Parental separation is common and the research on child outcomes is frequently misrepresented in both directions — as either catastrophic or inconsequential.
The more useful finding is about what varies within that population, because those factors are largely modifiable.
What the research broadly finds
On average, children whose parents separate show somewhat poorer outcomes across various measures than children whose parents don't.
The effect sizes are generally small, the distributions overlap substantially, and the majority of children of separated parents do not show significant problems.
There are also important confounding issues. Families that separate differ from those that don't in ways that predict outcomes independently — conflict levels, financial circumstances, parental mental health. Some studies have found that differences in child outcomes were detectable before the separation occurred, which complicates simple causal accounts.
And the comparison matters: children in high-conflict intact families frequently show worse outcomes than children whose parents separated. Separation from a high-conflict situation can be protective.
What predicts the difference
Consistently across the literature, a few factors emerge.
Ongoing conflict between parents. The strongest single predictor. Conflict that continues after separation — particularly conflict involving the children, or which they witness — is associated with worse outcomes than the separation itself.
Children placed in the middle, asked to carry messages, or hearing one parent criticise the other, are in the situation most reliably associated with harm.
Quality of parenting. Warm, consistent, involved parenting is protective, and it's frequently the thing that suffers when adults are in crisis. Maintaining routines and responsiveness through a difficult period is genuinely difficult and genuinely matters.
Continued relationship with both parents, where safe. The evidence associates ongoing meaningful contact with better outcomes, with quality mattering more than the precise division of time.
Economic circumstances. Separation frequently produces a decline in household income, particularly for the parent with primary care. Financial hardship independently predicts poorer outcomes and accounts for part of the association.
Stability. Multiple transitions — repeated changes in household composition, moves, changes of school — are associated with worse outcomes than a single transition.
What matters less than assumed
The precise arrangement. Considerable conflict occurs over exact division of time. The evidence suggests the arrangement matters less than the quality of the relationships and the level of conflict.
Shared arrangements are associated with good outcomes in many studies, and the association appears to be substantially explained by the fact that lower-conflict couples are more likely to adopt them.
Whether parents were married. The transition matters more than the legal status.
The stated reason. Children are generally more affected by how things are handled than by why it happened.
Practical guidance
Things with reasonable support behind them.
Keep children out of the conflict. No messages passed through them, no criticism of the other parent, no asking them to choose or to report. This is the single highest-value thing.
Tell them together, if possible, and simply. Age-appropriate information, clear that it isn't their doing, clear about what will practically change.
Repeat it. Children process this over time and will need to ask again. A single conversation isn't sufficient.
Maintain routines. Predictability is protective during a period when a great deal is uncertain.
Don't rely on them for support. Children, particularly older ones, frequently take on a supporting role for a distressed parent. This is associated with poorer outcomes and it's an easy thing to do without noticing.
Get your own support. Your functioning is one of the main variables. This is not self-indulgence.
Communicate practically with the other parent. Business-like arrangements about logistics, in writing where verbal contact is difficult. Parallel parenting — minimal contact, separate approaches — is a legitimate model where cooperation isn't possible and is better than continued conflict.
What children commonly need to hear
Repeatedly, and in various forms.
That it isn't their fault. Younger children in particular frequently construct explanations involving their own behaviour.
That both parents still love them and this isn't changing.
That they don't have to choose or take sides.
That it's alright to be sad or angry about it.
That it's alright to enjoy time with either parent, and to say so.
When to seek support
Persistent changes in mood, behaviour, sleep or school performance beyond the initial adjustment period.
Regression in younger children that doesn't resolve.
Withdrawal, or expressions of self-blame that don't respond to reassurance.
Any concern about safety.
Support for children through parental separation exists in many places, and involving a professional early is more effective than waiting to see whether difficulties resolve.
General information only. Concerns about a child's wellbeing should be discussed with a qualified professional.
New partners
A transition with its own evidence and one that is frequently handled badly through haste.
Research on repartnering suggests that multiple transitions in household composition are associated with poorer outcomes than a single stable arrangement, which argues for taking time rather than for avoiding new relationships.
Practical guidance that emerges from the literature: introduce a new partner only once the relationship is established, keep the introduction low-key rather than significant, and do not expect or require a child to form a relationship on any timetable.
Children frequently experience a parent's new partner as a threat to their own relationship with that parent, particularly early. Protecting individual time with the child through the transition addresses that more effectively than any amount of encouragement to get along.