Life Expression Lab | Does Maternal Depression Cause School Attendance Difficulties? Why Two Birth-Cohort Studies Say Schools Should Stop Blaming Families

# Life Expression Lab | Does Maternal Depression Cause School Attendance Difficulties? Why Two Birth-Cohort Studies Say Schools Should Stop Blaming Families

Author: 王莉如

This feature separates source-supported findings, editorial interpretation, and possible local implications. Its evidence boundary is explicit: Observational longitudinal data cannot exclude shared factors such as household economics, child health, or school conditions; mothers must not be treated as the sole caregivers or the cause of attendance difficulty.

The headline answer: the study did not show that mothers cause non-attendance

The study's central contribution is a test of temporal direction, not the identification of somebody to blame. It used two prospective UK birth cohorts: 6,375 participants from ALSPAC and 718 from WCHADS. Across the analyses, earlier attendance-related difficulties were associated with later maternal depressive symptoms, while earlier maternal depressive symptoms did not predict later attendance difficulties. That pattern supports the proposition that managing a child's attendance difficulty can add psychological strain for caregivers. It does not say that every family follows the same pathway, and it is not an individual causal diagnosis.

It would still be inaccurate to compress the finding into “school refusal makes mothers depressed.” ALSPAC relied on questionnaire-derived proxy measures for school avoidance and school-related separation anxiety, while WCHADS used a broader attendance-difficulty indicator. Longitudinal ordering is stronger evidence than a one-time correlation, but it cannot eliminate every shared influence, including poverty, child health, special educational needs, school climate, transport, housing, or the distribution of care work. The defensible conclusion is narrower: the familiar assumption that maternal distress comes first and produces the child's attendance problem was not supported by these two analyses.

Attendance difficulty is neither one diagnosis nor one behaviour

A child who reports stomach pain in the morning, panics at the gate, repeatedly arrives late, asks to go home, or misses only a particular class may be responding to very different conditions. Possible pathways include separation anxiety, peer evaluation, bullying, sensory overload, unsafe relationships, sleep problems, chronic illness, care responsibilities, or unreliable transport. Treating all of them as simple refusal reduces communication to compliance. A functional first assessment asks when the difficulty occurs, what the child expects will happen, which setting feels unsafe, and under what conditions participation becomes more possible.

Attendance percentages are useful signals, but they do not explain themselves. A rise in lateness, symptoms concentrated on school days, pronounced Sunday-night anxiety, or a decline in learning and social participation can justify an early, non-punitive conversation. Waiting until a statutory threshold is reached often means that a family has already spent its reserves of leave, supervision, employment flexibility, and clinical appointments. A better response begins with health and safety, maps classroom, peer, transport, and household pressures, and sets small, reviewable participation steps with the student.

Measuring mothers does not make mothers the only caregivers

Birth cohorts often followed mothers because of historical study designs. That is a measurement decision, not an ethical rule about who should carry care. These analyses did not comprehensively compare fathers, grandparents, siblings, foster carers, or other primary caregivers. A policy response framed only as “maternal education” would miss many people doing the work and convert institutional gaps into women's private responsibility. The more faithful translation is that the care system needs support: flexible employment, respite, confidential counselling, case coordination, and working referral pathways between education and health services.

The result also does not make caregiver mental health irrelevant. A parent or carer living with depression, anxiety, or exhaustion deserves accessible help that is not conditional on accepting blame. When one family's chronology differs from a population average, the difference is evidence to investigate. The child's difficulty may have preceded caregiver distress; both may reflect another pressure; or they may reinforce one another. Keeping multiple pathways open is what allows support to follow actual need instead of the person most visible to an administrative system.

From notices to relationships: what schools can do first

The UK Department for Education communication guide recommends messages grounded in belonging, friendships, and the child's daily experience, adapted to age, special educational needs, and prior attendance patterns. It does not claim that a friendly message resolves complex absence. For more complex reasons, communication has to sit inside deeper relationship-building and support. Practical steps include one stable contact, a quiet arrival point, a flexible arrival window, a phased timetable, access to a trusted adult, and a scheduled review so temporary adjustments do not become indefinite abandonment.

Harmful practices share a pattern: generic warnings without listening, public naming, recording help-seeking as non-cooperation, threatening sanctions to obtain attendance, or demanding an immediate return to a full timetable as the only acceptable outcome. They may improve a surface metric briefly while increasing shame, concealment, and disengagement. A support plan should record the student's account, workable times, triggers, effective adjustments, and caregiver load, with a clear statement of who can see each item and why it is being collected.

AI and school systems must not flatten context into a risk score

Attendance data look objective but embed policy choices: how lateness is counted, whether illness is accepted, whether remote participation counts, whose account is recorded, and whether reasonable adjustments exist. If those records feed a model that predicts “high-risk families,” a mistaken label can travel through pastoral, welfare, and even legal processes. Minimum safeguards include purpose limitation, data minimisation, role-based access, audit logs, retention limits, human review, and an effective correction process. Mental-health details should not become school-wide fields or training data by default.

Context and uncertainty must remain attached. “Five days absent” should be linked to transport failure, health needs, a bullying report, or an agreed adjustment where relevant, rather than surviving only as a red alert. Model outputs should disclose dates, missing fields, and plausible bias; they are not diagnoses. Students and caregivers need to know who accessed the record, how a decision was made, and how they can add their account. A useful digital tool helps teams remember promises and follow-up actions. It does not automate who deserves belief.

Implications for Taiwan: transfer the support principle, not a UK causal estimate

The UK cohorts cannot supply prevalence or causal estimates for every school in Taiwan. Leave rules, examination pressure, counselling capacity, work patterns, rural transport, and social protection differ. A responsible local application would begin with participatory definitions and local evidence, bringing students, caregivers, teachers, counsellors, special education, and health services into the design. Measures should include school climate, transport, service waiting times, belonging, anxiety, family load, sustained participation, and referral accessibility—not attendance alone.

The evidence supports one clear public choice: when attendance becomes difficult, provide understanding and coordination before assigning fault, and address health, safety, and barriers before sanctions. It does not support the reverse simplifications that mothers never matter or that children cause maternal depression. The research is most useful as a prompt for testable institutional questions: Did we notice early? Could the student speak safely? Could every caregiver access help? Did we include the possibility that school conditions themselves were contributing?

Sources and further reading

• Archives of Women’s Mental Health 原始研究
• University of Bristol 研究說明
• 英國教育部出席溝通指引