Ignoring face trustworthiness may get harder with age

How trustworthy we perceive someone to be at first glance can be based on somewhat arbitrary factors: how attractive someone is, what facial expression they happen to be making, or even how similar their face is to our own, for example.

Age can also affect who we choose to place trust in. Not only do we take cues from how old others look, but some research shows that our own age may tip the scales in how readily we trust. Older people are generally more likely to evaluate faces more positively than young people and, considering that older people are often the targets of financial scams, this may have significant ramifications.

In their recent paper, published in Scientific Reports, Marilyn Horta extend our understanding of the interaction between age and perceptions of trustworthiness. Moving beyond just facial evaluations, the findings of their study suggest that older adults are less likely to identify when a trustworthy looking person is behaving in a non-trustworthy way, potentially leaving them more open to harmful actions perpetrated by seemingly trustworthy figures.

Two groups of participants were recruited for the study: 143 younger adults, aged between 18 and 33, and 129 adults aged 67 or older. All took part in a version of the Iowa Gambling task, which presents participants with the choice of a card from two different decks: one, a 'good' deck, which is more likely to offer higher rewards, and the other, a 'bad' deck, which is more likely to offer penalties. Generally, participants start the task by selecting cards from different decks, before working out which is most likely to give them higher, more frequent rewards. The aim is to end the game with as many points as possible.

Some participants in this study saw the standardised gambling task, which is not associated with faces and has no social element. Others completed a different version, in which each type of deck was paired with faces which signalled different levels of trustworthiness. In congruous conditions, advantageous decks were paired with pictures of trustworthy faces and disadvantageous decks were paired with untrustworthy faces; in incongruous conditions, advantageous decks were paired with pictures of untrustworthy faces and vice versa.

All participants, regardless of condition or age group, were initially more likely to pick a deck paired with a trustworthy face: around 80% of first choices in the two social conditions were for a deck presented with a trustworthy face.

Unsurprisingly, trustworthy faces with advantageous decks were most appealing to participants. Those who saw this combination gathered more points than participants presented with faces that didn't match the nature of the decks. This was the case across both age groups.

However, when looking at data from just the incongruous condition, younger people generally collected more points overall than older people. In the initial stages of the study, participants learned which decks were likely to give them a reward and which a penalty, and performance across all three conditions was fairly similar at this point; yet older participants seemed less able to learn from this uncertain period, and were more likely to continue selecting from the deck associated with a trustworthy face, even when they were disadvantageous.

Applied to real life situations, these findings are pretty concerning. Older members of our communities are much more likely to be targeted in financial scams, many of which are perpetrated through establishing rapport. Being less able to spot when a trustworthy-acting character isn't acting in your best interests, therefore, poses additional risk when it comes to these cons. This study suggests more work needs to be done to understand how to mitigate such outcomes, and which (if any) psychological interventions can help develop such skills.

Read the paper in full:

Horta, M., Shoenfelt, A., Lighthall, N. R., Perez, E., Frazier, I., Heemskerk, A., Lin, T., Wilson, R., & Ebner, N. C. (2024). Age-group differences in trust-related decision-making and learning. Scientific Reports, 14(1). https://doi.org/10.1038/s41598-023-50500-x