Imagine you're A/B testing two landing pages, but your users have never seen a website before — they've never seen anything before. They're hours old. If Group A still clicks one page more than Group B, you can't blame the algorithm, the ad copy, or the remarketing funnel. The difference has to come from the hardware. That's the logic powering this Cambridge study, and it's cleaner than most research designs in the sex-differences literature precisely because there's almost zero cultural exposure to confound the signal. The committed claim: newborn girls, tested between eight hours and three days after birth, show proportionally greater visual attention to a human face versus a mechanical toy (Newton's cradle) compared to newborn boys. Both sexes preferred the face — the split isn't about whether babies like faces, but about the magnitude of that preference. The girls' face-preference was measurably stronger. Published in Biology of Sex Differences, the study tested 130 neonates at the Rosie Hospital in Cambridge using eye-tracking while the babies sat on a parent's lap watching side-by-side videos. The ladder here matters. This is not the first study to report neonatal sex differences in social attention. Simon Baron-Cohen's group (where this lab is housed) published similar findings in the early 2000s, but those earlier studies drew criticism for small samples and methodological questions about how gaze was coded. This study uses eye-tracking hardware rather than human coders, which closes one major objection. It sits in a lineage that includes prior work on prenatal testosterone exposure and sex differences in brain structure at birth — a 2025 Nature paper found structural brain differences in neonates, attributed to the 2x-plus testosterone differential male foetuses experience in the first two trimesters. The integrity profile is mixed in instructive ways. The validation is observational — no intervention, no pre-registration mentioned, no independent replication yet. The sample of 130 is decent for neonatal research (recruiting newborns is genuinely hard) but modest by general standards. The eye-tracker is a real improvement over hand-coded gaze, but the choice of stimuli — one face versus one Newton's cradle — introduces a confound: is the difference about faces specifically, or about any social versus mechanical stimulus? A single stimulus pair can't answer that. The researchers are appropriately cautious, emphasizing average group differences with substantial overlap between sexes, but the headline-ready framing ("girls prefer faces") risks outrunning the data. The milestone question is where the study gets genuinely interesting. The Cambridge team signals they want to follow these babies longitudinally — tracking whether neonatal face-preference predicts later language development, social skills, or vulnerability to autism and ADHD. That's the real payoff. A cross-sectional snapshot showing a small group-level difference at birth is suggestive; a longitudinal dataset linking neonatal attention patterns to developmental outcomes would be a different order of evidence. The gap between those two is years of follow-up and attrition management. The obvious experiment not run: testing whether the effect scales with prenatal testosterone levels in the same babies. The team mentions testosterone as a leading suspect and cites the literature on 2x-plus testosterone exposure in male foetuses, but doesn't report hormone measurements for their sample. This is almost certainly a resource and ethics constraint — measuring prenatal testosterone requires amniocentesis or cord-blood samples, and the study design appears to have been built around eye-tracking simplicity. A future study linking individual hormone levels to individual gaze patterns would be the decisive test of the proposed mechanism. What you're left with is a well-executed observational study that tightens one screw in a much larger argument. It doesn't settle the nature-nurture debate — nothing will, because the debate is about proportions, not categories. But it does establish, with better methodology than predecessors, that at least some sex-typed attention differences exist before the cultural machinery has time to spin up. The honest read: small effect, clean design, suggestive but not conclusive, and the real value depends entirely on whether the longitudinal follow-up happens.
