If a child is struggling to breathe, is floppy or will not wake, or their skin, lips or tongue look blue or grey: call 999 now. Colour changes are harder to see on brown and black skin, so check the palms, soles, lips and tongue. If you are worried but it is not an emergency, call NHS 111. In Northern Ireland use your GP out-of-hours service.

MarionSleep help for families

Responsive settling and gradual retreat

Staying with your child and withdrawing slowly, including the chair method.

The short version

  • Under six months, this is not the approach. For young babies, respond as soon as you can; feeding or holding to sleep teaches nothing unhelpful.
  • Gentler methods work as well in the long run: no appreciable difference in long-term effectiveness. A gentler method is mainly slower, not less effective.
  • Over six months, sit beside the bed holding a hand, then without contact, then further away, then the doorway, then outside, then alone.
  • Hold each position for several nights until settling is consistent. Going back a step is expected, not a setback.
  • If they wake in the night, return to the stage you reached at bedtime. If they get out of bed, be a boring parent.

The rest of this page explains where this comes from. You do not have to read it tonight.

About 5 minutes to read

2. Responsive settling and gradual retreat

Gradual retreat, also called gradual or graduated withdrawal, means staying with your child as they fall asleep and then, over a series of nights, being a little further away and a little less involved.

Under six months, this is not the approach. Cerebra says withdrawal is not recommended under six months; for young babies the NHS position is responsive settling. Newcastle Hospitals NHS Foundation Trust tells parents of babies aged 0 to 12 months: "When your baby wakes, respond to them as soon as you can so they don't get too upset," and says plainly that it is okay to let your baby fall asleep during a feed or while being held. Babies fed to sleep or held to sleep are not being taught anything unhelpful.

The six stages of gradual retreat A plan of a bedroom. The bed is on the left. A chair moves through five numbered positions: beside the bed with a hand held, beside the bed without contact, part way across the room, in the doorway, and finally outside the door. Stage six, marked over the bed, is the child falling asleep with nobody in the room. 123456 Asleep aloneBedDoorhand heldno contactacross the roomdoorwayoutside a few nights or more at each stage
NHS Lothian's six stages. Each position is held for several nights until settling is consistent. Going back a step is expected, not a setback.

For children over six months, NHS Lothian sets out six stages:

  1. Sit beside the bed, with contact Beside it rather than on it, a hand held rather than stroking, because stroking is stimulating.
  2. Sit beside the bed, without contact Same chair, same place, no longer touching.
  3. Move the chair away from the bed Still in the room, at a distance.
  4. Move the chair to the doorway In sight, at the edge of the room.
  5. Move the chair outside the room, by the door Close enough to answer, out of sight.
  6. Your child falls asleep alone The end point, reached gradually.

Two details matter more than the stages. If your child wakes in the night, go back to the stage you reached at bedtime, not an earlier one. And if they get out of bed, be a "boring parent": return them with the least possible engagement.

UK services pace this differently and none is the correct one: EPUT NHS uses nine stages, advancing every fourth night; Cerebra uses five stages of three nights; The Sleep Charity moves on after three consecutive nights without objection, using one set phrase ("It is time for sleep"), no eye contact, no conversation. NHS Borders suggests working out what contact your child actually needs and substituting a teddy, a blanket or an item of your clothing, and adds the line worth remembering on a bad night: do not worry if you have to go back a step or two.

The most reassuring finding on this page

Five studies have compared behavioural sleep treatments directly against each other. They found no appreciable difference in long-term effectiveness. A gentler method is mainly slower, not less effective. If you would rather take three weeks over gradual retreat than three nights of crying, the evidence does not say you are giving anything up.

Evidence. The 2006 review classes this as extinction with parental presence: four studies, 290 children, all positive, and noted as "more popular in England".

How long, and who it may not suit. About two weeks at Cerebra's pacing, over a month at EPUT's slower pacing. Not for babies under six months, and Cerebra advises against it where a child has a history of seizures. It asks for weeks of consistency, and ELFT NHS note that changing course part-way can unsettle things for a while. Read that as a reason to pick a method you think you can sustain, rather than a reason to carry on with one that has become unbearable. If you are parenting alone, being in the room every evening for weeks with nobody to take over is a real obstacle, and a fair reason to choose something else or to start when you have more support.

3. The chair method

The chair method is a popular name for gradual retreat rather than a separate technique. NHS Lothian calls the same thing "the disappearing chair". The name does not appear in Cerebra's guide or on The Sleep Charity's pages, and it has no evidence base of its own, so the evidence in the section above is its evidence.

  1. Do the routine, then into bed awake The routine ends at the bed.
  2. Sit in the chair beside the bed until they are asleep Present and calm.
  3. Keep interaction minimal One set phrase, repeated. No conversation.
  4. Hold each stage until settling is consistent Three nights in Cerebra's and The Sleep Charity's versions, four in EPUT's.
  5. Move the chair a short distance away A small move, not a dramatic one.
  6. Then the middle of the room, the doorway, then just outside Each move earned by consistent nights.
  7. Return them calmly if they get up Without comment, back to bed.
  8. Resettle night wakings at the bedtime stage Not at an earlier, closer one.
  9. Say goodnight and leave The final stage.

Be wary of any article giving a success percentage for "the chair method" by name. Those figures come from trials of graduated extinction or of extinction with parental presence and cannot be attributed to this label.

Where this comes from

  • NHS, Helping your baby to sleep, the bedtime routine sequence and the note on excitement before bed.
  • The Sleep Charity, Bedtime routines, the timed routine, "consistency is key" and the two-week rule.
  • NHS Lothian, Graduated withdrawal (PDF), the six stages, the pacing and the "boring parent" instruction.
  • East London NHS Foundation Trust, Sleep Basics (PDF), the camping out protocol including "start where you are", and the warning about abandoning a technique halfway.
  • Cerebra, Sleep: A Guide for Parents (PDF, version 4, October 2025), "choose whatever works for you", the controlled crying and bedtime fading protocols, the contraindications, the autism routine change, the sleep diary columns and "this is normal, keep going".
  • Newcastle Hospitals NHS Foundation Trust, sleep information for parents of babies aged 0 to 12 months: responsive settling, and falling asleep during a feed or while held.
  • Mindell et al. 2009, Sleep 32(5):599-606, the randomised trial of bedtime routines in 405 mothers.
  • Mindell and Williamson 2018, the dose-dependent pattern across 10,085 children.
  • Mindell et al. 2006, the American Academy of Sleep Medicine review: 52 studies, the 82 per cent average improvement, the classification of methods, and the comparisons finding no difference in long-term efficacy.
  • Hiscock and Wake 2002, BMJ, and Hiscock et al. 2007: the controlled comforting and camping out trials, and the loss of significance at four months.
  • Meta-analysis of behavioural sleep interventions, Scientific Reports 2022, 10 randomised trials, including the outcomes that were not significant.
  • Price et al. 2012, Pediatrics 130(4):643-651, the five-year follow-up of 326 children finding no marked long-lasting effects, positive or negative.
  • Gradisar et al. 2016, Pediatrics 137(6), 43 infants, cortisol and attachment at 12 months.
  • Douglas and Hill 2013, systematic review, the case against intervention under six months.
  • BASIS, Sleep training, the critique of the evidence base and the Middlemiss cortisol finding.
  • UNICEF UK Baby Friendly, their stated position that controlled crying is not recommended.
  • Trial report, 26 of 91 families discontinuing or not completing.
  • Pinilla and Birch 1993, Pediatrics, the 26-parent focal feed study.
  • Scope, Keeping a sleep diary, the practical tips and why memory is unreliable when you are sleep deprived.
  • Sleep Action, Scotland, gentle strategies without cry-it-out methods.

Last checked 1 August 2026. This page is general information, not medical advice. If you are worried about your child, speak to your health visitor or GP.