The halo effect

We like to assume that we are - as individuals - pretty unbiased, and that we can make truly independent and objective judgements. Hah! Sure we can. Enter Thorndike's "halo effect" theory (1920). The Halo Effect is where we experience pretty much a "constant error" of judgement, where we see another as the sum of their qualities, and may make an overarching judgment based on just a few characteristics - such as if they appear very bright; very able; are consistently late; or look unattractive (Nisbett & Wilson, 1977; Thorndike, 1920, 1982). We apply our general judgment of a person to a specific trait that we see (Nisbett & Wilson, 1977).

So if someone turns up late the first time we meet them, we keep referring to them as "The Late Mr X", or if someone picks up an error in a problem, we continue to think of them as having a great eye for detail (Nisbett & Wilson, 1977). We are all the sum of our attributes, skills, influences, experiences, and prejudices, and need to have done one hang of a lot of personal, reflective work in order to be even remotely unbiased: and, as we cannot get inside anyone else's heads, we make these sweeping judgements about others instead of continually re-evaluating them 'objectively'. 

Even competent managers find it difficult to 'see' us as a collection of individual qualities, instead clustering our skills into groups of traits which are not that accurate (Thorndike, 1920). A further complicating factor is that the halo effect process is "subterranean", a black-box process (Nisbett & Wilson, 1977, p. 252; Von Bertalanffy, 1968) like AI, so we cannot see it to interrogate it, and make sense of it. Because of this, we tend to be largely unaware that our biases affect our ratings of others (Nisbett & Wilson, 1977). Students watching videoed interviews of an instructor modelling a "bright and pleasant" or "cold and distrustful" demeanour were analysed (Nisbett & Wilson, 1977, p. 252). The students viewing the 'warm' instructor rated their physical appearance, mannerisms, and accent as appealing; those with the 'cold' instructor rated those same aspects as irritating... despite all students having viewed the same instructor. The students were not only unaware of their own biases, but also seemed to think that the influence ran in the opposite direction (Nisbett & Wilson, 1977). 

The halo effect has implications for a wide range of equity and parity processes, including blind reviews, avoiding conflicts of interest, ensuring all applicants are measured to the same standards (to avoid nepotism and cronyism), pay rating scales, and objective measures for performance reviews (Nisbett & Wilson, 1977; Thorndike, 1920, 1982). 

It also means that counts and other descriptive statistics are a necessary support when gathering qualitative evidence to help to avoid subjective ratings; and that whenever possible we should use a range of different qualities which can be judged as independently as we can (1920).

Sam

References:

Thorndike, E. L. (1920). A Constant Error in Psychological Ratings. Journal of Applied Psychology, 4(1), 25–29. https://doi.org/10.1037/h0071663

Thorndike, R. L. (1982). Applied Psychometrics. Houghton Mifflin.

Nisbett, R. E., & Wilson, T. D. (1977). The halo effect: Evidence for unconscious alteration of judgments. Journal of Personality and Social Psychology, 35(4), 250-256. https://doi.org/10.1037/0022-3514.35.4.250

Von Bertalanffy, L. (1968). General System Theory: Foundations, Development, Applications. George Braziller.