Tuesday, January 19, 2016

Perception Becoming Reality: The Effects of Framing Polls and Early Primary Election Results on Perceived Electability and Voting Behavior

National polls (and, in the coming weeks, the results of early primaries) present potentially misleading information about presidential primary candidates' chances of winning the eventual nomination. The actual likelihood depends on a several facets of the primary electoral process: how many delegates are assigned by the voters of each state; whether or not a state is winner-take-all; "triggers" and "thresholds" to allocate delegates to particular candidates; when a given state votes during the process. Add to that the effect of whether or not other candidates drop out of the race and who those voters then decide to vote for.

A lot of this can, and has, been modeled. You can model how many people would vote for each candidate in each state (even if there isn't accurate polling data in some states) based on what you know about the relationship between, say, education and likelihood to support a particular candidate. You can know who each voter's second, third, or fourth choice would likely be (i.e., how things will shake out when candidates start dropping out of the race). You can know what the rules are for delegate allocation in each state and how many delegates are in each state. When you take all of this into account, at least for the Republican candidates right now, you end up with a disjuncture between what the polls and what the early primary results will likely be (Trump and Cruz well ahead of Rubio) and who would actually get the most delegates if the primaries were to all be conducted today (Rubio, probably).

The crazy thing about this is that the emphasis on current national polls and early primary results in the  media (which, as far as I'm concerned, is a misleading picture of how people would vote if the primaries were all held today) might change later primary voters' perceptions of the electability of their favored candidate, causing them to abandon that candidate and switch to another one.

Surely, there will be some people voting in later polls who will "stand their ground" and still vote for their favored candidates, regardless of what national polls or early primary elections say. Also, there are many reasons why those voting in later primaries may change their opinion over the coming months: for example they may get more information about the candidates, or their favored candidate may say or do something they don't like. But I think at least one possible cause of switching candidates has to do with perceived electability, and that perceived electability could be based on the misleading information from national polls and early primary results.

So then, how will the misleading information sway voters?

My guess is that Trump and Sanders (and possibly Cruz) will keep referring to the polls and the early primary results, claiming it to be evidence of their electability. They would do this in hopes of a herding effect. For Republicans, people in late-voting states who would've voted for Rubio will see supporting Rubio as supporting a likely loser. Spending time and energy supporting him would be a waste, and possibly embarrassing. This would cause them to abandon Rubio and either fall in line with the herd developing around Trump and/or Cruz (likely due to an "anyone but Hillary" sentiment) or sit out the primary vote altogether. For the Democrats, Hillary supporters residing in late-voting states who were on the fence and perhaps supported Hillary only because they thought Bernie didn't have a shot would think that Bernie did have a shot, and switch over to Bernie.

However, this strategy of emphasizing national polls and early primaries might backfire for Trump. He'll keep saying he's winning and will successfully convince people he's likely to win the nomination, but this might freak other voters out ("oh my god, he could actually win!"). This might cause people who would have sat on the sidelines to vote against him. It might cause wealthy donors to throw more money at Cruz or Rubio. It might cause other candidates to drop out sooner and endorse Cruz or Rubio. Call this the "panic mode" reaction to the perception that Trump could win.

There are, of course, many X factors that could swing the election: the economy tanks, someone says something stupid, scandals, terrorist attacks, etc. But I think one factor is whether people think national polls and early primary results predict eventual electability. And whether people think this depends on what they hear both from the candidates themselves and from the news.

The news will likely present a "horse-race" framing of the election, not because they want Trump or Cruz or Sanders to win, but because they want a close race, because it's a simpler story, and because this will boost ratings. There is a chance that some news outlets (I'm looking at you, NPR and NYTimes) will try to convey the complex relationship between staggered primaries with various delegate allocation rules and public opinion. I think the likelihood of any of the above scenarios playing out depends on whether news outlets use the simple, misleading frame or the more nuanced one.


Saturday, January 02, 2016

The Awkwardness of Walking a High School Hallway (or, Digital Tribes: Gamers, Socialites, and Information Seekers)

This thought came to me while reading this New York Times article on app makers' attempts to understand how teens use smartphones and what they want out of the experience. In particular, I was struck by this sentence: "And when your phone is the default security blanket for enduring the awkwardness of walking a high school hallway, it feels nice to have a bunch of digital hellos ready with a swipe."

I thought of my own experience in high school. Indeed, it was awkward. I didn't have a phone as a security blanket. I suppose I just thought about the things that mattered to me as a way of escaping the awkwardness. I thought about the video games I'd play when I got home, or the movies or music I loved. Social media didn't exist. Maybe I thought about hanging out with my friends the following weekend.

Also while reading this sentence, I thought of my nephew (age 9) and niece (age 5). They're both too young for social media and smartphones, but I started thinking about what they'd be like when they are old enough to use these things. My nephew is already enamored with video games, in particular Minecraft. It seems unlikely that he'll be a heavy user of social media, and very likely that he'll spend a lot of time playing video games. My niece plays video games, and I honestly am not sure whether she'll stay interested in video games and/or develop an intense interest in social media, like many middle school and high school girls.

But as I read this article, and as I imagined how my nephew and niece would use media when they get to high school, a picture started to emerge in my head, a picture of at least two, maybe three, relatively distinct "tribes". One tribe spent most of their screen time using social media like Instagram or Snapchat. Another tribe spent most of their screen time playing video games. Of course, there would be some overlap: the gamers wouldn't totally forsake social media, and those who spent a lot of time with social media would also play some games. But they would differ in terms of how these media experiences fulfilled some fundamental needs or desires, how digital media provided a kind of default security blanket for them during the awkward teenage years.

For the gamers, video games would deliver a sense of challenge and accomplishment, and sometimes a sense of esteem (others see what you've accomplished and admire you). They also would provide camaraderie via the community of gamers.

For the social media users (let's call them "socialites"), social media would deliver a sense of social support and esteem, evidence that people are paying attention to you, that people like you, that you're not alone.

And perhaps there would be a third group: information seekers/entertainment consumers - people who use media primarily to consume rather than interact; consume news, consume educational material, consume movies, music, etc. I think I was one of these types of people in high school, and I think they still exist in high schools. Some kids aren't that into gaming or social media. They love movies, music, books, etc.

These are distinct groups driven by distinct desires. This brings me back to the Uses & Gratifications theory, a theory that I'm not too fond of (because I don't think people are very good at reflecting on why they use media), but might be of some use to help determine what the positive or negative effects of media might be.

So what? Why do these categories matter?

Well, for light-to-moderate users, all of these types of media use might help to keep young people happy and engaged with the world around them, give them a sense of belonging and fulfillment. The particular kind of media use that provides that sense of belonging and fulfillment won't be the same for everyone.

What about heavy media use? Well, heavy use is probably bad for all groups, but bad in different ways. For those in the social tribe, heavy use would be associated with a kind of fragile ego and need for validation from others, and preoccupation with this validation. For gamers, heavy use would be associated with not caring about accomplishments in the real, non-game world (i.e., not caring about grades, not caring about social connections with real-world peers, not caring about one's health, etc.), a kind of disappearing into the game world. For information seekers, heavy use might be associated with a kind of "filter bubble" problem: they get further and further into a particular view of the world without being forced to see messages from other perspectives or without interacting with people who, inevitably, will hold at least slightly different opinions.

If you just measure "internet use" or "smartphone use" as they relate to these outcomes, you might not find any effects, simply because the lack of effects in the other two groups "washes out" the significant effect in a single group. That doesn't mean the effects aren't there. By differentiating among these tribes (not necessarily by asking young people how they identify, but by measuring their actual use of video games, social media, and information/entertainment consumption), we would be able to see these different effects.

I'm usually quite skeptical of metaphors used to describe media technologies. Such metaphors tend to highlight the ways in which media technologies are similar to something else while ignoring all the ways in which it is not, and they seem chosen chiefly to support the pre-existing beliefs of the metaphor user. Do you think that a new media technology is harmful? Liken it to cigarettes or crack cocaine. Think that it's benign, or even helpful? Liken it to chess or painting or family.

But the security blanket metaphor seems a bit less...deterministic than these other metaphors. Of course, if one were to take it literally, it does have a negative connotation: the image of teenagers clinging to blankies evokes a kind of pathological arrested development, kind of like a pacifier. But what I like about the metaphor is that as long as you don't take it too literally, it helps you think about what young media users get out of the experience - security, comfort - and, at least for me, it doesn't dictate that this gratification come from a particular type of media experience.