Mothers should ignore unscientific claims that babies sleep better if left to cry

When an “expert” in one field weighs in heavily in an area she isn’t trained in, should we –and she – not be slightly cautious?

Take the American author, Emily Oster, who is also a professor of economics at Brown University.  In her huge bestseller Expecting Better, Oster looked at the scientific evidence around assorted bits of advice to pregnant women.

Her follow-up blockbuster success is Cribsheet in which she attempts to make sense of the evidence for popular strands of parenting advice and it’s already flying off the shelves.

At the weekend, one of our broadsheet Sunday newspapers picked up on one big reason why: Ms Oster says that leaving babies to cry until they fall asleep rather than soothing them with an adult presence does babies no harm at all.

For all those parents who are comfortable with the sound of a tiny baby screaming and screaming until it makes itself sick and falls asleep through exhaustion, this must be very reassuring. And she’s an economist! Economists are always right about everything, aren’t they?

After doing something called “crunching all the research data”, Oster claims that “studies show” that letting a baby cry itself to sleep “does not cause developmental damage”. You have to ask yourself, why didn’t anybody think to “crunch the data” before?

The answer is – plenty of people have tried to, but with very different results, because the truth is, there is very little data, from very few studies, looking at very few babies. I’m indebted to top sleep coach Lyndsey Hookway, author of Holistic Sleep Coaching (Praeclarus Press £14.43) for this eye-opening analysis. Full disclosure: she’s my sleep coaching tutor as well. And her take on the issue shows how difficult it is to make generalisations even when scientific studies are in the frame.  

Studies on the immediate effect of leaving babies alone to cry show fairly conclusively that the baby who is crying alone is awash with the stress hormone, cortisol. This is easy to check accurately, by use of saliva swabs taken at intervals through the night. Interestingly, those studies also show that in babies whose parents have “successfully” practised crying-it-out, the cortisol levels rise at intervals through the night, indicating that the babies are feeling distress, but not crying. Meanwhile babies crying, even though apparently inconsolably, in the arms of an adult or with an adult patting them, have lower cortisol levels.

Cortisol, by the way, when it remains high for too long, inhibits brain development. It also builds up when a child (or adult) is over-tired, accounting for that hyperactivity which parents dread in the child whose bedtime has been delayed even by a very short time.

Only three studies – yes, THREE, count ‘em, Emily – have been done of the long term effects of crying it out, or CIO, or “extinction method” as sleep consultants call it, and only up to age 6.  Two of these studies looked at the same sample several years apart.

The first, Hiscock et al 2008, used a sample of 328 Australian mother-and-baby pairs, and randomised to two groups, one which “received usual care” and the other completed a prescribed graduated extinction programme. It did find that fewer of the mothers in the CIO group had a significant depressive illness when their child was two years old; but it did not find that there was any statistically significant improvement in sleep outcome for either the control group or the CIO/extinction group.

There was a follow up study of the same sample, known as Price et al 2012. Nearly half – 46% – of the mums in the original sample refused to take part, leaving just 150 families, who were not representative of the general population (for example, non-English speakers were excluded and there wasn’t a representative ethnic balance). The study involved taking cortisol samples again; but the researchers did not have any recorded cortisol levels from before or immediately after the extinction method was applied, so there was no baseline to compare with.  

The third study of this kind was even smaller, involving just 43 babies aged 6-16 months, also in Australia, in 2016. Known as the Gradisar study, this also depended on taking cortisol samples from babies before programmes of night time care – a graduated extinction group and one where parents gave more supportive care (i.e. more responsive, with more cuddling).

The study claimed a “very large decline” in night time waking for the graduated extinction group but actually the difference was not very great – the number of times babies woke in the night went down over time in both groups, but more sharply in the graduated extinction group. Left to their own devices, all the children in the study might have been waking up with the same gradually diminishing frequency by the age of two, as all children do all over the world.

The Gradisar study also claims that “secure attachment” was not negatively impacted. Secure attachment is a child-psychology term referring to how confident toddlers are in their world, and it’s measured using a test devised by the psychologist Mary Ainsworth. You can find lots of examples of this test on YouTube if you search for, weirdly, “Strange Situation”.

In the Gradisar graduated extinction group, 46% of the children were judged to be “insecurely attached” (which we don’t want) and in the control group 39% of the children were.  Those are both quite high figures. In the population as a whole, about 30% of toddlers are thought to be insecurely attached. So the sample as a whole was not only tiny, but probably not even well-balanced or representative. You have to feel for the researchers, though: recruiting for these kinds of studies is the very devil at the best of times, and I’m struggling to think how they worded their outreach invitation, since they were effectively asking new mums to volunteer to try out parenting methods which might hurt or distress their babies.

So much for “crunching the numbers”. It would appear that Ms Oster is happy to extrapolate wildly from just three highly flawed studies which include fewer than 200 children, all from the same country but not representing its population make-up, and conclude that “crying it out” is harmless.  

It does make you wonder about economists.