Archive for August, 2008

Shared and Private Taste in Evaluations of Beauty

August 13, 2008 Leave a comment

The 2000-year-old adage, “Beauty is in the eye of the beholder” is roundly discounted in much of the contemporary literature on the psychology of beauty. There is consistent agreement on the universal human preference for features like averageness, symmetry, and sex-specific characteristics. Johannes Hönekopp, in a 2006 paper, argues that private taste (the eye of the beholder component) is being wrongly de-emphasized and that it is about as influential as shared taste (universal preferences component).

Simplifying a bit, Hönekopp approaches this topic this way. First, he had the same subjects rate the same photographs of faces twice. He then calculates what he calls “private taste:”

“The impact of private taste on judgments is reflected in the variance component for the interaction between faces and judges. To see why, imagine that the faces of Peter, Paul, and Mary receive face scores of 3, 4, and 5, respectively. Assume that Anne repeatedly rates these faces with 3, 4, and 7, respectively. One concludes that Anne’s preferences are somewhat different from the average in that she regards Mary as more attractive than the average judge does. Thus, an interaction effect between judge and face indicates private taste (Kenny, 1994).” 

He then calculates “shared taste:”  

“The strength of shared taste is reflected in the variance component for faces relative to overall variance. The more judges agree, the larger the fraction of variance attributable to faces.” 

From these, Hönekopp calculates the beholder index (bi) – the variance attributable to private taste divided by the variance attributable to private taste added to the variance attributed to shared taste. In this specific case:

“where varcomp J x F denotes the estimated variance component for the Judge x Face interaction, and varcomp
F denotes the estimated variance component for faces.”

Judges of and subjects in the photographs were found via opportunity sampling on or around college campuses in Germany. The data, including a reanalysis of homogenous faces in the (2nd and 3rd quartiles) and the heterogenous faces (1st and 4th quartiles) are shown in this figure:


As the black bars in the above figure illustrate, Hönekopp’s data show that shared taste and private taste explain about the same amount of variance. That is, the idiosyncratic preferences of the raters were as important as the characteristics of the faces themselves in evaluating relative attractiveness – when all of the faces in the stimulus set were included. When the range of attractiveness is reduced (the homogenous faces), private taste explains most of the variance. When the range of attractiveness is emphasized using the first and fourth quartile faces (the heterogenous faces), private taste becomes less influential.  

To illustrate, imagine asking subjects to rate the attractiveness of photos of Matt Damon, George Clooney, Matthew McConaughey, Jude Law, and Johnny Depp. Given that each of these celebrities has been a recent recipient of People Magazine’s “Sexiest Man Alive” title, it is likely that rater idiosyncratic taste will play a far more significant role in relative ratings than any of the differing qualities of the faces themselves. This thought experiment illustrates what happens when the range of attractiveness is restricted: idiosyncratic preferences in the raters themselves will become the most significant source of the variance in the attractiveness ratings. To illustrate the other extreme, imagine asking subjects to rate the attractiveness of the photos at this web site. This small sample set, with such extreme differences in facial characteristics, will likely produce very little evidence of private taste. [For a brief presentation of some research using this set of photographs and reaching this conclusion, see this video.]  

The conclusion – at least about methodology – is that as the range of facial characteristics relevant to attractiveness ratings is reduced, the influence of individual taste/preference will be increased.


Hönekopp, J, (2006).  Once more: Is beauty in the eye of the beholder?  Relative contributions of private and shared taste to judgments of facial attractiveness.  Journal of Experimental Psychology: Human Perception and Performance. 32 (2), 199-209.

Categories: Methodology, The Face