WPCRESEARCH

Why Can’t We Be Friends?

Diversity preference and social media
We all know the drill. Facebook suggests we add people we don’t know to our friends list. We look them over, deciding if we’re interested. However, what Facebook is interested in is building its business.

“From a profitability point of view, the platform grows the network with new connections and new inclusion of users, so advertisement propagation will be more effective,” says Olivia Liu Sheng, professor and W. P. Carey Distinguished Chair in Business in the Department of Information Systems.

The more people connect, the more time they spend on the platform, which makes them more likely to respond to advertising. Facebook mostly matches us with people like us. But is that all we’re looking for in people—a mirror of our interests and background?

The concept of diversity preference suggests otherwise. “Diversity preference is that you want friends with different backgrounds,” says Sheng. “I like my colleagues at ASU, but I want to have other kinds of friends, too.”

Olivia Liu Sheng wearing a W.P. Carey polo shirt and standing with her arms crossed

Individualization is a hit

But is diversity as simple as a yes/no preference? To explore this question, Sheng co-authored a study titled “Diversity Preference-Aware Link Recommendation for Online Social Networks.”

“Previous studies looked at diversity in a myopic manner,” says Sheng, “assuming everyone wants the same level of diversity. Either everyone wants more or less diversity” in their friend recommendations.

This study looked at individualized preference and found that when the diversity of the people recommended was modified for each person, the likelihood they would accept a friend recommendation increased—or, in technical terms, the hit-rate percentage was higher.

Of course, these kinds of online link-ups have to be reciprocal. “You want to connect to that person and vice versa,” says Sheng. “So it’s not one-sided.” The trick is to pair people who want the same level of diversity in their network of online friends.

Calculating individualized preference for diversity is much more complex than simply a yes/no preference. To analyze personal inclination toward diversity, the study had to transform the problem into a decision model, which could assess the decisions needed to optimize the process, given specific criteria.

To gain a broad perspective, the study used two data sets. One came from the gone-and-mostly-forgotten Google Plus social network, which is readily available to researchers; the other is a proprietary one gleaned from a global social media network that isn’t named.

Diversity preference is that you want friends with different backgrounds. I like my colleagues at ASU, but I want to have other kinds of friends, too.
—Olivia Liu Sheng,
Professor of Information Systems and W. P. Carey Distinguished Chair in Business

The profit in diversity

Critics lament that social media has divided us into narrow tribes that don’t want to hear perspectives different from our own. This study confirms that people aren’t quite so homogeneous—even when it comes to how diversified they want their networks to be.

So, customer satisfaction and engagement are at stake for platforms like Facebook. “The more satisfied you are, the more you use the platform,” says Sheng.

But what’s at stake for society might be much more significant. If there’s profit in being more flexible about diversifying people’s social media networks, then platforms have a clear motive to encourage at least some people to mix and mingle more. Diversifying social media networks could lead to less of an echo chamber effect online, in which people only hear back from like-minded friends.

Recently, Sheng presented her paper to an audience of information systems researchers at two Taiwanese universities. The researchers showed particular interest in how diversity preferences influence the spread of fake news.Sheng sees that these findings could support better business practices that benefit everyone. The more diversified our networks are, the more diverse our thinking can become, and the less prone people can be to misinformation. As Sheng puts it, “Both social media platforms and users will be better if we consider the heterogenous preference.”

—Joe Bardin