In Quillette, Bo Winegard explains how to define bias:
Bias is an important concept both inside and outside of academia. Despite this, it is remarkably difficult to define or to measure. And many, perhaps all, studies of it are susceptible to reasonable objections from some framework of normative reasoning or another. Nevertheless, in common discourse the term is easy enough to understand. Bias is a preference or commitment that impels a person away from impartiality. If Sally is a fervid fan of the New York Knicks and uses different criteria for assessing fouls against them than against their opponents, then we would say that she is biased.
There are many kinds of biases, and bias can penetrate the cognitive process from start to finish and anywhere between. It can lead to selective exposure, whereby people preferentially seek material that favors their preferred position, and avoid material that contradicts it; it can lead to motivated skepticism, whereby people are more critical of material that opposes their preferred position than of material that supports it; and it can lead to motivated credulity, whereby people assimilate information that supports their preferred position more easily and rapidly than information that contradicts it. Often, these biases all work together.
So, imagine Sally the average ardent progressive. She probably exposes herself chiefly to progressive magazines, news outlets, and friends; and, quite possibly, she inhabits a workplace surrounded by other progressives (selective exposure). Furthermore, when she is exposed to conservative arguments or articles, she is probably extremely critical of them. That National Review article she read this morning about abortion, for example, was insultingly obtuse and only confirmed her opinion that conservatives are cognitively challenged (motivated skepticism). Compounding this, she is equally ready to praise and absorb arguments and articles in progressive magazines (motivated credulity). Just this afternoon, for example, she read a compelling takedown of the Republican tax cuts in Mother Jones which strengthened her intuition that conservatism is an intellectual and moral dead end. (This example would work equally well with an average ardent conservative). The result is an inevitably blinkered world view.
The strength of one’s bias is influenced by many factors, but, for simplicity, we can break these factors into three broad categories: clarity, accuracy concerns, and extraneous concerns. Clarity refers to how ambiguous a topic is. The more ambiguous, the lower the clarity and the higher the bias. So, the score of a basketball game has very high clarity, whereas an individual foul call may have very low clarity. Accuracy concerns refer to how desirous an individual is to know the truth. The higher the concern, on average, the lower the bias. If a fervid New York Knicks fan were also a referee in training who really wanted to get foul calls right, then she would probably have lower bias than the average impassioned fan. Last, extraneous concerns refer to any concerns (save accuracy concerns) that motivate a person toward a certain answer. Probably the most powerful of these are group affiliation and status, but there are many others (self-esteem et cetera).
At risk of simplification, we might say that bias can be represented by an equation such that extraneous concerns (E) minus (accuracy concerns (A) plus clarity (C)) equals bias: (E – (A + C) = B).
