At 4am on the seventh terrible night, every parent has wondered whether sleep training is the answer or the problem. The internet will give you both opinions, loudly. Here's the actual landscape: there are four well-described methods, the strongest randomised evidence (the 2012 Hiscock-Price five-year follow-up in Pediatrics) shows none of them harm children emotionally or developmentally, and the variable that actually predicts success is how consistent the parents are. The "best" method is the one you can stay consistent with for two weeks straight.
If you want to track whether what you're doing is working, Healthbooq has a sleep log that lets you see the week-over-week trend instead of getting lost in individual rough nights.
What Sleep Training Actually Is
Sleep training is teaching a baby to fall asleep at the start of the night without a specific external prop — being fed, rocked, walked, or patted to sleep — so that when they surface between sleep cycles at midnight, they can drop back into sleep without recreating that prop.
It is not: cutting overnight feeds in a young baby who still needs them; ignoring distress; or training a baby to sleep through hunger or illness. Most methods are designed for babies from around 4 to 6 months. Before 4 months, sleep is governed by hunger, neurological immaturity, and the not-yet-formed circadian rhythm — not by learned associations — and structured sleep training won't work and isn't recommended by the AAP, NHS, or most paediatric sleep researchers.
Graduated Extinction (Ferber Method)
You put your baby down drowsy but awake, leave the room, and return at lengthening intervals — say 3 minutes, then 5, then 10, then 12, then 15 — to briefly reassure them with voice and a hand on the back. You don't pick them up. You don't feed them (unless it's a planned feed). You leave again. Intervals get longer over successive nights.
This is the method with the strongest evidence base. Hiscock and colleagues' 2012 RCT in Pediatrics (n=326) found graduated extinction improved infant sleep at 3 and 6 months post-intervention, reduced maternal depression, and showed no measurable difference in cortisol regulation, child emotional/behavioural functioning, or parent-child attachment at five-year follow-up. Most families see clear improvement by night 4–7 if they hold the method without breaking pattern.
Full Extinction ("Cry It Out")
You put your baby down awake and don't return until morning, or until a pre-set wake-up time. Faster than graduated — most babies are sleeping through within 2–4 nights — but with more concentrated crying in nights one and two. The same evidence base applies: no detected long-term harm; equivalent improvements in maternal mental health.
Full extinction works best when: the baby is well (no cold, no teething pain, no fever); the room is consistent (16–20°C, blackout, white noise on); and both parents are completely on the same page. Inconsistency drags it out and makes it harder than graduated extinction.
Chair Method (Sleep Lady Shuffle)
You start by sitting in a chair right next to the cot while your baby falls asleep — present, but not actively settling them. Every 2–3 nights you move the chair further away: by the cot → midway across the room → at the door → outside. By the end of 2–3 weeks you're not in the room at all.
Slower than the extinction methods. Less crying, more parental presence. Useful for families who genuinely cannot do extinction, or for slightly older toddlers who can be unsettled by the sudden absence of a parent. Requires real consistency — drifting the chair back closer because of a hard night usually resets the clock.
Pick Up / Put Down and Gentle Methods
You pick the baby up when they cry, hold until calm, put them back down. Repeat as many times as it takes. Variants include patting and shushing in the cot, or "fading" the existing sleep prop (shorter and shorter feeds, less and less rocking) over weeks.
The evidence base here is thinner — these methods are described in clinical practice but haven't been studied in the same large randomised trials. Many families succeed with them. They tend to take longer (3–6 weeks) and don't always work for older babies (10+ months) with well-rehearsed sleep associations. They're a real option for parents who genuinely cannot tolerate any extended crying.
What Almost Always Helps Before You Start
Before picking a method, set up the conditions that make any method work better.
A consistent 20–30 minute pre-sleep routine. Bath → feed → book → song → cot, in the same order, every night. Babies pattern-match faster than parents expect. Two weeks of a stable routine, on its own, improves settling for many babies before any "training" begins.
Bedtime that matches actual tiredness. For most 4–9 month olds, that's 6:30–7:30pm. Putting a baby down 45 minutes after they were ready takes you into cortisol territory — wired, not sleepy, harder to settle by any method.
A consistent sleep environment. Dark room (blackout if needed), 16–20°C, white noise if you use it, no rocking-while-feeding-while-walking variations from night to night.
A wake window that fits their age. A 6-month-old has roughly 2.5–3 hour wake windows; an 8-month-old, 3–3.5; a 12-month-old, 3.5–4. Putting a baby to bed undertired is a common, fixable cause of long settling and false-start wakings.
How to Pick
Quick filter:
- Both parents committed and want results in under a week → full extinction
- Want results in a week or so but need to provide some reassurance → graduated extinction
- Older baby, separation-sensitive, or you can't do extinction → chair method
- No crying tolerance and willing to spend 3–6 weeks → pick up / put down or fading
Whichever method you pick, give it 14 days before deciding it's "not working." Most progress shows up in week two. Switching methods mid-stream is the single most reliable way to extend the process and prolong the crying.
Key Takeaways
Every reasonably-researched sleep training method works for most babies — graduated extinction, full extinction, chair method, pick-up-put-down. The randomised trials don't show emotional, behavioural, or attachment harm at follow-up. The biggest single predictor of success is not which method you pick but whether you stick with it for the full 1–2 weeks. Most babies turn the corner by night 4–7. Don't try sleep training before four months.
