Sleep training — behavioral interventions to help infants and toddlers fall asleep independently — is one of the most emotionally charged and most extensively researched topics in pediatric science. The evidence consistently shows these methods are safe and effective, with no evidence of harm.
What Does the Research Say?
A 2016 RCT published in Pediatrics (n=43 infants, 6–16 months) compared graduated extinction (controlled crying), bedtime fading, and no-treatment control. Both sleep training groups fell asleep significantly faster, woke less frequently at night, and showed comparable cortisol levels to the control group — contradicting claims that sleep training elevates stress hormones. Parent stress was significantly lower in the sleep-trained groups. All differences in infant stress hormones had resolved by the end of the intervention.
A 2012 RCT followed sleep-trained infants to age 5 and found no differences in attachment security, emotional development, cognitive outcomes, or parent-child relationship quality compared to controls. A 2019 review of 13 studies covering 1,400 infants across extinction, graduated extinction, and fading methods all demonstrated improved sleep outcomes with no evidence of developmental harm. The American Academy of Pediatrics endorses both graduated extinction and unmodified extinction (Ferber and Weissbluth methods) as safe and evidence-based.
Key Findings
- 2016 Pediatrics RCT: sleep-trained infants showed identical cortisol levels to controls — no evidence of elevated stress
- 5-year follow-up confirmed no differences in attachment, emotional development, or cognitive outcomes
- AAP endorses both graduated extinction and full extinction as safe, evidence-based methods
- Parental mental health significantly improves with sleep training — exhaustion is a documented harm vector
Practical Takeaway
Choose the method aligned with your tolerance for crying and your infant’s temperament. Graduated extinction (check-ins at increasing intervals) and bedtime fading (shifting desired bedtime later then progressively moving it earlier) are both well-validated and gentler than full extinction. Consistency within a method matters more than which method you choose. Begin no earlier than 4 months, ideally 5–6 months.