do not know anyone and do not try to get involved than you will not make any friends. Obviously, finances can be a concern for many people. If someone loses a job and has financial concerns then they may need to change to a cheaper school.
Another possible reason for leaving is sports. Student athletes may attend a school with little chance of getting any playing time. As they have more practices and realize that their chance of playing are getting smaller then they will look to go to other schools. Athletics could also have an impact on non-athletes. Someone may go to a small school thinking that they do not care about having a big sports program then later realize that it is something that is important to them.
I would like to know what other factors they used and what was the most important factor. Overall, this shows the power and importance of data.