22 October 2014

John Rauser: How to Present Statistics Without the Agonizing Pain

Does drinking beer make you more attractive to mosquitos? This is the question John Rauser—data scientist at Pinterest—wanted to answer in his presentation Statistics Without the Agonizing Pain at the 2014 Strata Conference in New York.
As counterintuitive as this may seem, the answer to this question can save lives. Malaria is transmitted via mosquitos, so “if you can understand which people are at greatest risk for mosquito bites, you can target your interventions much more accurately and you can do a better job at fighting malaria.”



If you ask around whether drinking beer makes you more attractive to mosquitos, there will be people thinking you are crazy and people thinking it might well be the case. This can be seen as a statistical argument. The way you can solve such problems is through a statistical procedure called sampling distribution. I’m not going to explain you how this works, don’t worry. Even because I wouldn't be able to. But there is something remarkable in the way John Rauser presented the problem.

John Rauser's slide: Analytical approach 
to sampling distribution
He showed that there are two methods you can follow: an analytical approach and a computational one. The former is the traditional, painful method most professors would teach you—a method which is purely based on theory; the latter is more tangible. The analytical approach includes formula after formula; the computational one is more understandable because it shows you the statistical process unfold so that you can understand the meaning behind the formulae. It makes the numbers and statistics meaningful. According to Mr. Rauser, the analytical approach is agonizing because “the idea of a sampling distribution is really hard to understand […] and when it’s presented in pure mathematical formalism […] it’s just hopeless.”
Thanks to his effort to make the complicated simple, the audience could understand his topic at a deep level.

Mr. Rauser’s goal was to convince the non-statisticians in the room that the road to statistical fluency is shorter than they think. To do that, he put a great deal of effort into making a complicated topic easy to understand. He put himself into his audience’s shoes. By no coincidence he used simple visuals that amplified his message. He even showed a few pictures. Have you ever seen a statistician showing pictures during a presentation?
Examples of slides used by John Rauser 
As a bonus, he also closed his talk in a powerful way: “The message that I want to leave you with is this...” Whatever follows gets remembered, because the audience understands you are about to conclude, so they will inevitably pay attention.

Does drinking beer make you more attractive to mosquitos? Now you know how difficult it can be to answer such a question. Yet, the good news is no matter how difficult or technical your topic is, there is always a way to make it easier for your audience to understand. [Tweet this] Whether you are a professor teaching a tough subject, an employee presenting a project to some colleagues, an entrepreneur pitching a world-changing idea to investors or potential clients, present your topic without the agonizing pain. If it's possible to explain the sampling distribution of the test statistics under the null hypothesis in a simple way, anything is.

I'd love to know your thoughts on this article. Let me know in the comments below or meet me over on Google+ or Twitter to join the conversation right now!

IMAGE: John Rauser via Pinterest