HYPE: Sleeping separately improves sleep when sharing a bed is disruptive.
But what does the evidence actually show?
Sleeping separately may help when a partner's noise, movements or schedule is the actual source of disruption. Direct evidence for a general sleep benefit is limited, and some couples sleep well together.
The claim
Sleeping separately improves sleep when sharing a bed is disruptive. The rating below applies to this specific outcome and population, rather than every claim associated with the subject.
Why people believe it
A separate bed or room can remove a disturbance rather than trying to tolerate it. That is a practical idea, but the phrase sleep divorce can turn an individual adjustment into a universal health prescription. The relevant question is what is disturbing each person's sleep, and whether changing the arrangement addresses it. [1] [2] [3]
What the evidence actually shows
A small experiment does not favour separation for everyone
A 2020 sleep-laboratory study compared couples sleeping together and individually. Bed-sharing was associated with more and less-fragmented REM sleep, alongside more limb movements; other sleep stages did not differ significantly. The result challenges the assumption that a partner's presence necessarily makes sleep worse. [1]
Relationship status is not a randomised bedroom trial
A 2021 analysis matched 69 married people with 69 never-married people from an existing sleep study. Married participants had more REM sleep, while most other measures did not differ. It was observational and actual sleeping arrangements were not measured directly. It cannot prove that marriage, bed-sharing or separate bedrooms caused a particular outcome. [2]
The conditional question remains the useful one
Neither study directly establishes the benefit of moving to a separate room among couples selected because one partner repeatedly disrupts the other. That specific situation differs from studying healthy habitual bed-sharers. Evidence on sharing a bed and evidence on removing a known disturbance should not be collapsed into one rule. [1] [2]
The limitations
Laboratory nights are short and unfamiliar. People accustomed to sharing may respond differently when a partner is removed. Observational comparisons also mix relationship, health and lifestyle differences, even after matching. Large, long-term randomised comparisons of sleep-divorce strategies remain lacking in these sources. [1] [2]
Improving one sleep-stage measure is not a guarantee of better daytime functioning or relationship outcomes. The studies do not establish one correct arrangement for every couple, and they do not justify stigmatising either shared or separate sleep. [1] [2]
What this means for you
Identify the actual disturbance and agree on a comfortable, reversible arrangement. A simple diary of awakenings and daytime tiredness can help judge whether a change is useful. Treat it as a shared practical decision, not a diagnostic test of a relationship. This is an application of the evidence's limits, rather than a proven treatment protocol. [1] [2]
Separation does not treat a partner's underlying sleep disorder. Breathing pauses, gasping or persistent daytime tiredness warrant medical assessment for possible sleep apnoea, whether or not the other partner moves rooms. [3]
Research & Sources
1. Drews et al. (2020). Bed-sharing, REM sleep and sleep-stage synchronisation: laboratory study in couples.
Read the original source ↗2. Drews and Drews (2021). Couple relationships and REM sleep: matched observational analysis.
Read the original source ↗3. NHS. Sleep apnoea: symptoms, assessment and treatment.
Read the original source ↗Evidence reviewed: October 2026
This targeted editorial review considered evidence available by 25 October 2026. It is not an exhaustive systematic review or independent clinical peer review. Assessment used available full texts, research abstracts and authoritative guidance; some publisher full texts were inaccessible. Evidence vs Hype provides education, not individual diagnosis or treatment. Featured artwork is AI-generated editorial illustration, not a study result or product endorsement.