5 Academic Papers that Show How to Combat Attrition at Your School with Data and Analytics
Retention is critical for colleges and universities. It’s the key to better student outcomes and a solid revenue base to keep you operational. But often times, it’s neglected.
The good news is that you can do some simple things to start increasing retention. There is a wealth of academic literature that goes into proven predictive models, methods, and interventions that will help you decrease attrition.
Below we outline five papers about the topic. Glean from these researchers insights, or join us on our webinar about the topic and we’ll tell you exactly how to use this knowledge to increase your graduation rates.
Paper 1: Early Identification of College Dropouts Using Machine-Learning
In this paper, they looked at whether a machine learning-based prediction model could be made for the likelihood of university students dropping out. They tested different models based on data from the National Education Panel Study. They used information about a student, such as their basic information and subjective information that might be assessed when they enroll, to compare different models.
The results showed that prediction accuracy can increase when students are continuously surveyed about their satisfaction with the course and themselves.
Practitioners need to know different ways of measuring how effective a computer program is. They should not just look for the best one but find the most cost-effective one. They also discussed other hidden costs that might be involved with this type of system. Read more of this paper here.
Paper 2: Early Detection of Students at Risk – Predicting Student Dropouts Using Administrative Student Data and Machine Learning Methods
Education policies and student retention are important issues. Students who drop out of the university system waste the money, time, and effort spent by everyone involved. This paper develops a forecasting system to predict when students will drop out. The system is based on administrative data that is available.
To predict whether or not someone will drop out, they use different methods. They also use machine learning algorithms that are self-adjusting with new data. This helps to reduce the disadvantages of using any single method and the problems caused by heterogeneity between universities with different programs.
In this paper they use data from a state and private universities to develop and test the model. The more semesters, the better.
The Early Detection System (EDS) helps students stay in school. It can help prevent dropouts, or it can help speed up when students decide to drop out of school. This way, you don’t have to spend as much money on drop outs because the EDS starts with early intervention and support for at risk students. Read the full paper here.
Paper 3: What matters most for college completion? Academic preparation is a key predictor of success
This paper suggests that a student’s academic preparation in high school is one of the strongest predictors of college degree attainment. Research has found that students who take more rigorous courses and get higher grades are more likely to score well on college entrance exams, be better prepared for college, and have a higher likelihood of completing their degree. A student’s gender or race also have important correlations with the likelihood of completing a degree, but policymakers cannot change these characteristics.
Instead, policymakers should focus on working to increase the number of students who enroll in high quality postsecondary programs that are a good fit for them. For example: a four-year degree or vocational training program.
It is hard to know what policies to put in place when there is not much research. The problem is that it can take a long time for research to come out, and by the time it comes out, it might not be helpful anymore.
A solution that may not be perfect is to use research that links measures at different points in time. For example, test scores and grades in high school could predict a student’s success in college. Using this information, they can see if an intervention has a long-term effect on a student’s education.
Academics are not the only thing that matters when you go to college. For example, policies can also affect other things like extracurricular activities and school safety. This is important because it will help students succeed in college. The best plan is to create methods for measuring college readiness and then use them in policy and practice.
Paper 4: Beyond Early Warning Indicators: High School Dropout and Machine Learning
This paper combines machine learning with economic theory to predict which students will drop out of high school. The paper also shows that schools can do better by using even more data. This paper emphasizes that early warning systems are not good enough on their own. Other tools like machine learning are needed to be effective at predicting who will drop out of high school. They identify different clusters of students that might drop out, then they try to fix that by using a model.
Paper 5: Academic Performance and College Dropout: Using Longitudinal Expectations Data to Estimate a Learning Model
This paper is estimating the cause of college dropouts. They are using data to help make better assumptions. Their simulations show that 45% of college dropout in the first 2 years can be attributed to what students learn about their academic performance, and this type of learning plays a smaller role later in college. Poorly performing students leave because they don’t want to stay at school if they’re not doing well and they don’t think it’s worthwhile to keep going and trying harder. When people do poorly, it makes school less fun or enjoyable for them.
Conclusion
In conclusion, there is no question that data and analytics will help increase your school’s retention rate. The only question remaining is, will you let Teach Beacon do the work for you? Join us on November 2nd for our webinar and find out what we’re all about.
