Let’s say you are a man whose girlfriend has been sleeping with another man. Well, at least you think she has. You can’t be sure. And now she’s pregnant, and it seems a bit rude to demand a DNA test.

As a normal human, you’re reluctant to get on the blower to the Jeremy Kyle show, which leaves you with two options. Abandon the relationship and leave your former partner to raise a child that may or may not be yours on her own. Or you could stick it out in the hope that junior is yours after all.

Paola Bressan and Stefania Dal Pos of the University of Padua in Italy published some research on this question a couple of weeks ago. It deals with the conundrum faced by all men. While women know 100% that their child is their own – because they’ve been carrying it around for nine months – men can never be so confident. Not unless they keep their wives under lock and key. And that’s not been so easy since chastity belts dropped out of fashion.

Fathers judged parent–infant resemblance to be much stronger than mothers did

The researchers got together a group of men and women, some of whom were parents and some of whom were not. They showed each of these men and women an album of photographs. The photographs were of infants and adults. The task was to identify whether each child was the son or daughter of the adults they were pictured with. Some of the photographs were of genuine families, and some weren’t. Additionally, the participants were asked to judge how closely the children and adults resembled each another.

The results of the experiment showed that men and women, parents and non-parents alike, were no better than chance at identifying which baby belonged to which adult. They were terrible at judging family resemblance in children this young. But, revealingly, fathers judged parent–infant resemblance to be much stronger than mothers did, or men and women who hadn’t had kids. Fathers thought that the kids really did look like their parents, when nobody else was convinced.

How does this make sense? If fathers can’t be 100% certain of their paternity, surely they should be more hesitant about seeing resemblance where it doesn’t exist? If men make this kind of mistake outside the laboratory, in their real lives, surely they’re more likely to confuse another man’s child with their own, and end up investing in offspring to which they’re not genetically related. Hardly the sort of behaviour we’d expect natural selection to produce.

But perhaps, all else being equal, it makes more sense to assume any children your female partner bears are your own. Bressan and Dal Pos suggest that men might manage the risk that their children aren’t theirs by erring on the side of caution, assuming they are, and therefore avoiding the break up of their relationship and losing opportunities to produce more children in the future.