Visualizing Fall 2020 US Higher Education Plans
//Update 6/24: See postscript post that combines in-person and hybrid categories to address flaw in the data source.
The Chronicle of Higher Education has a useful page where they are updating a collection of individual institution plans for the Fall 2020 term in US higher education.
Notes: This list focuses primarily on institutions that serve undergraduates, and it excludes colleges that are 100 percent online. For purposes of tallying colleges’ intentions on the pie chart, The Chronicle leaves the “Category” column blank for university systems and instead marks it for the individual campuses, which are listed separately below their parent system. When a system has not enunciated a policy for all of its campuses, only the individual campuses are listed. Colleges that plan a mix of online, in-person, hybrid, or blended learning are classified as “hybrid.” Colleges that intend to convene only a few classes in person, such as labs, are classified as “online.” Because of rounding, figures in the pie chart might not total 100 percent.
I thought it would be useful to add some visualizations for this data beyond the aggregate percentages listed on that page. In order to combine with geographic and enrollment data, I combined the Chronicle data with IPEDS data. This method allowed me to analyze roughly 93% of the data set (currently with 1020 data entries), excluding systems, duplicates, and some institutions I just did not get around to matching up. 1The biggest challenge, for what it is worth, is matching the Chronicle’s naming system with the IPEDS institution name field. Please note that this scaling is done by the most recent IPEDS Fall 2018 data, and the actual enrollments for the fall are unknown and likely to shift.
Summary Table by Sector
When looking at summary tables based on count of institution and scaling each entry by its enrollment, we can get a better sense of how the different sectors have different plans.
The data skews in that smaller schools have a higher tendency to plan in-person for the fall, resulting in 64% of schools but just 51% of total enrollment of those schools. Likewise, 8% of reporting schools plan to go online, but 15% of total enrollment plan to do so.
We can also see significant differences by sector, with community colleges (i.e. public 2-year) much more likely to plan for online while private 4-year are most likely to plan for an in-person fall.
Map Views
When looking at the geographic location of the campuses, we can see patterns of online plans on each coast. Note that the size of each bubble represents Fall 2018 enrollment.
By highlighting the schools planning for online, we see just how much California dominates this type of planning for the fall. The large red bubble in New England is Southern New Hampshire University, with Harvard below it and tiny Bowdoin above it.
We shouldn’t forget Hawai’i and Alaska.
I’m not sure if there is any useful information here, but this view highlights the schools that either have not yet decided on modality or are considering a range of scenarios. Please note that much of the Chronicle data is self-reported and shows roughly one quarter of US Higher Ed institutions at this date.
Correlation to Online Factor
I found it interesting that if you just show schools where more than half of students took at least one online course in Fall 2018, as a measure of prior online experience, there is not an obvious correlation to Fall 2020 plans. In fact, you can see Liberty University, the University of Central Florida, the University of South Florida, and Ivy Tech Community College – some of the schools with the highest online usage 2 Remember that 100% online schools like Western Governors University are not in this data set – all planning for an in-person fall.
None of this analysis is exhaustive, but I hope these visualizations add to our understanding of Fall 2020 plans for US Higher Education. And I’d like to thank the Chronicle for updating and making this data available for analysis.
Update 6/24: As the commenters below have noted, the Chronicle data are not ideal – the data set misses much of the complexity of school decisions for in-person and hybrid, in particular. I will try to modify the visualizations to take this ambiguity into account in a future post.
Phil – I was looking at that Chronicle data yesterday. I think its misleading. I think most of the campuses that will have students returning to campus in the Fall will also have many, many more online courses and that “in-person” category needs to be far more nuanced. From my experience, campuses are planning for more many online, more blended, closing earlier, and have a pivot back to remote in the works if need be, even where they are advertising “in-person”. Lumping all that under planning for in person misses the bigger picture that is reshaping higher education nationally and globally.
Peter makes a good point. As a school referenced above (UCF), we are planning additional “regular” online courses, converted “remote” courses, blended courses, and face-to-face courses. Those courses with face-to-face elements will also include redundant online strategies to allow for social distancing and students swirling in and out due to expected needs to self-isolate. We are also going completely remote after Thanksgiving. Thus, to say that we are planning to be online or planning to be face-to-face misses the complexity of the actual campus plan.
To your point, last week, we announced 90% of all Fall 20 courses will be online, and all courses will be after Thanksgiving:
https://covid19.umbc.edu/fall-2020-course-information
John Fritz
Associate VP, Instructional Technology
University of Maryland, Baltimore County (UMBC)
Peter, Tom, John (and others on Twitter) – I agree that the Chronicle coding of data, particularly the in-person category, misses much of the complexity of school decisions. I’ll see if I can modify the visualizations accordingly, as long as I don’t have to recategorize all data points..
Hi Phil, interesting stats. I reach marketing research and see an opportunity to use your data to make a comparison of two population (proportions). Can you provide the number of institutions (n) for your tables?
https://www.socscistatistics.com/tests/ztest/default2.aspx
Clay – this visualization was based on Chronicle data that had 1,020 entries. After cleaning up, I used 950 data points.