Following up with our previous post, Six Steps to At-Risk Student Intervention: A University Case Study, found improved student retention post university intervention. To summarize, SightLine identified individual students at risk of dropping out of a Midwest university. Those students were clustered into groups with similar traits. Those common traits flagged them as at risk in the initial analysis. Intervention methods were applied. Interventions were specific to the needs of the targeted group of students. Finally, student retention post intervention was examined. Preliminary data suggests interventions were successful. Fewer students dropped out and retention and graduation rates improved.
Identifying Students At-Risk of Dropping Out
SightLine uses predictive analytics to identify individual students at risk of dropping out. The data was provided by the university. Data provided includes:
- Background data such as the student’s high school location or ACT scores
- Current academics, for example GPA or course load
- Financial status including loans, grants and other scholarships
- University engagement or participation such as athletics or clubs
Algorithms created by SightLine use this information to identify individual students at risk of dropping out. These algorithms can identify students a semester earlier than using GPA thresholds alone. Identifying students sooner gives student success professionals more time to intervene with at risk students. SightLine also provides at risk student forecasts during the summer semester. The University can engage with targeted students when typical engagement is low.
Grouping Students At-Risk of Dropping Out
SightLine understands that each university has finite resources. So, the next step after individual student identification is to group these students with their peers. Clustering these students will allow for different intervention methods by group. Using segmentation allows intervention methods to be unique to the group, but not as specialized as to create a plan for each individual. Different intervention methods allow for more efficient use of university resources.
Again, analytics are used to cluster these individuals based on traits. In our previous post we used two example groups but identified ten in total at this Midwest university.
Example Group A:
- First year students
- Receiving Pell grants
- Low academic performance during their first semester
- Did not participate in on campus activities
Example Group B:
- Highest amount of need-based institutional scholarships
- Receiving Pell grants
- Highest performing academically across at-risk student population
Intervention of At-Risk Students
Different intervention strategies should be applied to each group clustered by SightLine‘s algorithms.
Following our example, Group A would be nudged toward on-campus employment. The group has financial need. The group also needs increased on-campus engagement for academic success. On-campus employment would provide financial benefit as well as on campus engagement. Furthermore, most on-campus jobs have more flexible work schedules. Students struggling with finances and engagement can benefit from on-campus employment.
Conversely, Group B is high achieving with significant financial stress. This group is a wise place to invest more performance or even incentive based scholarships. Those extra scholarship dollars will likely give these students the boost they need to reach graduation.
Student segments at-risk of dropping out respond differently to different intervention methods. This is good news for the university. It means funds are not the only solution to improve retention results. Some groups may simply need more accountability or engagement. This could come in the form of more academic adviser meetings or group study opportunities. Identifying these groups through SightLine‘s predictive analytics leads to more efficient allocation of university resources.
Improved Student Retention in Preliminary Data
Fall to fall retention rates and spring to spring retention rates improved from both the previous year and the 10-year university average. An additional 101 students were retained throughout the past year at this Midwest university.
These improvements were realized after only one year of implementation. SightLine consultants spent approximately one month working with the University Student Success Team. Together they gathered and reviewed data. SightLine did all the heavy lifting. SightLine developed and validated a predictive model specific to the university. Then, SightLine used these models to forecast individual students as at risk and placed them into intervention groups. Next, intervention recommendations for each student segment were developed by the Student Success Team and SightLine together. This work and consultations were performed in June and July, allowing time for summer and fall intervention planning.
This is an ongoing effort and the SightLine team is encouraged to continue developing these intervention plans. SightLine will continue assessing what is working and which strategies deserve continued improvement to better serve students and universities.