does sleep training harm babies?

The evidence consistently says no. A 5-year follow-up study (Price et al., 2012) found no differences in emotional health, behavior, or parent-child attachment between sleep-trained and non-sleep-trained children. A 2016 RCT (Gradisar et al., Pediatrics) found cortisol levels and attachment security were identical between groups at 12-month follow-up. Sleep training does not cause lasting harm when used after 4-6 months in healthy infants.

what is the Ferber method and does it work?

The Ferber method (graduated extinction) involves putting baby down drowsy but awake, then waiting progressively longer intervals before briefly comforting without picking up. It works: the Gradisar 2016 RCT found significant reduction in sleep onset latency with no long-term harm. Most babies show marked improvement within 5-7 nights. It involves moderate crying, particularly on nights 1-2.

what is the gentlest sleep training method?

Fading and bedtime fading involve the least crying. Fading gradually reduces the sleep association baby relies on (nursing less each night, rocking less vigorously). Bedtime fading temporarily shifts bedtime to match when baby naturally falls asleep, then moves it earlier once they are falling asleep quickly. Both take 2-4 weeks and are as effective as cry-based methods at 12-month follow-up per Gradisar 2016.

when can I start sleep training my baby?

Most sleep specialists and pediatric guidelines recommend waiting until 4-6 months minimum. Before this, newborns lack circadian rhythm maturity, cortisol regulation is immature, and night waking is biologically appropriate to support growth. Starting before the window is ready typically means more crying, less success, and no developmental basis for it to work.

does cry-it-out cause attachment problems?

No — this is not supported by the research. Multiple studies with long-term follow-up find no difference in attachment security between sleep-trained and non-sleep-trained children. The most rigorous study (Price et al., 2012) followed children for 5 years and found them indistinguishable from peers on every measured emotional and behavioral outcome.

Sleep Training Methods Compared: What the Evidence Shows

It’s 3am and you’ve read twelve blog posts with twelve different opinions. One says Ferber will traumatize your baby. Another says anything less than full extinction is a waste of time. A third sells you a $200 course on the “gentle” method that’s actually just fading with a brand name.

Here’s what the actual research shows — not influencer opinions, not anecdata, not what worked for your neighbour’s cousin.

The Bottom Line First

A 2006 meta-analysis by Jodi Mindell reviewed 52 studies on behavioral sleep interventions. 94% of those studies showed clinically significant improvement in infant sleep. The key finding: no single method is clearly superior to any other. The best method is the one you can implement consistently.

That said, knowing the differences matters — for setting expectations, tolerating the process, and picking the right fit for your family.

Before You Start: Age Matters

Most sleep training methods are not recommended before 4–6 months. Here’s why:

Newborns lack circadian rhythm maturity — their internal clock hasn’t calibrated yet
Before ~12 weeks, night waking is biologically appropriate and feeds support growth
Cortisol regulation (the stress response) is immature in very young infants
Sleep consolidation naturally begins around 3–4 months as melatonin production develops

Starting before the window is ready typically means more struggle, more crying, and less success. Wait until your pediatrician gives the green light, usually around 4–6 months and when baby is gaining weight appropriately.

The Methods, One by One

1. Graduated Extinction (The “Ferber Method”)

What it is: Put baby down drowsy but awake. If they cry, wait a set interval before going in to briefly comfort (without picking up). Gradually increase the wait intervals over nights.

Typical schedule:

Night 1: wait 3 min, then 5 min, then 10 min
Night 2: wait 5 min, then 10 min, then 12 min
Night 3+: progressively longer intervals

What the research shows: The most-studied method. Gradisar et al. (2016, Pediatrics) ran a randomized controlled trial comparing graduated extinction, bedtime fading, and a control group. Both intervention groups reduced sleep onset latency significantly compared to control. Critically: at 12-month follow-up, there were no differences in cortisol levels, attachment security, or behavioral problems between the sleep-trained groups and the control group.

Honest reality: There will be crying — sometimes a lot, especially nights 1–2. Most babies show significant improvement within 3–7 nights. Parental anxiety is often the hardest part.

Best for: Parents who can tolerate some crying but want to feel like they’re “doing something” during the process.

2. Extinction (Cry-It-Out / CIO)

What it is: Put baby down awake. Don’t return until morning (or a predetermined feed time). No check-ins.

What the research shows: Extinction has the strongest and most consistent evidence base. It typically works faster than graduated methods — many babies show marked improvement within 3 nights. The same Gradisar 2016 data applies: no long-term harm found.

Price et al. (2012) followed sleep-trained children for 5 years and found no negative effects on emotional health, behavior, or the parent-child relationship compared to non-sleep-trained children. The children were indistinguishable from peers on every measured outcome.

Honest reality: The first 1–3 nights can involve significant crying. Parental follow-through is the main challenge. Check-ins (as in Ferber) can sometimes make crying worse by re-arousing the baby — pure extinction avoids this.

Best for: Parents who can commit fully and find check-ins make them (or their baby) more distressed.

3. The Chair Method (Sleep Lady Shuffle / Camping Out)

What it is: Sit in a chair next to the crib while baby falls asleep. Every 2–3 nights, move the chair progressively further from the crib — across the room, to the doorway, out of sight.

What the research shows: Less studied in isolation than extinction-based methods. It falls within the broader category of behavioral interventions covered by Mindell’s 2006 meta-analysis, which showed 94% improvement rates across methods.

Honest reality: Takes longer — typically 2–3 weeks. Baby still cries, but you’re present. The presence of a parent can sometimes be stimulating rather than soothing, prolonging the process. Requires significant parental time investment each night.

Best for: Parents who find leaving the room impossible but can handle a slower process.

4. Pick-Up-Put-Down (PUPD)

What it is: When baby cries, pick them up, calm them until drowsy (not asleep), then put them back down. Repeat as many times as needed.

What the research shows: Limited formal research specifically on PUPD. Associated with Tracy Hogg’s “Baby Whisperer” approach. Evidence is largely anecdotal and expert opinion. Some research suggests it can be more stimulating than calming for babies over 4 months — being picked up and put back repeatedly can increase arousal rather than decrease it.

Honest reality: Can work well for younger babies (4–5 months). Often exhausting for parents — can mean 30–40 pick-ups per night initially. Less effective for babies 6 months+.

Best for: Parents of younger babies who want to respond to every cry but are working toward independent sleep.

5. Fading (Graduated Withdrawal)

What it is: Gradually reduce the sleep association baby relies on. If you nurse to sleep, start pulling back a few minutes earlier each night. If you rock to sleep, rock less vigorously, then just hold, then sit next to crib.

What the research shows: Covered under the behavioral intervention umbrella in Mindell 2006. Gradisar 2016 found bedtime fading specifically to be as effective as graduated extinction with no difference in physiological stress markers.

Honest reality: Slower than extinction-based methods. Works best when you can identify a clear sleep association to fade. Requires patience and consistency over 2–4 weeks.

Best for: Parents who want to avoid significant crying and are comfortable with a longer timeline.

6. Bedtime Fading

What it is: Temporarily shift bedtime later (to match when baby actually falls asleep), then gradually move it earlier by 15-minute increments once they’re falling asleep quickly and independently.

The logic: If a baby’s natural sleep onset is 9:30pm but you’re putting them down at 7:30pm, they’ll fight sleep for 2 hours. That fight creates a learned struggle. By starting at 9:30 when they’re genuinely sleepy, they fall asleep quickly — building the association between lying down and falling asleep fast.

What the research shows: Gradisar et al. 2016 found bedtime fading equivalent to graduated extinction in outcomes, with similar cortisol and attachment findings at 12-month follow-up. Some parents find it more tolerable because crying is minimal when timing is right.

Honest reality: Requires tracking actual sleep times for a few days first. Works best when there’s a clear “sleep window” mismatch. Not ideal if your baby genuinely needs an earlier bedtime for developmental reasons.

Best for: Parents who are convinced their baby can’t fall asleep at the “recommended” time, and who want a low-cry approach.

The Honest Truth About All of These

All sleep training methods involve some crying. The question is: how much, for how long, and can you handle it?

This isn’t because all methods are harmful. It’s because:

Crying is how infants communicate frustration during change
Learning a new skill (self-settling) involves some discomfort
Your baby has strong preferences — and changing them takes time

The research consistently shows that parental consistency is the strongest predictor of success, regardless of method. Inconsistency — going in some nights but not others, caving after 45 minutes — often prolongs the process and increases overall distress for everyone.

Choosing Your Method

Method	Crying Level	Time to Results	Parental Involvement
Extinction (CIO)	High initially	3–5 nights	Low during night
Graduated Extinction	Moderate	5–7 nights	Moderate (check-ins)
Chair Method	Moderate	2–3 weeks	High (present nightly)
Pick-Up-Put-Down	Moderate	Variable	Very high
Fading	Low–moderate	2–4 weeks	High
Bedtime Fading	Low	1–2 weeks	Moderate

What the Research Does NOT Support

Sleep training causes attachment problems: Not supported. Multiple studies with long-term follow-up (including Price 2012’s 5-year study) find no difference in attachment security.
Sleep training is traumatic: Cortisol studies (including Gradisar 2016) show stress hormone levels normalize quickly and are no different from control groups at follow-up.
One method is clearly best: Not supported. Mindell 2006’s meta-analysis found consistent improvement across all behavioral methods.
Sleep training is appropriate for newborns: Correct — it’s not. Wait until 4–6 months minimum.

A Note on “Gentle” Methods

Marketing has made “gentle sleep training” a loaded phrase. In research terms, all behavioral sleep interventions are considered to have acceptable risk-benefit profiles when used after 4–6 months in healthy, typically developing infants.

“Gentle” doesn’t always mean less crying or faster results. It often means slower results with similar total crying across the process, just spread over more nights.

Choose based on what you can sustain. A method that takes 2 weeks but you can follow through on beats a method that promises 3 nights but you abandon on night 1.

Talk to Your Pediatrician

Before starting any sleep training, confirm:

Baby is gaining weight appropriately
No medical issues affecting sleep (reflux, ear infections, etc.)
Baby is developmentally ready (typically 4–6 months+)
Night feeds can be appropriately reduced for baby’s age and weight

Sleep training is a tool. Used at the right time, with consistency, the evidence says it works — and the evidence says it doesn’t hurt.

You’re not a bad parent for wanting everyone to sleep. You’re a tired parent doing the research. That counts.