Kevin Reuning is an assistant professor of political science at Miami University whose work focuses on political parties and social movements in the United States, and takes a particular interest in how political activity can be measured and quantified. He also maintains a website, mapping union elections in the United States.
Episode Description
Each election cycle experts appear in the news talking about polling data. Many of these experts are political scientists and are often the only contact audiences may have with political scientists. While polling research is important to the field, there are other kinds of quantitative research in political science that can get overlooked. We learned about some of them today on this episode of Stats and Stories with guest Kevin Reuning.
+Full Transcript
Rosemary Pennington
Each election cycle experts appear in the news talking about polling data. Many of these experts are political scientists and are often the only contact audiences may have with political scientists. While polling research is important to the field, there are other kinds of quantitative research in political science that can get overlooked. We learned about some of them today on this episode of Stats and Stories, where we explore the statistics behind the stories and the stories behind the statistics. I'm Rosemary Pennington. Stats and Stories is a production of Miami University's departments of Statistics and Media, Journalism and Film, as well as the American Statistical Association. Joining me as always is regular panelist John Bailer, emeritus professor of statistics at Miami University. Our guest today is Kevin Reuning, assistant professor of political science at Miami University whose work focuses on political parties and social movements in the United States, and takes a particular interest in how political activity can be measured and quantified. He also maintains a website, mapping union elections in the United States. Kevin, thank you so much for joining us.
Kevin Reuning
Thank you so much for having me here.
Rosemary Pennington
So we are a few weeks away from the midterm elections, as we are recording this episode. And I turned on the news this weekend, and there was a lot of discussion of polling data. What is something happening in your field that people should be paying more attention to? And journalists should be covering more closely?
Kevin Reuning
It's interesting, especially the journalists covering more closely aspects of that question. Because when it comes at least to two elections, let's just stay there. For now, a lot of political science research is about the fundamentals of the country, sort of like the economy, things like that, and how that informs who actually wins it and who actually loses. And so there's sort of like, I will say, two paradigms, but like two ways of thinking about elections, one is focusing on polls, and like, who's up and who's down. And the other one is focusing on this fundamentals aspect. And you can see, like 538, and some of the other ones will include this information there. But for a lot of political scientists, I think the daily polling is not that interesting. Just because there's a lot of variation in it. You know, one word examples like that, partially because it's so difficult to do. One thing that we know is a real problem with polling is non response bias. But I know this, this is kind of going back to polling. So let's stay there for a second, then we'll move on to something else. But non response bias is, you know, the fact that people don't respond to polls, right, which is a problem in general, because no one answers the phones anymore. But there is also some really interesting research showing that when your candidate is doing worse, you become less likely to answer the phone. Oh, yeah. And so it can create this, like the perception that things are moving even more than they are because, you know, most people, they're diehard Democrats died, Republicans are not gonna change their mind, but their candidates are doing poorly, they're not going to talk to a survey researcher. And that's gonna make the Election look much more volatile than it actually is. And so sometimes you will think that I spent all my day on like, 538, or all these other ones that I have some friends that I will occasionally talk about, but it's not, it's not commonplace discussion.
John Bailer
I find that really an interesting distinction that you're raising their distinction between the fundamentals of how this political process works, which is your field of inquiry, versus the way we're encountering it. and public life, you know that we're encountering the political process. It's a horse race, it's, you know, you want, a nose, you know, we're heading around the final, the final bend towards the finish. And there's this drama associated with these daily volatile numbers versus kind of this background structure. Can you talk a little bit more about some of the other foundational ideas that a political scientist might be interested in investigating?
Kevin Reuning
Well, it's interesting, because it's not another question like that. So let's not do that. It's hard to answer that question. Really simply because in reality, political science is sort of related fields together. So I do American politics, mainly, I study what happens in the US, generally. But political science itself has, you know, also international relations, which is studying how countries interact with each other, mainly wars, international organizations, things like that. And then there's also people who do comparative politics, which there's always a discussion of as American politics, just power of comparative politics. But you know, the name kind of gives it away that comparative politics is all about comparing countries that's looking at, you know, like, I had a master's student last semester last year, who looked at welfare programs in different countries, and how that was related to happiness, you know, happiness outcomes. So it's hard to say like, in general, were like the foundational things in political science, let's, let's stick to the foundational things in American politics at least. And we can even kind of keep it back towards sort of like elections and political participation, because that's what I do best. And I think one thing that we need to keep in mind and like it's a fundamental understanding is that there's a lot of variation in how much people participate in politics. And that variation is linked to a lot of other attributes, right. So the rich people have more income, more education, they tend to participate a lot more in politics, and those who are low education, low income. And these things have also changed over time, too. So, for example, there's been a growth, we talk a lot about polarization in this country. So meaning that like, the parties are moving further apart, or that people are becoming more entrenched in their party. And that's happening. But another part of that is a middle group that's kind of opted out of politics. Because it's a lot of anger, a lot of frustration, and you don't see a lot of change. And so if your choices are between being involved in politics, being involved in politics, and watching sports, or TV or whatever, it's easier to do the latter. And I kind of got distracted from the actual or the question.
Rosemary Pennington
You brought up this issue of participation, though. And I wonder, how do you define participation? And how do you measure it?
Kevin Reuning
Yeah, I actually do an activity. One of the activities with my students where we do in my applied research methods class is trying to figure out how we conceptualize and then operationalize. So actually quantify complex things. For example, political participation. Because there's a lot of different ways to participate, there's voting, which is what we generally think of, there's also calling elected officials writing letters to the editor, going out to a protest, and then the added sort of component is that some of those things matter more, and also some people we think about or more, say, and some people, like can't do, some of them, like if you're 17 years old, you can't vote. So all 17 year olds, then less politically active, maybe, but voting is a pretty minor form of political activity. And I would think if you are spending our day volunteering on a campaign, which some 17 year olds do, and then just can't vote, you're playing more politically active than someone who just votes, right. So what we often do is, instead of talking about political participation, in general, we focus on particular aspects of it, we talk about who goes to a protest or who votes. Because those are, I think, more interesting questions, because they're more specific. We know they're all linked together, right? Like, if you vote, you're more likely to do all these other things and so on. But when we're doing, you know, specific research, we rarely talk about political participation in general, instead, focus in on one one part of that.
John Bailer
You know, so Rosemary started with comments about kind of observing the flood of polls at this time as we're heading into these midterm elections. And you know, I found that, you know, I was at the gym and looking at, kind of watching the screens of just, you know, alternating attack ads seem to be the diet that the news pre-rendering news broadcasts seems to be there. I was fortunate I didn't have the sound on so I didn't have to endure them other than occasional viewing it. So my question is, you know, we certainly have seen an expansion of this type of campaign mechanism. And you know, does it work?
Kevin Reuning
It depends.
John Bailer
That's a typical statistician's answer. Wait, wait.
Kevin Reuning
Yeah. And it's often the answer I give my students, which they, I think, don't like, because life is complicated. Yep. Amen. So there's actually multiple parts to a campaign. And so what we're seeing generally now are more of the persuasion areas of the campaign. So at this point campaigns have ideas of who, you know, are their diehard supporters who are in the middle, and they've probably tested, tested, I'm using air quotes here, you can say different messages. The extent of those sorts of persuasion messages to get people to support one candidate or the other work. It's hard to know for certain which one works well, I should say that some of them definitely work. And there's often research afterwards. You know, there's some really interesting stuff coming out of UCLA. I think right now, where they looked at lots and lots of campaign ads, got people to watch them, they found that there definitely are some that work, but knowing what's going to work ahead of time is really hard to do. You know, pull or campaigns will do these things, really do surveys with people and they'll ask them a question that I actually really despise, which they'll give someone like, Dr. Pennington is a horrible person who hates puppies, would you put does this information make you more or less likely to support Dr. Pennington? And then they'll use responses to ask a variety of different things. They'll say also, like Dr. Pennington is a wonderful human being. She's only ever petted puppies and done nothing else. Does that make you more or less likely to support them? And they do that to figure out their messaging and what political science I think has come to accept is that that way of doing that is really kind of pointless. And instead, you've got to do full on experiments where you're giving people different, different treatments. And you're then asking them not does this make you more or less likely you're asking them and said just do you support this person? Because we're really bad as humans have fear of saying what's going to make us more or less likely to support someone?
Rosemary Pennington
Has there been something that's emerged that feels like it is a persuasive, useful, persuasive approach? You know, in communication studies, we often talk about emotion and how emotion is the thing that can trigger people to do things like if you give them something that makes them angry, or frustrated, they often will act on it, if they're happy, they won't like anger is a motivating emotion in a lot of Communication Studies. And I wonder if there's anything that's emerged in political science that's been shown to be a pretty persuasive approach?
Kevin Reuning
Yes. But this goes off what you're saying with the anger, I feel like framing it like making it very persuasive kind of has almost like a positive tinge on it. And what I'm going to talk about is not necessarily positive, what we know now, or what we've kind of identified as political scientists as the importance or growing importance of what's called negative partisanship. So we think of people as being partisans, whether you have an identity as a Democrat or a Republican. And traditionally, we thought of that as sort of like, This is who I am, I am a Democrat, or I am a Republican, it means xy and z. And research has pointed to the fact that well, by defining myself as one thing, I'm also defining myself as what I'm not. So it's not just that I'm a Democrat, it's that I am not a Republican, or I am vice versa. And that ties back to persuasion because what we're seeing and you know, I don't think this will be surprising to people watching. The ads that you're watching at the gym, is that people respond a lot to sort of this negative partisan framing so that it's a lot easier to get people angry at the other side than to get them very supportive of the side that they're in. Right, because we're seeing this, you know, to touch on at least some news about Georgia right now with Herschel Walker, right, Georgia. Yeah. With Herschel Walker. He's had lots of scandals at this point. And in the end of the day, though, he's probably not going to lose very much support, because as much as some of his voters might not like the fact that they have to vote for someone who is, you know, paid for abortions or whatever. They'd much rather have Herschel Walker and they're the Democrat. And in the end, at least he's not a Democrat. And the same thing happens for Democrats as well. I'm not going to pretend like there's perfect equality across the political spectrum on this sort of polarization aspect. But in general, a lot of negative partisanship is what drives people To support candidates and go vote.
Rosemary Pennington
You're listening to Stats and Stories. And today we're talking with Miami University political scientist, Kevin Reuning. Kevin, I know you've done some work on local political parties. And I wonder if you could talk about some research you've done specifically on Facebook and sort of how I think local Republicans were using the space.
Kevin Reuning
Well, the interesting part is not necessarily about how they were using it, but how the space was kind of using them, so to speak. So in the United States, again, those local problems are often seen as a joke. I don't know if you've ever gone to a local Democratic or Republican meeting. It's usually, you know, five people in a room, we've known each other for the last 40 years. And historically, we've had what are referred to as weak political parties. And what we mean by that, because I'm seeing your face there is they're organizationally weak, they don't have a lot of organization to them. But myself and a few other colleagues, were actually really interested in local parties, in part because this is a place where a lot of our students kind of get their first political involvement. It's one way they can get involved in politics. Literally, this project started with a colleague lehan at Wright State, who is their internship coordinator, and he had had trouble getting a hold of one of the local parties, I can't remember which to place and interns. And so this led to us, the Holy Hannah and Whitestone, I'd be interested in just what our local parties were doing. And so the first part of this was just collecting data on it. And there's 3000, some counties in the United States, most states have local parties at the county level, but not all, because the United States is unique. So we collected basically, 3000 are the other social media and email addresses for as many of the counties as we could find. And one of the things we are interested in doing is just looking at what they're posting on social media. Facebook is obviously important in our politics, right now, this was partially during COVID, I think we might have started at pre COVID.
John Bailer
Time as hard as last meeting,
Kevin Reuning
Yeah. And so we were just interested in what they're doing. And one of the things that we found here was a really interesting trajectory of what had happened over a relatively long timespan on Facebook. So Facebook actually is relatively good, relatively good about making data available for researchers. And we are looking at the response rates to Republican to democratic parties over time, going back to 2016. If I'm not sure, it'll be interesting to your readers or not. But I think this is interesting, a good example of how academia works. What we'd started doing was a smaller project, looking at just a year. We presented this at a conference, and one of the participants there said, you know, you have this data that you should really go further back. And so I did that a few weeks later. And what we found was that, you know, from 2016, to about 2018, ish, Democrats, Republicans had equal sort of responses to what they're posting on Facebook, they would allow them that get a lot of response, but they'd get equal number of likes, you know, shares, things like that. And then, starting in 2018, there was this really market change where Republican parties started getting a lot more attention on Facebook. And so this kind of opened up a question about why that was happening.
John Bailer
That's great. I mean, so what changed?
Kevin Reuning
Well, you know, so the truth of the matter is, I can't fully tell you what changed, right? Because we don't know Facebook, they're getting again, slightly better, even now doesn't make available their, their algorithm, though, I did actually review a piece that got a lot of information about it, but it's not out yet. So I probably shouldn't talk about it. So Facebook doesn't make a lot of their information available on this. But I had remembered that around that time period, or after this time period, there's all this news that came out about Facebook having they're changing their algorithm. The Wall Street Journal had this great investigative reporting about Facebook realizing that when they changed their algorithm, it led to a lot more anger on their site. And it was this whole back and forth that they did. And so what we did then is like, well, this kind of lines up with the time period we saw, but it could be other things. So we tried to generate a list of other possible things or ways we could sort of limit it down to being this. So one of the things we did is we had similar data for Twitter, as we looked to see if we saw the same trend on Twitter because it could just be that the parties are posting better. No changes really on Twitter, at least not in the same way. Um, so Okay, well, he asked me something on Facebook. Well, these posts are becoming on our Facebook all the time because they're not well organized. People literally forget their password, things like that. So we checked to see like, didn't have to do with how old or new the parties were or didn't really matter. All these Republican parties saw this strand. And then we did some modeling to try to figure out if we could identify sort of like, around specifically when this change had happened, because you know, you could see it in a graph, we want to know like, is this talking about something more specific, like the end of fall, right. And we were able to identify that it was basically August September's when this real change happened in 2018. And that actually linked up really well with some of the stuff that came out from the Wall Street Journal. For example, there was a, they had an email from a Buzzfeed editor at the time being like, stuff that doesn't go viral is starting to go viral now on your on your, like, on Facebook, or stuff that we didn't use to see going viral is going extremely viral. And so all these things point to the fact that there is something about this change in the algorithm. And something about Republicans pre existing, you know, how they're posting on Facebook that led these Republican Party posts to become much more like shared much more widely, widely on Facebook, and it happened for about a year. And again, what we can tell from all this information is that Facebook spent a lot of time tweaking their algorithm. And so after a while, it sort of settled back to normal. And so again, we can never say certainly, like what happened here, but we can say that there was this, you know, going from being even to Republicans getting double the shares that Democrats were getting to then being about Evilish. Again, and that lines up with this time period.
John Bailer
You know, you were mentioning, this the idea of of sampling and doing polling and, you know, there's there's been this incredible change going from times when you could study a community using random digit dialing where everyone had an area code that lived in that area, to now with cell phones, all bets are off. But now, as you've talked about social media, you know, I wonder about the impact on political science and on studying political processes that social media has, whether it's Facebook, or Twitter or any other kind of outlet?
Kevin Reuning
Well, so I think there's a couple of interesting answers to that. So one is that, you know, there's, whenever something new like this comes along, I think there's one set of questions that sort of obvious, like how things change, right? People drag politically on Twitter, in real life houses changed on Twitter and Facebook. And so like, that's an obvious one. But the other thing that that sort of, I don't know of more interesting is more interesting to me, is that one of the wonderful things about something like Twitter and Facebook is that it's, it's led people to actually quantify things that they wouldn't normally have quantified, like their relationships, right? You know, we are really interested in how connections matter. And sure, like a Facebook friend, and a Twitter for like, mutual follow might not be the exact same thing as a real life friend. But we can get that data a whole lot easier. And so it's led to a lot of research on things like the sorts of following patterns or friending patterns. Historically, Facebook's been less, you've been less hard to get personal data on Facebook. But for Twitter, like any researcher, basically now can go and download anyone's publicly available tweets from going back since if you're an academic, you go back as far as they have data, you can get their entire network of who they're following into, they're not following. It's led to this plethora of research on these networks, though, as much as I think it's interesting, it has sometimes been hard to figure out what this means on a broader scale, right? Twitter is unique, Facebook is unique. That's actually one of the things in that paper that we did that was a little bit different, or a different paper associated with local parties is that we collected Facebook, Twitter, and Instagram data as well as website presence for these local parties. And so we're able to look at one of the other papers just like, where are they actually online? And so we look to see if like, are there differences across differences like social medias? Right? And we generally found that, you know, if a party's if, if a player is more likely to be on Facebook, they're also more likely to be on Twitter and all these other things. So it's, it's kind of reassuring that there isn't sort of like I like being unique, not unique. I can't think of the word. They're not like specializing in any way. And we can learn from one take it to the other, but it is your social media that has led to some theoretical questions. given us a whole lot more data we can use, but it's not as wonderful as we hoped, because we still have to figure out like, what, what this all means in the end.
Rosemary Pennington
Kevin, as you've been talking, you've been sort of providing a lot of caveats, you know, in this conversation, which is not to say that it's a negative, but I wonder like, you know, journalists can sometimes we'll have to simplify things, right. Like, that's part of the job is to simplify and understand and often don't have space for caveats. And I sort of wonder, as we are going into this midterm election, and then as soon as the midterm election is over, we're going to be in the run up to that next general election, right? Cuz it just starts immediately, what advice would you give to people who are going to be covering news stories about elections to help them sort of, perhaps produce more more nuanced reporting based on polling data or data, like what you've produced? around various things?
Kevin Reuning
That's an interesting question. And so and this is hard to say, in this day and age, but this day and age, it sounds like I'm so old. Things aren't gonna change as much as anyone expects them to or anyone fears. I almost want to say, hopefully, at the end of that, assuming that our democratic institutions remain, things won't change that much. Because in the end, when we talk about campaigns, there's so much focus on things like, what candidates are promising, and how they're engaging with people. But in the end, they're one person among many that are making no decisions. And as we've learned with, you know, Biden, Trump, Obama, and George W. Bush have helped push the president to do more and more, but the President can only do so much in our system. And as long as we have the Senate as it is with, you know, basically needing 60 votes to do anything representing states, the Senate is not going to do that much. And so, like, I'm not saying all this is pointless, because, you know, go vote, I'm gonna vote, I think it's important to vote, I think it's part of people who engage. But the changes that are happening are going to be relatively minor. I was talking to my students about this actually, a few weeks ago, presidents really can do one big thing. In their four years, Biden might have actually gotten away with like one and a half thing, which is interesting. But it's not as you know, there's, there's so much I think, like, wrapped around this, because that's what campaigns need to do. They need to get people engaged, they need to get them to think like, Oh, we're gonna go change the world, because that's how you get people to be motivated. But at the end of the day, like, that just isn't what usually happens. Then the other. The other thing like, this is probably broader advice than just like, you know, campaigns, elections or politics, but like, the simple stories are often missing a whole lot. Which, again, is something I think my students find annoying about my classes. We talk about how I do a class on social moves and protests, social movements are incredibly complex phenomena, there is no easy answer to why a protest appears when it does. We can point to certain factors, but I can tell you a lot of reasons why that factor isn't the only explanation. And so we often want to say, well, this is the cause. There's often lots of causes. And we're gonna see this after the election, what will probably happen is there will probably be some, some reports showing that this certain voter group didn't turn up in certain amounts, and that that's wise, you know, Democrats, Republicans did better or worse, right? If you cut the data enough, you can always find those demographic groups, because, you know, there's enough people, there's enough ways of cutting things, that there will be multiple ways that people didn't show up. And so I would just, I would always think about, like, what, what else could be causing what you're seeing whenever I'm talking about these things, what other factors might play into it that matter?
Rosemary Pennington
That's all the time we have for this episode of Stats and Stories. Kevin, thank you so much for joining us today.
Kevin Reuning Thank you. It's a lot of fun.
Rosemary Pennington Stats and Stories is a partnership between Miami University's Department of Statistics and media journalism and film and the American Statistical Association. You can follow us on Twitter @statsandstories, Apple podcast, or other places where you find podcasts. If you'd like to share your thoughts on the program, send your email to statsandstories@miami.oh.edu or check us out at statsandstories.net and be sure to listen for future editions of Stats and Stories, where we discuss the statistics behind the stories and the stories behind the statistics.