With the 2020 Joint Statistical Meetings in the rear view mirror and the fall semester under way, we wanted to take some time to reflect on our first virtual conference of this scale. While we were pleased to see the progress in the field of statistics education, we left feeling a little disappointed in the experience of the virtual conference overall. Below Adam and guest contributor Ann Brearley (University of Minnesota) share some of their favorite takeaways from the conference, which include topics like social justice, consulting and collaborating, teaching math stats courses, and approaches for talking about more difficult concepts…
Takeaway #1 (Ann): GAISE: It’s not just for the intro course anymore
Session #136 (Tuesday, August 4), titled “Modernizing the Mathematical Statistics Course” was a delight; five excellent talks about rethinking the traditional math stat course sequence, both at the undergraduate and the M.S. levels. Between them, the five speakers gave multiple examples of incorporating all six of the GAISE college guidelines (who knew they could be applied to statistical theory!) The ideas included interweaving probability with statistics throughout both semesters, focusing on statistics as an investigative process [GAISE #1] with real examples, using visuals to both aid and assess conceptual understanding [GAISE #2], using technology [GAISE #5] (including apps, resampling methods and R) from the start, and incorporating interactive learning activities in class [GAISE #4]. They also discussed changing from the traditional homework + exams assessment approach to a mixture of low-stakes formative assessments and higher-stakes summative assessments [GAISE #6]. Ideas for assessments included doing analyses of real data [GAISE #3], describing why each step of a proof is needed, creating a quiz question and its solution on a specific distribution, or using simulations to address questions about estimators.
Takeaway #2 (Ann): Teaching Collaboration
Session #115 (Monday, August 3), a panel discussion on “Strategies for Training the Next Generation of Applied Statisticians”, had a lot of good suggestions for teaching both undergraduate statistics majors and graduate students how to be effective statistical collaborators. It was noted that training in collaboration/consulting varies much more across institutions than training in theory and methods, and it was very helpful to see the wide range of approaches that are being used, including role playing, extensive writing and speaking assignments, capstone projects, internships, and working in a consulting center. They advertised having new training videos (created by Emily Griffith, Julia Sharp, and others), which have since been posted!
Takeaway #3 (Ann): Teaching using Social Justice Examples
Beth Chance’s talk in Session #447 (Thursday, August 6) on “Building Statistical and Multivariable Thinking Using Social Justice Investigations” described working with high schools in East Boston to develop statistics modules for use in non-AP statistics classes in racially and ethnically diverse urban high schools. Students use a browser-based educational software package, CODAP, to visualize and explore the data. She described one module where students explore income inequality by sex and education level, and another where they explore the unemployment rate by immigration status and education level. The second module also uses background information from TeachingTolerance.org. Her slides can be found at RossmanChance.com/JSM2020/Chance.pptx.
Takeaway #4 (Adam): Don’t shy away from difficult conversations
Building off Ann’s takeaway, Maria Tackett and Jo Hardin offered some good advice about how to discuss difficult topics in your courses. You can find their slides here. Maria discussed how she uses the U.S. Census to discuss both difficult statistical topics in Stat2 (e.g. missingess and imputation) and difficult social topics (e.g. Who is missing and why? What implications does this have on apportionment?). I’m still trying to figure out how to have difficult conversations in my classes, and framing them in the context of the Census seems more doable for a first attempt. Jo talked about a project she ran using the Stanford Open Policing data set and how she is approaching difficult issues head on through data analyses that allow students to develop advanced tools. I found the discussion to be very thoughtful, and I need to think more about the language I use around sensitive data to emphasize equity and justice.
Takeaway #5 (Adam): It might be interesting to think about intro stats from the “inside out”
Kelly Nicole Bodwin discussed her version of intro stats, and while the approaches were not earth shattering, they gave me some good ideas for my course. For example, instead of taking a lot of time to stress how to do things, take more time to discuss why, and make assessment align with that! Kelly discussed her use of imaginary results, where students imagine what the data would look like for a problem before even opening R, and encouraging students to verbalize what they would expect if certain claims were true. I found myself wanting to see her course materials, which is a good sign!
Takeaway #6 (Adam): I don’t like virtual conferences!
OK, maybe this doesn’t deserve to be a takeaway, but I was pretty disappointed with JSM. I missed running into colleagues in the convention center and the conversations over a pint that occur every day at in-person conferences, there’s just no good substitute for that. Also, I was disappointed by the volume of technical issues, ranging from presenters not realizing there was a delay between when they shared their screens and when the audience could see it (this derailed at least 50% of the sessions I attended) to the URLs linking to the wrong session.
What did you think of JSM?
Share your takeaways or any other thoughts in the comments below!