Kent de Grey · Annals of behavioral medicine : a publication of the Society of Behavioral Medicine 2019 · cross-sectional dyadic study · n=184 (92 couples)

Strained Bedfellows: An Actor-Partner Analysis of Spousal Attachment Insecurity and Sleep Quality.

Level 4 - case-series / case-control

Cross-sectional observational dyadic study without intervention or longitudinal follow-up

PubMed 29788062 · doi:10.1093/abm/kay037 · record verified 2026-08-26

What was done

The study examined the associations between adult attachment insecurity (anxiety and avoidance) and sleep quality in 92 married heterosexual couples (184 individuals). Attachment was measured with the Experiences in Close Relationships (ECR) scale, and sleep quality was evaluated using the Pittsburgh Sleep Quality Inventory (PSQI). Actor-partner interdependence models were used to analyze dyadic effects, controlling for marital satisfaction and evaluating depression as a potential mediator.

What was found

Actor attachment anxiety predicted poorer individual sleep quality, whereas actor avoidant attachment was unrelated to individual sleep quality. At the dyadic level, couples where both spouses reported higher attachment anxiety experienced better sleep quality (b = -0.74, SE = 0.28, p = .0082, 95% CI [-1.287, -0.196]). Conversely, couples where both spouses reported higher attachment avoidance had poorer sleep quality (b = 0.56, SE = 0.23, p = .0188, 95% CI [0.095, 1.016]). These effects remained independent of marital satisfaction and depression. Depression showed evidence of acting as a mediator.

Why it matters

This study shows that the association between romantic attachment and sleep quality depends on dyadic combinations of partner characteristics, with shared anxiety and shared avoidance showing opposite links to sleep quality.

Limits

The study used a cross-sectional design, precluding causal or directional inference. The sample size was modest (92 couples) and restricted to heterosexual married couples. Sleep quality was evaluated solely through retrospective self-report rather than objective measures such as actigraphy or polysomnography.