Teaching Statistics in Public Health During COVID-19: Lessons Learned From Spring 2020 and Recommendations for the Future

General Teaching/Pedagogical Approaches

While statistics instructors are perceived to be more technologically savvy, that does not necessary apply to teaching using Zoom, using asynchronous approaches to teaching, and coping with the multitude of changes that happened so quickly to all of our lives due to COVID (e.g., dependent care). Even if you had taught online before, we were suddenly teaching students who did not sign up for an online class. While many schools provided resources to help with the transition, the suddenness of the switch to online left many overwhelmed. As an example, shifting group work from in-person to Zoom is possible, but requires thoughtful adaptations.

For us, the challenges were different between the courses taught for MPH and PhD students. Our MPH courses tend to be larger, and required different adaptations compared to the smaller PhD courses. For the MPH course, some lab-based activities had to be shifted to instructor-led demonstrations. Some individual assignments were dropped in favor of group-based assignments. The group assignments involved applying statistical principles and code demonstrated in class to analysis projects unique to each group. Instructor and group interactions were managed through Zoom. We found that when group sessions occurred during class time, they could be managed with Zoom breakout rooms allowing the instructor and teaching assistant to circulate among groups much like in an in-person setting. Overall, this resulted in more “hands-on” interaction for the instructors, but allowed us to ensure that every student was able to engage.

One of the PhD courses is heavily based on readings and discussions about advanced statistics. The main goal for this course is for fledgling future researchers (i.e., first year PhD students) to become comfortable with breadth of statistical methods and their very limited knowledge and abilities at this point in time. Compared to previous cohorts, it was harder for students to cope with being uncomfortable venturing into the unknown. The readings are intentionally demanding. With the new uncertainties of COVID life, the instructor had to rethink about how much discomfort to add. We spent more time discussing the readings and how to approach them during class. The instructor also provided more guiding questions. The students responded well to this shift mid-semester.

Supporting Students Outside of Class Time

There are many students, especially those who come with less privilege, who do not seek support from the instructor outside of class (and often even during class time). As discussed in detail in the seminal book The Privileged Poor: How Elite Colleges Are Failing Disadvantaged Students (Jack, 2019), those who need more support are often less likely to seek it. With this in mind, we adapted our approaches. Across our classes, we intentionally talked about the importance of contact time with an instructor outside of scheduled class time. We increased both time and modes of instructor access across both sets of courses. At times it was necessary to be flexible and meet with students at odd hours.

The larger MPH courses required options that leaned more heavily on group interactions. We increased our use of discussion posts on Canvas and encouraged students to meet with their instructor in their assigned groups. This included using both prescheduled office hours and flexible meeting schedules outside of regular office hours. This required flexibility on both our part as instructors and for our students, as everyone juggled both class and sudden changes in daily responsibilities (dependent care, etc.). By meeting with students in their assignment groups, we were able to manage regular contact with each student while also keeping the burden manageable for the instructor.

For the smaller PhD courses, a more individual approach was feasible. To add more informal opportunities to ask for help, one instructor opened their Zoom room and emailed students the times when they could “stop by,” mimicking an ajar office door. The invitation explicitly offered an opportunity for casual chats, group discussion, or more structured meeting.

Technical and Space Problems

Our doctoral students have the benefit of receiving a laptop with statistical software packages installed on them (our MPH students are only given a list of requirements for a laptop and have to acquire it themselves). However, not everybody has high-speed internet access, and heavily used tools such as Zoom and Citrix/Apporto do not function well otherwise. Paying close attention to technical issues (e.g., students dropping out of Zoom) and providing solutions (e.g., switching their camera off or calling in via phone) was critical. When teaching statistics, coping with technological issues was an added barrier and took away important classroom time. It also exacerbated anxiety for many who already have high levels of anxiety around statistics. Additionally, many students had less than ideal work spaces in their homes. Trouble-shooting with students on how to find a good space for class time and other work hours was necessary to support students’ learning. Being cognizant that many students did not have the luxury of having a screen beyond their small laptop was important throughout. Some students would not have brought it to our attention if they could not read what was shared on Zoom or that they are attending class while sitting in their car outside a business offering free Wi-Fi.

Assessing Learning

Across our MPH and PhD courses, we strive to make assessments as authentic as possible, culminating in answering an original research question through a poster presentation.

Across courses, we chose to recreate as much of the in-class experience as possible via Zoom. For the presentation of final products in MPH course, two groups of students were assigned together into 30-minute slots. Each group alternated presenting their findings to the audience (instructors as well as the other group) and asking critical questions about the study’s methodology and results. Everyone still received the benefit of presenting a virtual poster to an unfamiliar audience, but restricting it to two groups at a time kept it manageable. For the PhD course, the final assignment was adapted to be asynchronous. Students used Microsoft PPT to create a 5-minute recorded audio presentation of either a poster or a brief multislide presentation. Over several days, students reviewed a set number of their peers’ presentations and engaged in Q&A using discussion boards. They finally provided their peer assessment using a Qualtrics form. The process taught students valuable skills in creating online presentations for online conferences such as APHA (American Public Health Association) in late Fall 2020.

Many of the challenges we encountered can be alleviated in the future.

Teach Statistics Remotely

With the recommendation of keeping at least 6 feet of distance, teaching statistics online is much more in line with students’ needs. In a classroom, instructors cannot social distance and approach a student’s computer to see why their code might not work or why their output looks different from their peers’. They cannot ask them to let them take over their keyboard/mouse to trouble-shoot either. On Zoom (as well as other platforms), students can share their screen and allow an instructor or teaching assistant to take control of their computer. This is very similar to what happens in a non-COVID classroom when a student asks for help during a lab. If possible, supply incoming students with laptops that have all the necessary requirements. Make sure that if this is not possible, that lower income students have support to acquire an adequate computer and software.

Create and Curate More Content for Asynchronous Learning

This helps alleviate issues related to connectivity and schedules of students and instructors that are affected by COVID-19 needs such as dependent care (Daniel, 2020). It is important to review and maybe caution against some statements made in content created by others, but it is much less work to locate such materials than create them from scratch. Additionally, with a decent webcam and headset with a noise canceling microphone, creating your own content is possible, albeit time-consuming. If teams of instructors share the workload, we can create asynchronous materials more efficiently. Ideally, if schools can hire additional instructional designers, the burden on the individual instructors can be lessened. Less financially straining solutions could include hiring contractors and pooling resources across schools and possibly institutions.

Improve Students’ Technology Readiness

Talk to your IT department about the needs we have as statistics instructors—preferable as a group. Make sure that incoming students know what the computer requirements are. This can mean the need for pushing only using Windows computers, as it is easier for instructors to troubleshoot on multiple platforms. Communicate clearly how better internet access impacts learning—but also work on creating ways to support who either do not have high-speed internet access where they live or cannot afford to pay for it.

Support for Instructors

With more technological problems and more need to help troubleshoot, there is a greater need to improve the instructor/teaching assistant to student ratio. If possible, have IT support present/on call during synchronous portions of the course (Baran & Correia, 2014).

Provide More Diverse Opportunities for Office Hours

Especially for statistics, individual help is essential or a small problem can lead to a larger one. When we are on campus, we can have the door ajar. As described earlier, using Zoom or other platforms can be used to allow the students to virtually drop in. It will reach some students that might not be able to make it to office hours or are more comfortable with a less formal way to approach an instructor (Jack, 2019).

Talk About Mental Health

Many students have anxiety around statistics (Morsanyi et al., 2016). This anxiety is exacerbated by COVID-19 (Zhai & Du, 2020) and by other current events such as recent police killings of people of color (Bor et al., 2018). It is imperative for instructors to acknowledge the world we live in and how it impacts all of us. Being vulnerable ourselves opens the doors for students to feel comfortable to admit when they are struggling. Teaching remotely makes this harder, but it is not impossible to overcome. Not all of this has to be around “heavy” topics. Including some fun activities into your classroom and encouraging sharing (e.g., pictures of their home work spaces or their favorite COVID outfits) can help build community. In this time, it is more important than ever that you are aware of your institutions’ mental health support systems and refer students if needed. Ask for help from your institution’s experts if you are unsure of how to support a student who concerns you.