Sleep Training Methods for Older Infants: What the Evidence Shows

The argument about sleep training is, by 2026, one of the more outdated arguments in parenting culture. The research base has grown to dozens of randomised trials and several long-term follow-ups, and they all point in the same direction: structured behavioural sleep interventions help babies sleep, help parents sleep, and don't produce the harms that the older debate worried about. Cortisol studies, attachment studies, and child outcome studies five years post-intervention have all come back reassuring.

That doesn't mean every family should sleep-train, or that any particular method is right for every baby. It does mean parents who want to try are not making a developmentally risky choice — and parents who don't want to try are not condemning their family to a year of broken sleep with no alternatives.

The Healthbooq app is useful for tracking sleep before and during a sleep-training plan — knowing the baseline (how many wakes, when they're feeding, how long settling takes) makes it much easier to see what's working.

What Sleep Training Actually Is

Sleep training, in the clinical and research literature, refers to behavioural interventions that help an older baby fall asleep without active parental intervention — and resettle themselves between sleep cycles without needing to be fed, rocked, or held back to sleep.

The goal is not to make a baby sleep through the night by force, or to eliminate normal night waking. Babies wake repeatedly through the night by neurological design — sleep cycles in older infants run about 50–60 minutes, and a brief surfacing between cycles is normal. The goal is to make those surfacings brief and self-resolving, the way they are in adults, rather than requiring a parent to recreate the conditions the baby fell asleep in.

The cause of most "sleep problems" in healthy older babies is not a sleep disorder. It's a sleep-onset association: the baby fell asleep being fed or rocked at bedtime, surfaces between cycles in the night, and now needs the same conditions to get back to sleep. The brain is doing exactly what it has been trained to do.

Sleep training is the relabelling of those conditions: instead of bedtime = "in mum's arms with a bottle", bedtime = "in my cot, awake, on my own, falling asleep here." Once that's the association, the baby can resettle themselves at the inter-cycle surfacings.

What sleep training is not:

• It is not sleep deprivation. The baby gets the same total sleep, often more.
• It is not ignoring a hungry, ill, or genuinely distressed baby.
• It is not a single method called "cry it out." There is a spectrum.
• It is not appropriate for babies under about 4 months — sleep architecture isn't mature enough, night feeds are still developmentally needed, and a young baby cannot yet self-soothe in the way the methods rely on.

When to Consider It

The standard window is 4–12 months, though many families try earlier or later. The pieces that should be in place before starting:

• Age 4 months minimum, ideally 6 months or older. UK and US paediatric sleep groups (Mindell, Hiscock) typically recommend 6 months as a comfortable start point.
• Health is settled. Not in the middle of a cold, ear infection, teething pain, vomiting bug, or recent illness.
• No major recent change. Ideally not the week of starting nursery, a house move, a new sibling, or a parent returning to work.
• Adequate daytime feeding. A baby who genuinely needs night feeds (often the case in 4–6-month range, sometimes longer in smaller babies) shouldn't be expected to drop them as part of sleep training. Talk to your health visitor or GP if uncertain.
• A consistent bedtime routine. 20–30 minutes of predictable wind-down: bath, feed, story, song, cot. Sleep training works inside a routine, not instead of one.
• Both parents on board. A method only one parent believes in is unlikely to be implemented consistently for the 5–10 nights it usually takes to work.

The Main Methods

There are three families of approach, with many named variants. The differences matter less than they sound — what differs is the amount of crying tolerated and the speed of effect.

1. Unmodified Extinction (sometimes called "cry it out" or CIO)

The baby is put down awake at bedtime and the parent leaves the room. No checks are done unless something is genuinely wrong. The same principle applies at night wakings: no parental intervention to settle.

Typically resolves within 3–7 nights. Crying is most intense on nights 1–2 and reduces sharply.

This is the fastest method. It is also the method most parents find hardest to tolerate emotionally, even when they understand it is safe. Many sleep researchers and practitioners recommend it as effective if a family is comfortable; few use it as their first recommendation.

2. Graduated Extinction (Ferber method, "controlled crying", "checks")

The most-studied and most widely-used method. Richard Ferber's 1985 book Solve Your Child's Sleep Problems gave the protocol that bears his name.

The parent puts the baby down awake, leaves, and returns at progressively longer intervals — typically 3 minutes, then 5, then 10, increasing on subsequent nights. Each check is brief and minimal: a brief touch, a calm reassuring phrase ("It's bedtime, time to sleep"), then leaving. The baby is not picked up; rocking and feeding are not used.

Typically resolves in 5–10 nights. Crying decreases rapidly, usually halving by night 3.

The variant that's slightly easier on parents and equally effective is "checks at fixed intervals" — every 5 or 10 minutes, regardless of night, rather than escalating intervals. Studies show no meaningful difference in outcome.

3. Gradual Fading / Chair Method / Sleep Lady Shuffle

The parent sits next to the cot at bedtime, providing presence but not active settling, and gradually reduces involvement over many nights. Variants:

• Chair method (Kim West's "Sleep Lady Shuffle"): a chair beside the cot for 3 nights, then halfway across the room for 3 nights, then by the door, then in the corridor, then gone.
• Pick up / put down (Tracy Hogg's "Baby Whisperer" approach): pick up if crying, put down when calm; repeat. Slow but gentle. Some families find it works; others find it stretches both nights and tempers.
• Verbal-only fading: stay near the cot, use only voice / brief touch, reduce the duration over nights.

These methods take longer (often 2–4 weeks) and involve less crying at any one moment. They are emotionally easier and arguably better-suited to children with separation anxiety, sensitive temperaments, or families that find extinction methods unbearable.

Camouflage Approaches

Some methods (the "no-cry sleep solution" by Elizabeth Pantley is the most popular) attempt to change associations gradually without any crying. They tend to work for families with relatively mild sleep issues and lots of patience; they often don't work for entrenched, multi-wake-per-night patterns. Honest framing: very gradual approaches sometimes don't move the needle at all, and families end up exhausted and back at the same problem six months later.

What the Evidence Actually Shows

The headline findings, from the literature parents are most likely to hear about:

• Hiscock and Wake (BMJ 2002, Australia). Randomised trial of 156 mothers/babies, controlled crying vs. usual care. Significant reduction in infant sleep problems; significant reduction in maternal depression. Effects sustained at 4 months.
• Hiscock et al. (BMJ 2007). Larger 692-baby cluster RCT, infant sleep group vs. control. Sleep problems halved in intervention group; maternal depression scores improved. No adverse effects on infant or family functioning.
• Price et al. (Pediatrics 2012). Five-year follow-up of the Hiscock 2007 cohort. No differences in child mental health, child-parent attachment, child stress regulation, or family functioning between sleep-trained and control children at age 6.
• Gradisar et al. (Pediatrics 2016). Australian RCT of 43 babies aged 6–16 months, randomised to graduated extinction, bedtime fading, or sleep education control. Both intervention groups showed faster sleep onset and fewer night wakings. No differences in infant cortisol, attachment security, or behaviour at 12 months between any of the three groups.
• Bilgin and Wolke (Child Development 2020). Longitudinal analysis of 178 mother-infant dyads. Found no association between extinction-based sleep training in the first 18 months and attachment security or behaviour at 18 months.
• Mindell et al. (Sleep 2006). Comprehensive review of behavioural sleep interventions across 52 studies. Strong evidence base for graduated extinction; sustained improvements; positive effect on parent mental health.

The cortisol question — does sleep training elevate stress hormones harmfully? — has been studied multiple times. Brief elevations during the training nights themselves are seen in some studies but not others; longer-term cortisol patterns post-training are normal. The original 2012 Middlemiss study often cited as showing harm has well-documented methodological problems and has not been replicated.

The bottom line: sleep training, done in older babies with reasonable care, is not a high-stakes developmental decision. It is a reasonable behavioural intervention with reasonable evidence behind it. Whether to do it is a values and family-fit question, not a developmental safety question.

Before You Start, Rule Out the Imitators

A baby who isn't sleeping well isn't always having a sleep problem. Things that look like sleep problems but won't respond to sleep training:

• Hunger. A young or smaller baby genuinely needing 1–2 night feeds at 5 months. The fix is feeding, not training.
• Reflux. Babies who arch and cry when laid flat may have GORD (gastroesophageal reflux). Worth a paediatric GP review before sleep training.
• Cow's milk protein allergy. Persistent unsettled sleep with eczema, blood in stools, or ongoing colicky symptoms warrants a feeding review.
• Iron deficiency. Older babies (>6 months) on a diet low in iron-rich foods can have iron-related restless sleep.
• Ear infection or teething. Recent-onset waking that wasn't there before, especially with daytime irritability or fever, deserves a check.
• Obstructive sleep apnoea / large adenoids. Snoring, mouth breathing, restless tossing, night sweats — worth a paediatric ENT or GP review.
• Developmental leap or sleep regression. The 4-month, 8-month, and 12-month "regressions" are real (changes in sleep architecture) and usually self-resolve in 1–2 weeks. Sleep training during one is harder.

A short health visitor or GP appointment to flag these before you start is well worth the time.

Practical Implementation

Whichever method you choose, the consistency rules are the same:

1. Pick the method you can actually do for 7–14 nights without flipping. Inconsistent training is worse than no training — it teaches the baby that crying for long enough produces the desired response.
2. Both parents on the same plan. Decide who will respond to wakings and how. If one parent goes in and feeds while the other follows the protocol, the baby learns to wait for the feed-parent.
3. Have a clear bedtime routine. Same order every night. The cues are part of what builds the new association.
4. Put the baby down awake. Not drowsy-but-sleeping; awake. The skill being built is falling asleep, not staying asleep after falling asleep on you.
5. Treat night wakings the same as bedtime. A baby who self-settles at bedtime but is fed/rocked back to sleep at 2 a.m. won't make progress.
6. Track it. A simple log of wake times and settling times for the first two weeks tells you whether you're trending in the right direction. Often progress feels less than it actually is in the moment.
7. Plan for it to get worse before it gets better. Most methods produce an "extinction burst" on night 2 or 3 — the baby protests harder before giving up. This is normal and not a sign the method isn't working.
8. Allow 2 weeks before judging. A method abandoned at night 4 because "it isn't working" usually was working.
9. Don't sleep-train through illness or major disruption. Pause and resume.

Methods That Don't Suit Some Families

There are families for whom none of the standard methods are a good fit:

• Co-sleeping families by design or choice. Sleep training as described doesn't fit a bedsharing setup. Different strategies (shared room, separate sleep surface, gradual association change) apply.
• Highly sensitive babies for whom even brief crying produces escalating distress that doesn't extinguish. Some of these babies do better with very gradual fading; some with no formal training at all.
• Babies with significant developmental, medical, or sensory differences. Sleep training protocols need to be adapted; talking to a paediatric sleep specialist is worthwhile.
• Parents with significant trauma history for whom hearing a baby cry triggers their own unresolved experience. Pushing through this is rarely the right call; gentler approaches or working with a perinatal mental health practitioner first is usually better.

The "right" approach is the one you can actually do, that fits your child, and that lines up with your values. The wrong approach is a half-implemented version of a method you don't believe in.

When to Seek Professional Help

If you've tried structured sleep training consistently for 2–3 weeks with no improvement, or if waking is associated with snoring, gasping, sweats, or other features that suggest a medical cause, seek input from your GP or health visitor. Persistent severe sleep problems in children sometimes need specialist paediatric sleep services — many CCGs / Trusts have sleep clinics through paediatrics or community CAMHS.

Private paediatric sleep consultants can also help, particularly with complex cases. Look for someone with paediatric clinical training (rather than only certification from a sleep coaching course); the evidence base for paediatric sleep is enough that a qualified clinician's input is more useful than generic coaching.

Sleep training in older babies (typically 4–12 months) has been studied in dozens of randomised trials, including multi-year follow-up. The headline: graduated extinction (Ferber), unmodified extinction ('cry it out'), and gradual fading approaches all work — they reduce night waking, shorten settling time, and improve parent mental health. Repeated, large-scale follow-up (Price et al. Pediatrics 2012, five years out; Gradisar et al. 2016 with cortisol monitoring) has not shown harm to infant attachment, cortisol patterns, or development. The right method is the one you can actually carry out consistently — and consistency is the single biggest predictor of success. Before starting, rule out hunger, reflux, illness, ear infection, or developmental disruption, all of which mimic 'sleep problems' and don't respond to behavioural training.