Q: What kinds of hands-on topics most effectively help you meet the goals of the course?
We cover topics like A-B testing, which has become ubiquitous across industries. Essentially, they are micro-experiments. A firm like Airbnb will send out different promotions to customers to see which one leads to more clicks or more bookings. A-B testing allows companies to very quickly test different strategies, before implementing them on a wide scale. And, as customers, they’re all around us. Every time you sign on to your Uber app, you might be A-B tested without even knowing it.
The interesting thing is that A-B testing is just hypothesis testing, which is an old tool in statistics. But it’s sort of been rebranded. When I used to just teach hypothesis testing, students would ask, “Where am I going to use this? Why is this relevant to my career?” I’ve tried to repackage it and show students that A-B testing is hypothesis testing. As soon as you explain this, they immediately understand the importance of it and the usefulness of it. That’s a general teaching strategy that I try to employ across a range of topics: How can I connect material that, on its own, can be quite esoteric? How do I make it interesting and relevant to them for their career? That’s one big goal.
The second big goal is to help our students think more critically as consumers of stats and data.
I’ll give you an example: In my very first class, I start with this story that was on NPR’s Planet Money a few years ago, making the claim that investing in horror films is the “best deal in Hollywood.” They looked at data for the top 100 highest-grossing movies and found that a disproportionate number were horror films. So, I ask my students: “Is that a valid argument, to say that this is the best deal in Hollywood?”
Some students will get it right away, while others have to think about it. The punch line is that you cannot only look at the top 100 films. Those are, by definition, the extreme outliers, and before you make a film, you don’t know if it’s going to be hugely successful or not. So, what you really need to do is look at the universe of all films and then ask yourself, “What are the chances it turns out to be a high-grossing film?” We go through the data and we show that, in the end, horror films no longer look like the best deal in Hollywood.
In teaching this way, my hope is that after their time here at Anderson, when they encounter a story like that, they think critically about it — and think what could be an alternative explanation.