Sleep Training Methods: What the Research Actually Compares (Not What Instagram Tells You)

Sleep training is probably the most contentious topic in the infant parenting space, and I say this as someone who also works in a field where people argue about attachment theory for a living.

The arguments are fierce, they are emotional, and they frequently invoke child psychology — often inaccurately. People who oppose sleep training cite attachment research. People who support sleep training cite cognitive development research. Both groups are often citing things selectively. As a psychologist, this bothers me professionally. As a parent who has been through it twice, it also bothers me because I remember what it felt like to be in that state of exhaustion while everyone online had a very confident opinion.

Let me try to give you an honest account.

The main approaches

Extinction methods (Cry It Out / CIO, Weissbluth): The parent establishes a bedtime routine and then places the baby in the cot awake and does not return until morning, regardless of crying. Marc Weissbluth's approach in Healthy Sleep Habits, Happy Child advocates this for infants from about four to six months.

Graduated extinction (Ferber method): The parent places the baby down awake and returns at progressively longer intervals to briefly reassure but not pick up. Richard Ferber's method involves a structured schedule of waiting intervals that increase over nights.

Fading methods: The parent gradually reduces their involvement over time — moving a chair progressively further from the cot, reducing the amount of physical contact during settling, or gradually extending wait times. The Camp method and various "chair" methods fall here.

No-cry approaches: Elizabeth Pantley's The No-Cry Sleep Solution involves responding to every cry but gradually and gently adjusting sleep associations. Typically slower results, lower distress during the process.

What the research actually shows

There are several good randomised controlled trials here, which is more than many parenting topics can claim.

A 2016 study in Pediatrics by Michal Gradisar and colleagues randomised infants aged six to sixteen months to graduated extinction, bedtime fading, or control groups. Both intervention groups showed significant improvements in infant sleep compared to controls. Critically: there were no significant differences between groups on infant cortisol, attachment security (measured at one month post-intervention), or parent-child relationship quality at twelve months. The graduated extinction group improved sleep faster. There was no evidence of psychological harm.

A comprehensive 2020 review in the same journal (Middlemiss et al.) examined cortisol research in sleep training studies. The finding most cited by sleep-training opponents — that infant cortisol remains elevated even when crying stops — was reported in one small study (n=25) with significant methodological limitations and has not been reliably replicated.

The evidence base supports this summary: standard behavioural sleep training methods (graduated extinction and fading approaches) produce faster sleep improvements than no-cry approaches, with no evidence of harm to attachment, cortisol regulation, or developmental outcomes at the timepoints studied. No-cry approaches produce slower improvement with less distress during the process.

What the research does not show is long-term follow-up beyond twelve to twenty-four months. It does not measure the experiences of children who did not respond to these methods. And it cannot tell you how the trade-offs feel in your specific family.

The things the research doesn't measure

Research measures sleep, cortisol, attachment Q-scores, and parental wellbeing on standardised scales. It does not measure:

The specific quality of a parent's distress while listening to their child cry — which varies enormously by individual, and which is a real cost regardless of what happens at the twelve-month developmental outcome assessment.

Whether a family's specific circumstances (one bedroom flat with grandparents in the next room; two children sharing a room; a parent with postnatal anxiety) affect the feasibility or experience of different approaches.

The cumulative effect of sustained sleep deprivation on the very parental sensitivity that attachment theorists correctly identify as important. Severely sleep-deprived parents are less regulated, less responsive, and less able to provide the warm, consistent care that matters most.

Where I land

I did a modified graduated extinction approach with my daughter at around six months. It worked in four nights. It was hard. I monitored her more carefully than the protocol recommended because I am a psychologist who could not turn that off.

With my son, at fifteen months, we are in the middle of something. I will not call it "sleep training" because we are mostly just surviving, which I believe is an equally valid approach.

I am not going to tell you which method to use. The research does not give me the standing to make that decision for your family, your baby, or your nervous system. What I can tell you is that if you choose a standard behavioural approach, the evidence does not support the most catastrophic claims made against it. And if you choose not to use one, you are also fine, and the evidence doesn't support the apocalyptic predictions about children who were never sleep trained.

This is genuinely a decision where your family's specific situation, values, and tolerances matter more than the research.

Which I find, as a researcher, both humbling and honestly a relief.