Halo and Horn Effect

What the halo and horn effect is

The halo effect is a single positive impression of a person carrying over into the assessment of all their other qualities. Its opposite is the horn effect, in which a single negative quality colours everything else about the person. The horn effect is also known as the horns effect.

In both directions, one observation stops being a judgement about one quality. A favourable impression of one thing the person does well, or an unfavourable impression of one thing they do badly, becomes the basis for judging the rest of what they do. The other qualities are then assessed less on what the rater has seen of each of them and more on the impression left by the first. In the halo effect that impression lifts the other ratings; in the horn effect it pulls them down.

In performance appraisal, the halo effect is one of the patterns that our definition of rater bias groups together, alongside the recency effect in appraisal and central tendency error. Rater bias is the wider term. It covers any systematic tendency in appraisal that makes a rating reflect the rater or the method instead of the employee’s performance. The halo and horn effect names one particular route by which a judgement goes wrong: a single quality spreading across the whole assessment. It is also not confined to appraisal, because it appears in interviews as well, as the first case below shows.

How the halo and horn effect appears in appraisals and interviews

The halo and horn effect shows itself in cases such as these, two in the positive direction and one in the negative:

• In an interview. A candidate who speaks fluently is rated highly on analytical skills that were never tested at all. The interviewer heard the fluency, so that part of the rating rests on something observed. Nothing in the interview examined the candidate’s analysis, and yet the score for analysis rose together with the score for fluency. Fluency and analytical skill are two separate qualities, and the interview produced evidence on only one of them.
• In an annual appraisal. An employee who delivered a successful project in the last quarter is rated highly on items that concern the beginning of the year. The project and its success both happened, but they happened at the end of the period. The items on the first months of the year are meant to record a different stretch of time, and the success of the final quarter has been carried back into them. The rating on those early items then describes the end of the year while standing for its beginning.
• In the opposite direction. An employee who arrived late twice at the start of the year has every later note on their work interpreted through “lack of discipline”. This is the horn effect, the negative counterpart of the halo effect. The two late arrivals did happen, but they have turned into a label, and the label then decides how everything recorded about the employee afterwards is understood.

Why the halo effect misleads the rater

The halo effect misleads the rater because the rater does not feel it working. A rater under its influence does not say “I liked this person, so I gave them a higher score.” Instead, the rater finds themselves convinced that the employee is good at everything. To the rater, that conviction looks like an assessment of each quality, not like one impression carried across all of them.

That is why alerting raters to the halo effect is not enough on its own. A rater who does not feel the effect cannot correct for it simply by being told that it exists. Addressing it therefore starts from the design of the appraisal instrument itself.

What limits the halo and horn effect

Because the remedy starts with the instrument, the measures that limit the halo and horn effect are changes to how a rating is collected:

• Rating item by item, with evidence. Each item on the form has to be accompanied by a specific example of the employee’s work. The example is asked for with each item, not once for the whole form. The general impression then has no gap to fill. An item that the rater cannot support with an example is left visibly unsupported, instead of being completed from the rater’s view of the person as a whole.
• Rating by item, not by person. The rater assesses the same item for the whole team before moving on to the next item. In that order, the rater considers one quality across several people at a time, instead of one person across several qualities at a time, so an impression formed on one item for one employee is not carried straight into the next item for the same employee.
• Several raters. Ratings are gathered from more than one rater, as in 360 degree appraisal, where the ratings of one employee come from different sources. It is difficult for a single halo to appear identically across different sources, so a quality that has coloured one rater’s assessment is less likely to colour every assessment in the same way.

How the halo effect differs from implicit bias

The halo effect arises from real information about the person, which is then generalised too far. The rater saw something the person actually did or actually is, as in the three cases above: the fluency in the interview, the project in the last quarter, the two late arrivals at the start of the year. The error lies in extending that observation to qualities it says nothing about.

Implicit bias arises from the person’s membership of a group, not from anything the person has done. It needs no observation of the person to begin with.

The halo effect is therefore a mistaken generalisation from a correct observation, while implicit bias is a judgement made before the observation.

Our definitions of rater bias and first impression error draw the same line between the halo effect and implicit bias.

This is an explanation of the concept, not legal advice.