Statistician and journalist Nate Silver is the founder and editor-in-chief of FiveThirtyEight, a news analysis website that uses statistics to explain prominent stories. His accurate predictions during the 2008 and 2012 presidential elections elevated him to national prominence. He will be the commencement speaker at the 190th Commencement on May 19.

What’s one of your pet peeves when you see members of the media interpreting polling data? Is there one particular thing that gets miscommunicated that you’d like to clarify?

I think after 2016, I think people not communicating the margin of error in polls is a big one, and not explaining [that] when a candidate is only a couple of points ahead of that candidate, [he or she] is going to lose a fair amount of the time, although not most of the time. There’s still a lot of cherry picking of polls. If you have five or six surveys of the race, you should probably be trying to represent the average, or at least give people what the range is. Instead, people are trying to write stories that fit their narrative and so they’ll kind of pick and choose which stories and which polls they use a lot.

I think right now people are maybe a little bit too obsessed with the polls of the midterm elections in November, but as we get into September, October, then those are a lot more reliable.

You suggested right after President Trump’s election that perhaps his polling numbers would affect the President’s judgment, and he might abandon some of his signature policies if his polling numbers didn´t appear good to him. Do you think that’s been happening?

The healthcare bill they tried to pass last year, along with firing Comey, were maybe the things had the biggest downward effect on Trump’s approval ratings. One can imagine a scenario where Democrats take the House and Trump starts to actually be cooperative and bipartisan and that could actually help him, maybe. But he has improved from a 39 percent approval rating to 41 percent, which doesn’t seem like much, but 41 percent is closer to being “normal.” It’s closer to being where Obama was at this point eight years ago, or where Reagan was at this point two years into his term. Now it’s a little different in that that 40 percent doesn’t seem to move for any reason, people are very fixed in their views of Trump.  But from an approval rating standpoint, being big outlier in the first year where Presidents are usually very popular, he’s kind of caught up to the back of the pack. Not to Eisenhower or whatever, presidents who were really popular, but there were several presidents who were kind of embattled at this point in time, and also those presidents learned on the job.  I don’t come to a judgment of whether Trump can learn on the job in the third or fourth year, but in the abstract, it wouldn’t be that surprising if he has a bad midterm and then recovers and then potentially wins reelection, like what happened to Clinton and to Reagan and to Obama. It’s a very common track.

Let’s go back to when you were an undergraduate and preparing to launch into the real world. How do you remember feeling and how do you reflect on that time, as someone who has experienced post-grad success?

I went abroad and I came home my senior year. Going abroad is great, but it sets up your senior year in a really weird way. You’ve taken the bulk of your coursework, you kind of have one foot out the door already, so I had a fair amount of senioritis. I wish I had done a little bit more thinking about what I want to do next.

Can you give us a preview of the speech?

I have no idea what I’m talking about yet. I haven’t given any thought to it until this call today. [laughs]

This interview has been edited for length and clarity.