Halo effect

The confident candidate

A job candidate is articulate, confident and well dressed. Afterward, the interviewer rates them highly on technical skill, reliability and leadership.

The interview didn’t test any of those three. The ratings rose together because one strong impression was doing all the rating.

The likable lecturer

A student enjoys a lecturer’s jokes and energy. On the end-of-term evaluation, they also give top marks for “grades fairly”, “chose a good textbook” and “course was well organized”.

The student has direct evidence about the lecturer’s humor. Grading fairness would need a comparison with how others were graded, which a single student rarely has. The warm impression fills in the answers that the student can’t actually assess. Teaching evaluations are one of the settings where researchers have repeatedly found halo (see Evidence).

The rude host

A restaurant’s host is curt and makes a couple wait without explanation. Later, they describe the food as bland and overpriced, although other diners that night rated the same dishes highly.

This is the negative direction, often called the horn effect. The irritation with one person spreads to a judgment (the food) that the host had nothing to do with.

• The attractiveness halo (“what is beautiful is good”). In a 1972 study, people shown photographs judged the more attractive people to have more desirable personalities and expected them to have more successful careers and marriages.
• The horn effect: the same thing in the negative direction.
• Brand and product halos: a well-liked product raises opinions of the rest of a company’s products, including ones the buyer knows nothing about.

• When the traits really do go together. Someone careful about one kind of work is often careful about others; a well-run company often makes several good products. When ratings move together because the underlying qualities do, the judgment is accurate. Researchers call this true halo, as opposed to illusory halo, where ratings correlate more strongly than the traits themselves.
• When each trait was actually observed. Rating several qualities highly isn’t a halo if you have evidence for each one.
• When a general impression is all you have. Using it is reasonable, as long as you hold it loosely and don’t let it override specific evidence.

The test: did you observe this trait, or infer it from a different one?

A track record in the same kind of work

A contractor built a neighbor’s extension on time and to a high standard, and three earlier clients say the same. You expect the contractor to handle your kitchen renovation well.

The good impression isn’t spilling over from an unrelated trait. It comes from repeated direct evidence about the same quality you’re predicting: doing building work well. That’s inference from a track record, not a halo.

Rating what you saw

After an appointment, a patient rates their doctor as warm, a good listener and clear at explaining the treatment plan.

All three ratings are high, but all three were observed during the appointment. It would be a halo if the same warmth led the patient to rate the doctor’s diagnostic accuracy highly too, which a single appointment can’t show.

Judging many traits one at a time is effortful; forming one overall impression is fast and automatic. Daniel Kahneman explains the halo effect through “what you see is all there is”: the mind builds the most coherent story it can from the information in front of it, and a coherent story about a person is one where their qualities agree with each other.

Order matters too. What you learn first sets the impression that later information is read through, so the same facts can produce different judgments depending on which came first.

Status: replicates robustly. The halo effect is one of the longest-standing findings in psychology.

• Thorndike (1920) named it after noticing that military officers’ ratings of soldiers on separate qualities (physical, intellectual, leadership and personal) correlated far more strongly than those qualities plausibly could.
• Nisbett and Wilson (1977) showed students one of two videotaped interviews with the same instructor, who spoke with a European accent. In one he was warm; in the other, cold. Students who saw him warm rated his appearance, mannerisms and accent as appealing; those who saw him cold rated the same features as irritating. They were also unaware of the influence. Many of those in the cold condition believed their dislike of his features had lowered their overall opinion of him, when the experiment showed it worked the other way round.
• Cooper (1981) reviewed the literature and found halo across performance appraisals, interviews, and peer and teaching evaluations. He introduced the true/illusory distinction.

The effect itself is not in doubt. What’s harder to pin down in any real-world rating is how much of the correlation is illusory, because that requires independent measures of the traits being rated.