Research, evaluation, analysis

Looking Beyond the Halo

Let's suppose you've just been introduced to a tall man. I'm not going to suggest that you would be more likely to conclude he was intelligent than you would be if he were short, but research shows that in fact many people are. There is, of course, no known relationship between height and intelligence (I'm 6' 2" myself, so I make this assertion disinterestedly). Similarly, people often conclude that someone is likely to be excessively emotional because she's a woman, or emotionally undeveloped because he's a man.

Leaping to conclusions about people because of a single characteristic of theirs is known in psychology as the halo effect. We know that someone in a painting with a halo must be a saint, and we tend to make similar evaluations of people who aren't in paintings. I'm often amused in my middle age by other leading-edge baby boomers who've decided that the clothing of the contemporary young is a sign of moral or intellectual inadequacy. People who perpetrated the clothing of the 1960s and 1970s should know better.

The halo effect frequently creates serious and unnecessary obstacles to communication and decisionmaking. Management is often made difficult, for example, by the tendency of individual managers to believe that anyone who is an employee must be shiftless and incompetent, and the tendency of individual employees to believe that anyone who is a manager must be incompetent and shiftless. And of course the phenomena of racial, ethnic, and other forms of prejudice are manifestations of the halo effect.

The effect is often unconscious, which makes the problem even more difficult. The basic problem, of course, is slipshod analysis. We are not only assuming relationships of which there is no evidence, we are assuming relationships which often can readily be shown not to exist. Even when there is a known relationship it is rarely strong enough to justify leaping to conclusions about individuals.

For an example of this problem we can look at the article on this site about suicide genes. Even if we accept that the gene mutation isolated by the researchers is more common among suicidal depressives, the fact remains that they still found it in fewer than half their sample of suicidal depressives. You therefore cannot leap to the conclusion that suicidal depression is a genetic problem.

The standard procedures of scientific inference provide a good safeguard against the halo effect. Framing hypotheses, specifying the evidence necessary to disprove them, and then collecting evidence not only offer you the opportunity to determine whether a relationship exists, they also let you determine just how strong the relationship is if it does exist. Of course, real inference does take more time than leaping to conclusions, but it is far less likely to lead to poor decisions.

Looking Beyond the Halo © 2000, John FitzGerald
Home page | Decisionmakers' index | E-mail