Sleep training is a topic where the parenting internet is at war with itself. One side says it scars babies for life. The other side says one bad night fixes everything. Neither is true. The peer-reviewed evidence is, frankly, more boring and reassuring than either camp wants. Sleep training works modestly. It does not harm attachment. It is not a magic switch. The method matters less than your ability to do it consistently for a week.
This piece walks through the main methods, what the research actually shows, when sleep training is and is not appropriate, and how to pick something you can stick to.
Healthbooq lets you log sleep before and after a change so you can see whether it is actually working — useful when 3am feels like nothing is.
What the Evidence Shows (and Does Not)
The big studies:
- Hiscock et al. (2007, BMJ) — randomised trial of behavioural sleep interventions vs usual care in 328 infants. Improved infant sleep, reduced maternal depression. No adverse effects on the baby.
- Price et al. (2012, Pediatrics) — followed the Hiscock cohort to age 6. No differences in stress, behaviour, sleep problems, parent-child relationship, or cortisol patterns between sleep-trained and control children.
- Gradisar et al. (2016, Pediatrics) — RCT of graduated extinction, bedtime fading, and control. Both methods improved sleep. Salivary cortisol patterns were not elevated. Attachment security was unchanged.
- Mindell et al. systematic reviews — consistently find behavioural interventions (extinction, graduated extinction, fading) improve sleep onset latency, night wakings, and parent-reported sleep quality.
What the evidence does not show: sleep training producing better-sleeping children years later. By age 5 or 6, the differences mostly wash out. It helps now. It is not a one-time fix that imprints sleep skills forever.
What it also does not show: long-term harm to attachment, stress regulation, or behaviour. The argument that sleep training is harmful relies largely on theoretical extrapolation (Narvaez and others) rather than data on actual sleep training. That is not the same as saying every baby tolerates it equally — some clearly find it harder than others — but the population-level evidence is reassuring.
So the honest framing is: this is a tool. It works. It does no measurable harm in the studies we have. It is not the only way. Use it if you need it.
When to Start (and When Not To)
Most experts say 4–6 months minimum. Why:- Before about 16 weeks, babies do not have the circadian and homeostatic sleep regulation in place to consolidate night sleep.
- Most babies still need at least one night feed for nutrition until 4–6 months. Night feeds are not "training problems" — they are calories.
- Self-soothing capacity (the ability to fall back asleep without intervention) starts to come online from around 4 months.
- Baby is under 4 months
- Baby is unwell, recently vaccinated, or in the middle of a developmental leap or illness
- You are about to travel, move house, or start childcare
- You cannot be consistent for at least a week (one parent on board, the other not — postpone)
- Night feeds are still needed for weight gain (check with health visitor or GP)
The Main Methods
Full Extinction ("Cry It Out")
What you do: bedtime routine, place baby in cot awake, leave the room, do not return until morning (except for safety or feeds you have planned).
Speed: typically settled in 3–5 nights. Night 2 is often worse than night 1.
Pros: fastest. Clearest signal to baby. No mixed messages.
Cons: hardest for parents emotionally. The first 2–3 nights involve sustained crying. If you cannot tolerate that, do not start — going in halfway through teaches the baby to cry longer next time.
Ferber, who is often associated with cry-it-out, did not actually advocate full extinction. He recommended graduated.
Graduated Extinction (The Ferber Method)
What you do: bedtime routine, baby in cot awake, leave. Return at set intervals — for example 3 minutes, then 5, then 10, then 10 thereafter — to briefly reassure (a few seconds, voice, gentle touch, no pick-up, no feed). Leave again. Repeat.
Speed: usually 5–7 nights.
Pros: the most studied method. Lets you check on baby. Less emotionally brutal than full extinction.
Cons: the check-ins can wind some babies up rather than soothe — they see you, you leave, they cry harder. If your baby is one of those, switch methods rather than persist.
Chair Fading (Sleep Lady Shuffle)
What you do: bedtime routine, place baby in cot, sit in a chair next to the cot. No picking up, but you can shush, pat, give verbal reassurance. Every 2–3 nights, move the chair further away — to the middle of the room, by the door, just outside, then gone.
Speed: typically 2 weeks.
Pros: parent stays present. Less crying for many babies. Suits parents who cannot do extinction.
Cons: slow. Requires real consistency. Easy to backslide ("just one cuddle").
Pick-Up-Put-Down
What you do: place baby down. If they cry, pick up briefly to calm, put down again as soon as calm. Repeat as many times as it takes. No feeding to sleep.
Speed: variable. Can take 2–4 weeks.
Pros: most "gentle". High parental presence.
Cons: physically exhausting. Can become a power struggle that wakes the baby up further. Often more effective from around 6–7 months than younger.
Bedtime Fading
What you do: identify the time the baby actually falls asleep (not the time you put them down). Shift bedtime to that time. Once they are settling within 15 minutes consistently, move bedtime 15 minutes earlier every few nights until you reach your target time.
Speed: 2–3 weeks.
Pros: works with the sleep drive rather than against it. Almost no crying. Effective in research (Gradisar et al.).
Cons: requires accurate observation. Baby often goes to bed late initially (8.30pm or 9pm) which feels wrong.
"No-Cry" Approaches (Pantley and similar)
What you do: a bundle of low-intensity changes — gentle routine adjustments, gradual feed-to-sleep weaning, cot-tolerance building during the day, paced bedtimes.
Speed: weeks to months.
Pros: very low distress. Suits families where any crying is a deal-breaker.
Cons: results are modest and slow. Some babies are not changed by it at all. Honest framing: this is more a way of life than a training protocol.
Choosing a Method
Three honest questions:
- Which one can I do for 7 nights without breaking? The method you can stick to consistently is the one that works. Inconsistent extinction is worse than consistent chair fading.
- What does my partner agree to? A divided household sabotages every method.
- What is the cost of doing nothing? If both parents are functional and the night wakings are tolerable, you do not have to train. Plenty of babies sleep through eventually with no training. If the lack of sleep is wrecking mental health, capacity to drive safely, or work — that is a real cost too.
What Sleep Training Will Not Do
- It will not change your baby's temperament. A spirited baby will still be spirited.
- It will not eliminate night wakings forever. Teething, illness, developmental leaps, travel — all reset things. Plan to do brief tune-ups.
- It will not work the same on every child. Some babies sleep-train in 3 nights; some take 3 weeks; some find it hard regardless of method.
- It will not produce a "good sleeper" by 8 months that you can take credit for. Sleep is mostly biology.
Safety, Always
Whichever method you use, the safe sleep rules do not relax:
- Back-sleeping
- Firm flat surface
- Nothing in the cot but a fitted sheet and a sleep bag
- Room temperature 16–20°C
- Same room as parent for the first 6 months
Sleep training is about how you respond to crying. It is not a reason to put a soft toy in the cot to comfort baby, or to add a wedge, or to switch to tummy sleep "to help them settle". The environment rules are non-negotiable.
Key Takeaways
Sleep training works modestly and does not damage attachment — multiple randomised trials and follow-ups (Hiscock 2007, Price 2012, Gradisar 2016) say so. The main methods are extinction (cry-it-out), graduated extinction (Ferber), chair fading, pick-up/put-down, and bedtime fading. Pick the one you can do consistently. Do not start before 4–6 months — younger babies still need night feeds and the developmental machinery is not there. Sleep training improves sleep but does not produce a different child.
