Sleep Training Methods
Every method compared - How each works, cry level, speed, and what the evidence actually says about long-term outcomes.
When Can I Start?
Most paediatricians and sleep consultants recommend 4–6 months corrected age as the earliest point for formal sleep training. Before 4 months, babies are not developmentally capable of self-settling and physiologically need night feeds. See the sleep by age guide for what is developmentally appropriate at each stage. The 4-month regression is a permanent brain restructuring - many parents use this as the prompt to begin building self-settling skills for the first time.
Responsive settling only. Swaddle, shush-pat, white noise, contact naps. Build the routine foundation. Focus on safe sleep practices. See newborn schedule for what this looks like day-to-day.
Drowsy-but-awake placement at the start of each nap. Consistent bedtime routine. Gentle fading if ready. Avoid starting during an active regression or illness.
Any method below can be used. Baby is physiologically able to sleep longer stretches without feeds. Check that night feeds are habitual rather than nutritionally needed before starting.
Before You Start: Prerequisites
Sleep training works best when the foundations are right. Skipping these steps is the most common reason training fails or takes much longer than expected.
Correct wake windows
If your baby is going to sleep with wake windows that are too short (undertired) or too long (overtired), training will not work. Confirm the right windows on the sleep by age page and verify with the Nap Tracker.
A consistent bedtime routine
The routine is the cue. Bath, massage, feed, song, dark room - in the same order every night. Aim for 20–30 minutes. Without a consistent routine, the baby has no signal that sleep is about to happen.
Appropriate sleep environment
Dark room (blackout curtains matter enormously in Singapore where the sun rises before 7am - see the Singapore heat guide), white noise at 50–65dB, room temperature 24–26°C. These alone sometimes resolve frequent night waking.
Baby is not sick, teething badly, or mid-regression
Training during a sleep regression significantly increases crying and reduces success. Wait until the baby has had a relatively stable week before starting. Do not begin if the baby has a fever or illness.
Both caregivers aligned on the method
Inconsistency between caregivers is the single most common reason training fails. If one parent intervenes after the other has held back, the baby learns to persist longer. Agree on the exact method and commit to the same approach every night.
Extinction (Cry It Out / CIO)
3–7 daysBaby is placed in the crib awake after the bedtime routine. Parent does not return until morning or the next scheduled feed. No check-ins during the night. This is the fastest method, but it involves the highest initial crying. Studies consistently show it is safe (see evidence section below).
How it works:
Complete the routine (bath, feed, song, crib). Leave the room. Do not return. Baby learns to self-settle within 3–7 nights. Crying peaks on night 1–2 and drops sharply thereafter. The silence on night 4–5 can feel abrupt but is normal.
Ferber / Graduated Extinction
5–10 daysTimed check-ins with increasing intervals. Parent returns to reassure without picking up, then leaves again. Intervals grow each night. Baby hears your voice while learning to self-settle - which many parents find easier to tolerate than full extinction. This is the most commonly recommended method by paediatricians in Singapore. It is compatible with maintaining a scheduled night feed if the baby still needs one.
How it works:
Night 1: wait 3 min, check in (no picking up - verbal reassurance only), wait 5 min, check in, then 10 min intervals. Night 2: start at 5 min. Night 3+: increase intervals again. Keep check-ins brief and calm. Do not pick up during check-ins - this resets the process.
Chair Method (Fading / Sleep Lady Shuffle)
2–3 weeksParent sits in a chair next to the crib on night 1 and gradually moves the chair toward the door over 1–2 weeks until they are out of the room entirely. The physical presence of the caregiver reduces separation anxiety while teaching the baby that sleep happens without being held. This is the gentlest of the formal methods but requires strict consistency to work - it is very easy to slip back into rocking or feeding if the chair is right there.
How it works:
Night 1: chair beside crib. Minimal interaction - no eye contact, shushing only. Every 2–3 nights, move the chair one position closer to the door. Be consistent at every nap and night sleep. The same schedule applies to naps during the training period.
Pick Up Put Down (PUPD)
Varies widelyBaby is put down awake. When crying, parent picks up until calm, then puts down again. Repeated until baby falls asleep. Parent never leaves the room. This works best for younger babies (3–6 months) where the wake windows are shorter and the training window per nap is brief. For older babies, being picked up and put back down repeatedly can become overstimulating and actually extend the process significantly.
How it works:
Put baby down awake after the routine. If crying, pick up until calm (not asleep), then put down immediately. Repeat. This can take 30–60 minutes initially but shortens each night. Ensure wake windows are correct first - if the baby is not tired enough, PUPD will not work regardless of how consistently it is applied.
No Cry / Pantley Pull-Off
4–8 weeksGradual removal of sleep associations while nursing or rocking. As baby is nearly asleep, unlatch or stop rocking. Break the last-thing-remembered association over many nights. This is the approach for parents who cannot tolerate any crying and who have the time and consistency to apply it for weeks. It requires keeping a sleep log to track progress, as improvements are subtle and slow. It is compatible with breastfeeding and does not require reducing night feeds before starting.
How it works:
Begin unlatching or reducing rocking just before deep sleep - baby may rouse and re-associate. Repeat with decreasing dependency each night. Track with a sleep log (note start time, how long to settle, night waking frequency) so you can see the gradual improvement. Without a log, the slow progress feels like no progress.
What Does the Evidence Say?
The short answer: sleep training is safe. A 2016 Australian randomised controlled trial (Price et al.) followed 326 babies randomly assigned to graduated extinction, bedtime fading, or standard care. At 6-year follow-up, there were no differences in:
- ✓Child emotional and behavioural outcomes
- ✓Parent-child attachment security
- ✓Child stress hormone (cortisol) levels
- ✓Parental mental health
The American Academy of Pediatrics (AAP) supports behavioural sleep training from 6 months. The decision is personal - the best method is the one you can apply consistently and that fits your family values. Sleep regressions will temporarily undo progress, but the self-settling skill is not lost - it comes back faster than the first time.
Common Sleep Training Mistakes
Starting too early (before 4 months)
Newborns and young babies are not developmentally ready for self-settling. See the sleep by age guide for what is developmentally appropriate before 4 months.
Inconsistency between nights
If you intervene one night and do not the next, the crying actually increases - the baby has learned that persisting long enough works. Both caregivers must apply the same method every night, even during regression periods when the temptation to revert is highest.
Wrong wake windows
Putting a baby down with the wrong wake window (too early = undertired, too late = overtired) makes any method fail. Check the Nap Tracker to calculate correct windows. If naps are short or bedtime fights are long, the wake window is usually the problem.
Training when sick or teething badly
Baby needs extra comfort when unwell. Wait until recovered, then restart. Any sleep training done during illness typically needs to be redone anyway. Extra night waking during illness is expected and appropriate - respond freely.
Not fixing the sleep environment first
In Singapore, early morning light from a non-blackout window and road or corridor noise can wake a baby mid-cycle no matter how well they have been trained to self-settle. Resolve the environment first. See the Singapore-specific sleep environment guide for HDB-specific solutions.
Quitting after 2–3 nights
Most methods take 5–14 nights to show full results. Stopping early is the worst outcome - all the distress with none of the benefit. If the method is not working after 14 nights with full consistency, the setup (wake windows, environment, associations) needs reviewing, not the method itself.
More from the Sleep Guide: