“One of the ironies of having more data is having more misinterpretation of that data.” — Nate Silver at SMU
FiveThirtyEight.com founder Nate Silver, who calls himself an “applied statistician-journalist or something like that,” was peppered with non-stop questions during SMU’s Willis M. Tate Lecture Series Student Forum May 5. Here are a few of the most memorable exchanges (his answers italicized).
What issues would you like to explore more on FiveThirtyEight.com?
Education, including higher ed, is ripe for study, especially the U.S. News and World Report rankings. I’d also like us to take a more data-driven approach to healthcare as well as crime.
How might data science transform the healthcare industry?
There are two constraints: One, physicians and others in healthcare tend to be a tight-knit, reticent group and two, there are serious concerns about privacy issues.
What’s something you can tell us about the runup to the 2016 presidential election?
No party has had a long-running advantage because they adapt to changes of circumstances. For instance, some Republicans are de-emphasizing gay marriage and immigration issues – and the GOP has moved from talking about social issues to the economy. It’ll be a long road ahead.
Any advice for SMU students embarking on their careers?
The first thing I’d say: You can’t escape hard work. Another thing: Find problems you’re interested in and the find a way to solve them; you’ll break down your own barriers in the process. Also, be skeptical of the thing that everyone else is doing; keep an open mind.
How do we keep ourselves from just finding facts that “dress up” our own agendas?
Be skeptical of your own belief system.
SMU Meadows Asst. Prof. Jake Batsell mentioned his new media for entrepreneurship class and asked if FiveThirtyEight.com had experienced a “pivot” moment that’s led them to change their approach.
Beyond staying hyper-focused on the long-form pieces we call “jumbo” stories, we’re looking at ways to more quickly and intelligently respond to the news. We’re also spending more time on audience development. We know our subject matter appeals to smart, young people, so we first thought, hey, we’ll publish stuff and those people will just find it. We’re realizing now it doesn’t quite work that way.
How would you re-design a typical university statistics class (typically seen as gruelingly boring)?
I say teach it like a good literature class: Spend 75 to 80 percent of the time focused on great works of writing and don’t have a lot of conversation about syntax.
Dedman Law Professor Emeritus Fred Moss asked, “I want to get down to the nitty gritty. Do you think the Texas Rangers will finish … in the basement?”