Alleged Bias in Trending on Facebook: Facts and Ideology

Alleged Bias in Trending on Facebook: Facts and Ideology

Last Friday, I wrote about allegations of ideological bias in Facebook’s Trending feature. I focused then on questions of law and of possible impact.

Today, I want to take a closer look at the question of ideological bias and its relationship to perceived source credibility.

Those are loaded words, suitable for a college classroom. Let’s break them down a bit.

We throw around the word “ideology” quite casually, often meaning something more like “philosophy.”

A philosophy is a way of thinking, a way of approaching certain subjects or all subjects. We know what our philosophy is, whether we call it a philosophy or not. If pressed, we can give some form of articulation to it.  It’s our consciously accessible approach to a subject, or the world, or life itself.

Ideology is something deeper, more encompassing and typically less consciously understood. Our philosophy or philosophies are embedded in it and supported by it, but it is wider and deeper and more “given” than they are. We can modify or even reject a philosophy without touching its ideological foundation.

A manifestation of a widely shared ideology is the concept of privacy.

For most of us living in the U.S., the notion that somethings are private . . . well, that’s a given, isn’t it? Of course my personal finances are private. Of course my relationships are a private matter. Of course my thoughts and dreams are private. Nobody has a right to those except me.

Of course.

Except . . . that’s actually not a given. It may be the law (at least for the most part) in this country, and it may be what most if not all of us want to have accepted, but that doesn’t mean that it is a given.

I’m not referring to the familiar question of just how far privacy should extend. I’m referring to the claim that there is something called privacy, or that the thing called privacy is clearly something that should be protected.

There have been societies, especially smaller ones, in which little if anything related to wealth was private. Similarly, there have been societies in which even intimate elements of interpersonal relationships were expected to be known by all. There even have been societies (think of the Cultural Revolution under Mao, or the “re-education” efforts in the Soviet Union under Stalin, or the Reverend Jim Jones’ Jonestown cult) where thoughts and dreams were understood to belong to all.

I’m not going so far as to say that any of these were “right” in their approach to the question of privacy. Ideologies can exist that are profoundly dysfunctional or immoral. And people can believe them, live by them, not even realize how they are being guided by them, even though they are dysfunctional or immoral.

The “problem” with ideology is two-fold. First, we all are ideological in the sense I’ve just described. We all believe, even to the point of not realizing it’s a distinctive belief, certain things to be true, or right, or good, or even simply the way things are. Those convictions are, in this sense, ideological.

Second, because much of ideology operates below the level of conscious choice, in our basic assumptions and perceptions, we can intend to be totally without bias, and be biased in practice.

Which may be exactly what is happening with Trending.

Facebook acknowledges that human curators have an important role in what “trends.” What they do, FB says, is weed out stories that lack credibility.

What if most of the curators are ideologically inclined (in the sense described above) toward a set of beliefs about the way of the world and of human beings in it? What if that ideological inclination is such that they find the kind of journalism (in subject matter, rules of evidence, and approach to writing) found on conservative-leaning sites to be less credible and more suspect than the kind of journalism found on liberal-leaning sites? What if they look to other sites, sites that follow the “rules” they associate with good journalism, to verify those stories, and don’t find them there (perhaps in part because the reporters and editors at those sites are ideological disposed to reject those stories)?

In such a situation, the curator would exclude a story appearing only on a conservative-leaning site from Trending. The rationale would be credibility, not ideology. But the origin of the judgment, deep down, would be ideological.

The resulting stream in Trending would have a distinct “bias” in terms of sources of stories and, if my scenario is realistic, that would mean that some stories would never appear in Trending, even if they were popular with users.

Furthermore, for Facebook users whose ideology was more in sync with the conservative-leaning sources, the stories that did not appear in Trending would seem to reflect an anti-conservative ideological bias. And, at least in the sense I’m working with, they would be right.

Even if no one set out to be anti-conservative. Even if there wasn’t some corporate directive operating.

Facts and ideology. Not as neatly distinct as we might wish.