What does a baby know about right and wrong? A foundational finding in moral psychology suggested that even infants have a moral sense, preferring “helpers” over “hinderers” before uttering their first word. Now, nearly 20 years later, a study that tried to replicate these findings calls this result into question.
In the original study, Kiley Hamlin and her colleagues showed a puppet show to six- and ten-month-old babies. During the show, the babies would see a character — which was really just a shape with googly eyes — struggling to reach the top of a hill.
Next, a new character would either help the struggling individual reach the top (acting as a “helper”) or push the character back down to the bottom of the hill (acting as a “hinderer”).
By gauging babies’ behaviour — specifically, watching how their eyes moved during the show and whether they preferred to hold a specific character after the show ended — it seemed that the infants had basic moral preferences. Indeed, in the first study, 88% of the ten-month-olds – and 100% of the six-month-olds – chose to reach for the helper.
But psychology, and developmental psychology, in particular, is no stranger to replicability concerns (when it is difficult or impossible to reproduce the results of a scientific study). After all, the original study sampled only a few dozen infants.
This isn’t the fault of the researchers; it’s just really hard to collect data from babies. But what if it was possible to run the same study again — with say, hundreds or even thousands of babies? Would researchers find the same result?
This is the chief aim of ManyBabies, a consortium of developmental psychologists spread around the world. By combining resources across individual research labs, ManyBabies can robustly test findings in developmental science, like Hamlin’s original “helper-hinderer” effect. And as of last month, the results are in.
With a final sample of 567 babies, tested in 37 research labs across five continents, babies did not show evidence of an early-emerging moral sense. Across the ages tested, babies showed no preference for the helpful character.
John Locke, an English philosopher argued that the human mind is a “tabula rasa” or “blank slate”. Everything that we, as humans, know comes from our experiences in the world. So should people take the most recent ManyBabies result as evidence of this? My answer, however underwhelming, is “perhaps”.
This is not the first attempted replication of the helper-hinderer effect (nor is it the first “failure to replicate”). In fact, there have been a number of successful replications. It can be hard to know what underlies differences in results. For example, a previous “failure” seemed to come from the characters’ “googly eyes” not being oriented the right way.
The ManyBabies experiment also had an important change in how the “show” was presented to infants. Rather than a puppet show performed live to baby participants, researchers instead presented a video with digital versions of the characters. This approach has its strengths. For example, ensuring that the exact same presentation occurs across every trial, in every lab. But it could also shift how babies engage with the show and its characters.
I appreciated the recent remarks made by Michael Frank, founder of the ManyBabies consortium, on social network BlueSky: “Some people will jump to the interpretation that [the results of ManyBabies] shows that the original finding was incorrect (and hence that the other replications were incorrect as well, and the earlier non-replications were right). This [is] one possibility – but we shouldn’t be so quick to jump to conclusions.”
Rather, we can take this finding for exactly what it is: a well-executed large investigation (senior-authored by Kiley Hamlin herself) of the hypothesis that infants prefer helpers over hinderers. In this instance, the hypothesis was not supported.
This could be because, underneath it all, Locke was right. Perhaps the babies tested hadn’t had enough time in the world to learn “right from wrong”, so they wouldn’t make any distinction between a helpful character and a harmful one. Or perhaps there’s something more complicated going on. Only more science, with many, many more babies, will tell us.
At the very least, a question mark now hangs over one of the most famous experiments in developmental psychology.
Madeline G. Reinecke does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.
