An Analytics Perspective on AACRAO’s Guidance for Changes in Transfer Student Admissions Policies Impacted by COVID-19

On April 1st, the American Association of College Registrars and Officers (AACRAO) released its Guidance on Admissions from Novel Coronavirus COVID-19 Impacted Terms. This resource makes several suggestions for postsecondary institutions to consider, considering the unusual circumstances that we are all facing.

Many institutions have already incorporated some or all the guidance from AACRAO into their admissions policies. I want to speak to each from a data and analytics perspective, specifically for Admitting Undergraduate Transfer Students. Hopefully this will set you up to better understand later what really happened during these challenging times. Below are my thoughts on each guideline:

1. Extend any deadline for submitting final college transcripts until the end of their first enrolled semester
   • One of the big questions is will undergraduate transfers who apply near the new deadline be more or less likely to enroll than usual? The truth is no one knows the answer to this question yet, but you may find that these students have a newfound reason to transfer as a result of COVID-19, such as being closer to home. Depending on the outcomes, it’s conceivable that institutions will continue to push deadlines to later dates in subsequent academic years. Incorporating submission dates into your models will likely yield some interesting findings down the road.
2. If your institution has accepted only paper college transcripts or any other supporting documentation needed for admission, develop policy and practice for accepting secure alternative formats
   • Manual data entry from paper college transcripts to your CRM/SIS may be causing you some data quality and operational headaches. This suggestion is a good one, regardless of the current climate, in order to prevent transcription and transposition errors. Nobody wants to spend all that time cleaning data when they could be analyzing it instead.
3. Consider providing and moving to online, other remote or virtual summer orientation, advising, and registration soon
   • Are you collecting virtual orientations, advising appointments, and registration as separate fields from your in-person equivalents in your data? It will come as no surprise that students will react to these types of engagements differently, despite having the same intended results. You will thank yourself later when you’re trying to analyze just how different the impacts of in-person versus virtual interactions really were.
4. Waive standardized testing admissions requirements
   • Many institutions have already been weighing this decision and now they have another good reason to try it. Some of the more analytically advanced schools have used predictive models to assess how different their class shape would be if they were to remove SAT and ACT scores from admission decisions. How much would the absence of standardized test data affect your enrollment class? What will it do to your retention efforts? These are complex questions that require data science expertise to have trustworthy answers.
5. Transition interview and/or audition requirements to a virtual platform
   • Similar to suggestion #3, it would be ideal to know which interviews and auditions were conducted virtually in your data for future analysis.
6. Consider moving to online and/or remote course placement testing options OR develop a rubric of course placement eligibility based on other available data, if applicable
   • Document any and all changes like this and, if you can, create flags in your data for students who were placed outside of normal circumstances. You may find that your approach for placement of these students led to different outcomes than would have been expected.
7. Be flexible in your international credentials guidelines admission guidelines and refer to example protocol for international evaluation and Inclusive Admissions Practices for Displaced & Vulnerable Persons
   • As if international student enrollment wasn’t suffering enough… chances are you were already being flexible here. Variation in your admission decisions can ultimately be a good thing for predictive modeling, especially with international students where less information is typically available for admissions decisions.
8. Two scenarios for calculating transfer GPA for admission into the institution from a post-secondary transcript that includes any pass/fail (P/F) or other non-qualitative, binary marks for the terms impacted by COVID-19
   • Scenario A: Student has completed enough credit hours to be admitted as a transfer student under current policy
     • Count Pass(P) and other equivalent non-qualitative, binary marks grades as earned credit
     • Exclude all non-qualitative, binary marks from the transfer GPA calculation used for admission regardless of whether credit was earned for the course
   • Scenario B: Student has not completed enough credit hours to be admitted as a transfer student under current policy
     • Same as above PLUS use the stated admission policy for these students as you would under normal circumstances
     • Calculate a transfer GPA for admission into a selective major
   • For this suggestion, it’s especially important to make sure you’re tracking which students have been allocated credits outside of the normal procedure. Similar to suggestions #3 and #5, changing the definition of what counts as earned credits will likely lead to issues in predictive models. Knowing which students have which number of credits earned from abnormal circumstances will prevent these issues. From there, the expected outcomes of these scenarios would be difficult to measure unless your model is already incorporating credit hours as a feature towards predicting transfer enrollment. As we’ve observed in our own enrollment models, the number of transferrable credit hours is not only important in the admission decision and potentially for determining scholarship offers, but also a strong indicator of how likely a transfer student is to enroll. This is why accurate tracking of credits is especially important.

On the surface, each of these guidelines is very practical given the circumstances, however, following the best practices I've provided should help your institution to maintain long-term data quality.