Study challenges idea that we prefer faces we've seen more often

The more often you're exposed to something, the more you like it. This is the well-established 'mere-exposure effect', which has been found to apply to everything from sounds to paintings.

It occurred to Jason Chow at Vanderbilt University and colleagues that this effect might be exploited to help older people to settle more quickly in a care home; repeatedly showing a new resident photos of images of caretakers and fellow residents would hopefully allow them to adapt more easily.

Based on other research, the team suspected that a mere-exposure effect for faces might actually be stronger in older people than in younger people. So they decided to first run a study to explore this. But as they report in a paper in Psychology and Aging, to their surprise, they found not only no evidence for this effect in older people, but no evidence for it in younger people, either. For the researchers, this raises some serious questions, even: does the mere-exposure effect really exist?

Chow and colleagues recruited 50 participants aged 18-35 and also 50 about 65-80 for each of four very similar experiments. In all the experiments, the team used the same images of 48 neutral faces of men and women. Half of these faces fell into the same age range as the younger participants, half in roughly the same range as the older participants. Each participant first viewed 24 of the faces individually, with a varying number of repetitions. Each time, they were asked to identify the gender of the person; this was to ensure that they were paying attention and processing the face. After a 5 to 10 minute gap, they were then shown all 48 faces, one at a time; each time, they had to rate how much they thought they 'would like' that person on a scale of 1 - 5.

The team expected to find higher liking ratings for the faces that the participants had seen more often. But this did not happen. In fact, though they report following methods and procedures used in experiments that have found mere-exposure effects, in this case there was "substantial evidence against a main effect of repetition on liking ratings in each experiment."

Why might this be? The team ruled out a few possibilities. Some people had done poorly on the gender identification, so may not have paid proper attention to the photos – but the team found the same results even when excluding these participants. Similarly, the researchers found no evidence that the pattern of results was influenced by whether or not a participant had done the tasks at the time of day when they tend to be most alert.

But they did find something that has been reported before: both age groups liked the younger faces more than the older faces (and this was especially true for the younger adults). This finding shows, they argue, that their procedures could detect an experimental effect — so the non-replication of the mere-exposure effect "is unlikely due to an error in the task or an anomaly in the participants."

So, what is it due to? The least palatable suggestion, as far as researchers in the field would be concerned, is the 'decline effect'. As Chow and his colleagues note, some effect sizes in studies (not just in psychology but also medicine and biology) have faded or even vanished as experimental data has accumulated. "Often, effects seem to decline in magnitude because testing in broader populations reveals nuances not observed in the original work," the team notes. This can mean that an initially clear effect eventually fades into background noise.

Some other recent studies of the mere-exposure effect have produced mixed results, the team notes. This could fit with the idea of a decline effect.

But it may be that faces, which carry such social importance, are more resistant to the mere-exposure effect — at least, in the lab — than other stimuli, such as songs or pictures. The researchers note that people taking part in a study might be reluctant to indicate that they 'dislike' or even 'like' a person based solely on their face — while they probably wouldn't feel the same reluctance to judge a song or an image in this way. If this is the case, though, it might not necessarily be what happens in the real world.

The team was "disappointed" not to find a mere-exposure effect. But, they write: "we hope that others see its absence as an opportunity to better understand the efficacy and incidence of the mere-exposure effect and, accordingly, develop a sharper understanding of how repetition and familiarity influence liking of other people's faces."