Types of audience fragmentation

 I'm embarking on a new large-scale project relating to audience fragmentation. Or rather, I have been embarking on it for the past year - such is the leisurely pace of the post-tenure research agenda. It started as a refutation of the echo chamber as an intuitive but overly simplistic characterization of audiences' media diets in the age of information abundance. Then I realized that someone already wrote that book. 

In researching the idea, I was surprised to find how few studies about fragmenting audiences and echo chambers even tried to capture what I felt was the right kind of data: data capturing the whole of people's media diets - not aggregate audience data, not what individual users post on a particular platform, not even the amount of time or what individuals see on a particular platform, but ALL of what they see across all platforms and media. Unless you capture that, you really have no way of knowing whether individuals have any overlap with one another in what content they consume and/or how many of them are sequestering themselves in ideologically polarized echo chambers. 

In defense of researchers, this is a hard kind of data to get. What media content people consume is often a private matter. It's just hard - for an academic researcher, a company, a government - to get people to trust them enough to get that data. Observing people might cause them to change their behavior. Still, some researchers have made in-roads - working with representative samples, trying to get precise, granular data on precisely what content people are seeing - and I think that if we start to piece together what they have gathered and supplement it with new data, we'll be able to get a better sense of what audience fragmentation actually looks like. 

I've started the process of collecting that data. In a survey, I've asked a sample of college students to post URLs of the last 10 TikTok videos they watched, the last 10 YouTube videos they watched, and the last 10 streaming or TV shows they watched. I'm anticipating that there will be more overlap in the TV data than in the YouTube of TikTok data. But I wonder what counts as meaningful when it comes to overlap or fragmentation. I return to the age-old question: so what?

Let's say you have a group of 100 people. In one scenario, 50 of them watch NFL highlight videos, 25 watch far-right propaganda videos, and 25 watch far-left propaganda videos. In another scenario, all 100 of them watch 100 different videos about knitting. The latter audience, as a whole, is more fragmented than the former audience. The former is more polarized in terms of the content it consumes - half of the sample can be said to occupy echo chambers, either on the right or left. 

It's clear to me while the polarization of media diets matters - it likely drives political violence, instability, etc. But why does fragmentation, in and of itself, matter? 

I guess one fear is that we will no longer have any common experiences, and that will make it harder to feel like we all live in the same society - not as bad as being ideologically polarized, but it's plausible to think that it might lead to a lack of empathy or understanding. But what counts as a common experience? Do we have to have consumed the same media text? Stuart Hall would tell you, in case you didn't already know, that different people watching the same TV episode can process it in different ways, leading to different outcomes. But at least there would be some common ground or experience. 

But what if we watched the same genre of television show, or watched the same type of video (e.g., videos about knitting)? If we contrast the 100 people who all watched different knitting videos to 100 people who all watched videos about 100 very different topics (e.g., knitting, fistfights, European history, coding, basketball highlights, lifestyle porn, etc.), I would think that the former group would have more to talk about - more common ground and experience - than the latter, despite the fact that there is an equal amount of overlap (which is to say, no overlap) in terms of each discrete video they watched. 

Instead of just looking at fragmentation across discrete texts, it would also be useful to look at it across genres or types. It could get tricky determining what qualifies as a meaningful genre or TikTok video. Some TikTok videos share a set of aesthetic conventions but may not convey the same set of values, or vice versa. There will be some similarities across the texts in people's media diets, even if there is no overlap in the discrete texts. The challenge now is to decide what similarities are meaningful.