Tuesday, June 07, 2005

What Bias? This Bias.

On a link from the marriage and self delusion study from Marginal Revolution, I found this article about framing and perception at Econlog. The article points out how the use of adjectives alone can influence perception:

Economists have heard a fair amount from psychologists about "framing effects." Redescribing your options sometimes changes your choice. Firms would rather advertise the sale of "half-full glasses," than "half-empty glasses," though of course they're the same thing.

Aldert Vrij's book on lying describes a particularly striking example:

Participants saw a film of a traffic accident and then answered questions about the event, including the question 'About how fast were the cars going when they contacted each other?' Other participants received the same information, except that the verb 'contacted' was replaced by either hit, bumped, collided, or smashed. Even though all of the participants saw the same film, the wording of the questions affected their answers. The speed estimates (in miles per hour) were 31, 34, 38, 39, and 41, respectively. . . .

Pretty neat, but there's more. Given a little time, framing effects can engender false memories:
One week later, the participants were asked whether they had seen broken glass at the accident site. Although the correct answer was 'no,' 32% of the participants who were given the 'smashed' condition said that they had. Hence the wording of the question can influence their memory of the incident.

Another example of this: in preparing for trials, many lawyers cultivate a habit of selectively using 'power' words to describe the assault, accident, etc., to sway the jury into accepting their version of the facts. The defense in a fender-bender may call it a "bump" or a "tap" but the plaintiff will use words like "crash." A caveat: overusing these words could tip the argument into overkill/hysteria and undercut the credibility of the case.

This theory can also be applied to articles in the media, which use words that even if chosen carefully to be as descriptive and neutral as possible, inevitably imply a placement of the subject on a sliding scale of seriousness or degree. A brief excerpt from this chapter from a book on Cultural Selection outlines some of what I learned back in journalism school regarding media and belief systems:

Media scientists have often discussed how much influence the media have on people's opinions. People tend to selectively read what they already agree with and to rationalize their preformed opinions in the face of contrary arguments. Experimental evidence seems to indicate that the mass media have little power to change people's opinions on issues for which they already have formed a strong opinion, but they have a profound influence when it comes to setting the agenda and priming people on new issues. The way an issue is framed determines how it is discussed, which causes a social problem is blamed on, and which of the possible remedies are entered into the discussion (Sasson 1995, Beckett 1994, Pan & Kosicki 1993, Wanta & Hu 1993, Iyengar 1991, Nelson 1984, Howitt 1982, Weimann & Winn 1994).


Given all this, it's natural that two people from two ends of the political spectrum can watch/read the same media and each see a bias in favor of the opposition. First, each naturally tends to rationalize the stories to their respective perceptions of how the world works. They want to make the story fit into their frame of reference, not continually reinvent their world view to accomodate each new minor event or occurrence. Second, each naturally tends to "read through" the framing that concurs with the world view that they hold. If you feel that a particular event was "tragic," for example, seeing that word used in conjunction with it doesn't register as being odd or out of place. But someone who felt the event was not tragic would focus on the word as demonstrating bias.

My theory: if both sides of the political spectrum are unhappy because they feel like they've sacrificed too much to the opposition, it's likely that something close to equilibrium has been reached.

No comments: