Night shifts: Revisiting blanket restrictions on children’s overnights with separated parents
Journal of Divorce & Remarriage
Report summary
Do the studies used to support blanket overnight restrictions justify those restrictions when their designs and results are examined closely?
Warshak argued that prominent cautionary claims exceeded what the observational designs, measures, and reported findings could establish, and that overnight decisions should not be governed by blanket age rules.
No single overnight study settles policy. The more defensible synthesis is conditional: relationship history, caregiving competence, conflict, safety, transitions, and developmental stage all matter.
Researchers' conclusions
The review supported regular overnights in ordinary circumstances while retaining exceptions for safety, poor care, and other child-specific concerns.
This is a plain-language summary of the researchers' conclusion. It is not advice for an individual family.
How the report was produced
Studies differ in age bands, frequency thresholds, comparison groups, prior caregiving, conflict, measurement, and whether outcomes are concurrent, later, or retrospectively reported.
What this paper supports
Reasonably supports
- Treating prior hands-on caregiving and the child’s familiarity with each parent as central evidence.
- Considering frequency, transitions, conflict, safety, parental responsiveness, and developmental stage together.
- Rejecting claims that one observational paper proves a universal rule for all infants or toddlers.
- Monitoring the child’s functioning and adapting arrangements rather than treating a schedule as an entitlement or diagnosis.
Does not establish
- A claim that any overnight with a second familiar parent necessarily disrupts attachment.
- A claim that equal overnights are proven optimal for every infant regardless of care history or family conditions.
- Using an attachment classification by itself to prescribe a custody schedule.
- Ignoring violence, abuse, neglect, coercive control, severe conflict, or a caregiver’s lack of established competence.
Important limitations
- Critical review rather than a preregistered systematic review
- Interpretive judgments differ from authors of the studies being critiqued
- No new causal evidence was generated
Evidence context and debate
Open the sections below for the wider research context, points of agreement, and areas that remain contested.
Why this question matters
The debate intensified when attachment concepts were used to support either restriction or expansion of infant overnights despite a small, heterogeneous direct evidence base.
The overnight-care research does not come from one experiment producing one answer. Different studies asked different questions of different families. Solomon and George examined attachment classifications in small separated-family groups and a married-parent group. The Australian work examined parent-reported settling and emotional regulation. Tornello and colleagues studied a large urban group and raised an infant attachment concern but found no link with emotional or behavioural problems at ages three and five. Fabricius and Suh asked young adults and parents to remember earlier schedules and later relationships. These findings concern different outcomes and should not be treated as if they measured the same harm.
The studies also define exposure differently. One or more nights per week, frequent overnights, shared care, and approximately equal care are not the same category. Children entered arrangements with different histories of daytime care, parental conflict, cooperation, breastfeeding, socioeconomic conditions, and familiarity with each parent. Because families were not randomly assigned, a schedule may be associated with an outcome while prior relationships or family conditions account for part of that association.
The most useful synthesis therefore does not ask whether an overnight is inherently safe or harmful. It asks which children, at what developmental stage, with what established caregiving relationships, under what transition pattern, and in what safety and coparenting environment experienced which outcome. Pruett, McIntosh, and Kelly’s 2014 integration and the later international attachment consensus both move in this conditional direction, even though authors still differ over how cautious the starting position should be.
Where researchers agree and disagree
Most authors reject a universal schedule and agree that safety, familiar caregiving, relationship continuity, and the quality of coparenting context are relevant.
The sharpest disagreement concerns how much caution infant findings justify, how heavily to weight competing observational studies, and when association warrants policy guidance.
Comparison with other research
Solomon & George, 1999
Reported group differences in attachment classifications and found that disorganization within the overnight group was associated with poor communication, conflict, and low reported psychological protection. The small, non-equivalent groups make it difficult to isolate overnights from family conditions.
McIntosh and colleagues, 2010
Reported regulation concerns for some very young children in shared-care patterns. Critics dispute whether the measures, thresholds, and analyses justify broad policy conclusions, while supporters view the findings as a reason for developmental caution.
Tornello et al., 2013
Infants who frequently stayed overnight were more often classified as insecurely attached. The toddler result was uncertain, and frequent overnights did not predict more emotional or behavioural problems at ages three and five.
Fabricius & Suh, 2017
Reported that more overnights were associated with stronger later relationships with both parents after statistical controls. The retrospective design addresses longer-term relationships but cannot independently establish that overnights caused them.
Warshak, 2014 and 2018
Argued that cautionary studies were overextended into blanket restrictions and gave insufficient weight to father-child attachment and the risks of weakened relationships. These are evidence reviews and critiques, not randomized outcome studies.
Pruett, McIntosh & Kelly, 2014
Sought common ground by treating attachment security and early involvement as mutually attainable. Their framework supports both relationships while retaining caution where care is unfamiliar, conflict is frequent, or the child’s regulation indicates difficulty.
How the evidence developed later
The 2014 integration papers replaced a simple attachment-versus-involvement conflict with a triadic secure-base framework.
In 2015, McIntosh, Smyth, and Kelaher defended the purpose and limits of the Australian study while rejecting the claim that it supported primary maternal care or discouraged shared parenting by fathers.
McIntosh's 2015 to 2017 work widened attention toward what helps children feel secure in shared-time arrangements and toward parent education based on the circumstances of each young child.
Warshak’s 2018 Night Shifts review made the methodological dispute explicit and challenged blanket age restrictions.
McIntosh and Booth's revised 2021 guidance retained conditional infant caution while stating that the long-term benefit of safe involvement by both parents was not in question.
The 2022 international consensus endorsed familiar non-abusive caregivers, continuity of good-enough care, and attachment networks while declining to derive exact schedules from attachment theory.
McIntosh's 2024 population cohort examined attachment change and caregiving sensitivity rather than post-separation schedules. It found no difference in caregiving sensitivity scores for mothers and fathers of preschool children.
Read the paper and related research
This report
Richard A. Warshak (2018). Night shifts: Revisiting blanket restrictions on children’s overnights with separated parents. Journal of Divorce & Remarriage, 59(4), 282–323. https://doi.org/10.1080/10502556.2018.1454193
Record checked against the linked scholarly source. Last editorial review: 22 July 2026, as part of a batch review of the current collection.