Using Predictive Analytics to Impact Policy and Practice

Ask eAIR invites questions from AIR members about the work of institutional research, careers in the field, and other broad topics that resonate with a large cross-section of readers. If you are interested in writing an eAIR article, or have an interesting topic, please contact eAIR@airweb.org.  

This month’s question is answered by Jessica Egbert, Vice President of Institutional Effectiveness & Community Engagement, Rocky Mountain University of Health Professions.

The ideas, opinions, and perspectives expressed are those of the author, and not necessarily AIR. Subscribers are invited to join the discussion by commenting at the end of the article.

Dear Jessica: How have you used Predictive Analytics on your campus to impact policy and practice?

JessicaE.PNGEvidence-based practice is an integral part of the mission at Rocky Mountain University of Health Professions (RMUoHP; Provo, Utah). We teach from this model in our degree programs (all graduate degrees in healthcare fields) and integrate evidence-based practice into institutional operations. Predictive analytics inform data-driven decision making; this type of evidence both improves the institution and serves our students.

As a tool for integrating data into policy and practice, predictive analytics support student learning and long-term success. While business and sales models have long adopted these tools, the higher education field has an opportunity to strengthen its use of predictive analytics outside of financial and formal research capacities. Additionally, historically cumbersome processes have been replaced by technology that empowers diverse functions to effectively translate predictive analytics into policy and practice across the institution.

Fortunately, technology has brought with it new methods to improve the effectiveness of related research and its translation to policy and practice. We have included several areas in which we are currently integrating predictive analytics to address student- and institutional success:

  1. Physical Therapy Degree Program – The faculty developed (and use) predictive analytics to assist in setting internal doctoral program admission standards and to predict the likelihood of national licensure exam pass rates. Statistical correlations on multiple factors and regression analyses, as well as ROC curve analyses, determined admissions cut-off points. A combined overall, multifactorial preadmission score set minimum thresholds for different criteria/factors that help predict student success. Additionally, this helps rank candidates for the admissions process. By utilizing predictive analytics to set policies and practices, the outcomes have yielded student success through an overall 100% national exam pass rate.

  2. Nursing Practice Degree Program – This doctoral program is engaged in a multi-institutional, longitudinal research study on student retention that builds upon existing research on strength-based models. Students participate in a strengths inventory assessment and results are used to develop leadership and practitioner skills throughout the curriculum. The personalized model, which emphasizes magnifying existing strengths and talents over improving perceived weaknesses, has been implemented into course design, forum instruction, and grading feedback to students.

  3. Speech-Language Pathology Degree Program – For this developing graduate program, the faculty implemented into practice the use of predictive analytics to ensure student success. Using an internally developed numerical index that includes scores for two faculty reviewers, GRE component scores, and overall undergraduate GPA, the program will analyze final program GPA and national exam scores. A regression analysis will predict the amount of variance of these scores that may be accounted for by each variable. Due to the quantity of variables and the admissions cycle, it will take three full student cycles to collect adequate data for these analyses. The analyses will reveal best practices in admitting and preparing students for successful employment in their degree field.

  4. Satisfaction and Loyalty – Institution-wide surveys are processed through the office of Institutional Effectiveness. Faculty, alumni, student, and employee surveys that have historical yielded only means and trends regarding satisfaction and loyalty (via Net Promoter Score) are now reviewed through more sophisticated lenses. By integrating correlation and regression to the survey analyses, the University is empowered with actionable data on what variables affect overall satisfaction and loyalty. For example, while a mean score may reveal some dissatisfaction, the predictive analytics allow us to dig into the relationship of what variables are related to or may cause that dissatisfaction. By understanding relationships, policy and practice changes become both justifiable and meaningful.

  5. Resource Allocation – In a complex higher education environment, the competition for resources may be fierce. By utilizing predictive analytics to create models, establish trends, and forecast the needs of students and employees, the university creates resources pathways towards desirable outcomes for all constituents. While this may not always imply RMUoHP can provide all desired resources, the use of predictive analytics helps the institution close the gaps between plans and surprises.

By integrating predictive analytics into higher education, we are empowered to make meaningful, positive change for our institutions and for our students. Expanding our practice may not be simple, given how we are often stretched too thin already. Below are some tips to help you move forward:

  1. Stay connected to AIR. The Association for Institutional Research has great resources on this topic – including recommended reading, events, and experts. Reach out!

  2. Do a literature review. There is no need to reinvent the wheel; most areas of interest for this audience likely already have some related data upon which your research or data may be grounded. Use your library resources if you do not have time to conduct a search and build upon this existing knowledge.

  3. Check out resources on your campus. Rather than using a time consuming database application or learning how to do statistics online, find out what access you have to sophisticated web-based and analysis tools on your campus. You may be only a few clicks from regression! Check with your IE/IR offices, research or statistics faculty, software support, or library personnel for resources.

  4. Finally, ramp up your knowledge and your network. Take a refresher stats class on your campus, from another institution, or online (for free!). Demo some new technology. Reach out to your colleagues and ask questions to inspire analysis: “Wouldn’t it be cool if we knew X?” You’ll find a lot of data junkies out there ready to chat!

     

     Comments

     
    To add a comment, Sign In
    There are no comments.