Fathers of
autistic children
report more
distress than
mothers —
and the more
affirming
they are, the
heavier it gets

What was found — distress higher, and burden rising with affirmation

The parental-wellbeing literature in autism has a default subject, and she is a mother. So a small study in Research in Neurodiversity from Liza Anderson, Finley Breazeale and Grace Binion — 91 American parents of autistic children aged two to eighteen, 63 mothers and 28 fathers — earns attention for who it asked before what it found. What it found is stranger than the design.

First: fathers reported significantly more psychological distress than mothers (b = −12.02, p = .045). That inverts the expectation the field’s own prior work set up, in which mothers, carrying more of the caregiving, carry more of the cost. The authors note earlier evidence pointing the same way — elevated physiological stress markers in fathers, a tighter link between paternal stress and mental health — but the headline remains: in this sample, the parent the literature ignores was the parent reporting the worse time.

Second, and stranger: neurodiversity-affirming attitudes did not behave like armour. The foundational study here — Ferenc and colleagues, 2023 — found that mothers who saw their child’s autism as difference rather than deficit reported less distress and burden, and mothers and fathers in this sample endorsed affirming attitudes at similar rates. The raw association with lower distress appeared here too. But once the models controlled for the child’s autism characteristics and the parent’s social support, the attitude–distress link stopped being significant. What predicted distress was the child’s support needs and how much backup the parent had — the same variables the Directory found doing the work in parental resilience and outcomes. Attitude, in this sample, was not an independent shield.

For caregiver burden, the sexes split. For mothers, affirming attitudes and burden were unrelated. For fathers, more affirming attitudes predicted more reported burden (b = .47, p = .032). The authors’ interpretation is careful and plausible: mothers tend to become the household’s autism expert regardless of ideology, so for fathers — typically more peripheral to caregiving — endorsing neurodiversity may be a marker of actually engaging, and engagement is what the burden scale detects. On that reading the finding is not that affirmation harms fathers. It is that the disengaged father reports no burden for the same reason a person who never lifts anything reports no back pain. Which reframes the deficit entirely: the family system’s problem is not affirming fathers who feel the weight, but the structure that assigned the expertise to one parent in the first place — the same system the landmark parent-treatment trial showed runs through the whole household, and the same one visible in how mothers’ expectations get studied while fathers’ absence doesn’t.

One more detail worth keeping: general neurodiversity attitudes — beliefs about neurodivergent people at large — related more strongly to parental functioning than autism-specific attitudes about one’s own child. How a parent thinks about the world beyond their caregiving apparently matters more to their wellbeing than how they frame their own child. Ideology, it turns out, is load-bearing.

The instrument problem — burden, according to 1980

Now look at the ruler. Caregiver burden here was measured with the short-form Zarit Burden Interview — an instrument developed in 1980 to quantify the strain on relatives of the impaired elderly. Psychological distress came from the DASS. The child’s autism was quantified with the GARS-3, a symptom checklist keyed to DSM-5 criteria; the authors, to their credit, explicitly reinterpret its scores as a count of characteristics and support needs rather than pathology, which is more than most of the literature manages.

But an instrument built to measure the burden of tending decline cannot help importing that frame when pointed at the parents of a six-year-old. Every question it asks assumes the cared-for person is a weight; the only variable is how heavy. There is no item on which a child’s monotropic joy, communication or company registers as anything but absence-of-strain — no channel through which a father’s engagement could show up as something other than burden. So when engaged, affirming fathers score higher on the ZBI, part of what is being detected may simply be that the scale has one register, and involvement lands in it. This corpus has made the same point about engagement scales that measure the mask and about self-tracking that converts neurodivergent life into deficit ledgers: the instrument encodes the verdict, and here the verdict — this child is a load — was cast in 1980, for a different relationship, at the other end of life.

None of that makes the distress finding artefactual. The DASS asks about symptoms, not caregiving, and fathers still reported more of them. The population deserves the attention: paternal mental health in autism families connects to threads this corpus has followed in male loneliness and social connection, in the anxiety and depression that lifelong friction produces, and — given heritability — in the strong possibility that a portion of these fathers are undiagnosed neurodivergent men measuring their own perinatal-era abandonment by services without a word for it. The study did not ask whether the parents were neurodivergent. In 2026, in this field, that is a remarkable question to leave unasked.

What 28 fathers can and cannot carry

Scale the claims to the sample. Ninety-one parents; twenty-eight fathers; recruitment through advertisements, so the fathers who opted in are plausibly the engaged subset; race and socioeconomic status not collected; cross-sectional, so causation runs uncontrolled in every direction — parents with better mental health may adopt affirming views rather than be protected by them, as the authors say themselves. The interaction result needs replication before anyone builds a programme on it, and the mother–father comparisons carry wide intervals. This is a signal, not a structure.

But the signal is structurally instructive, because of what it does to a comfortable assumption. The affirmation-as-protection story was always convenient: it costs institutions nothing to recommend an attitude. What this data sketches instead is that attitudes wash out once you account for support and the child’s actual needs — that is, once you account for the things that cost money — and that for the parent whose engagement is optional, taking the affirming stance seriously means picking up weight the system left on the floor. An ideology that survives only in the parent who doesn’t do the care is not protection; it is distance.

Two things follow, and the first is a guard. Nothing here is a charge against any parent. The disengaged father is inside the same structure as everyone else in this account — the workday without walls, the division of caregiving labour he inherited rather than chose, an entire research literature that never thought to ask him anything until now. Siting a structural condition in the nearest available individual is the same move as siting it in the child, and I decline it in both directions. There is no villain in a family carrying more than the state will fund.

The second is the shape of the whole error. Every system in this chain measures participation: hours of care delivered, symptoms rated, burden scored, attitudes endorsed. What matters to a person is something else entirely — whether their life holds together, whether the steering is their own, whether what they do alters what produced the difficulty. Systems measure only what matters to systems, and a research programme that keeps measuring the parent’s model of their child has chosen the cheapest available variable. Parents are not distressed because they hold the wrong picture of their child. They are distressed in proportion to how much support their child needs and how little of it anyone else provides, and no attitude survey will ever be permitted to return that answer, because the answer has a price attached.

Citations

Anderson, L., Breazeale, F. & Binion, G. (2026) — The associations of mothers’ and fathers’ neurodiversity-affirming attitudes with psychological distress and caregiver burden — Research in Neurodiversity

Ferenc, K., Byrka, K. & Król, M. E. (2023) — The spectrum of attitudes towards the spectrum of autism and its relationship to psychological distress in mothers of children with autism — Autism

Zarit, S. H., Reever, K. E. & Bach-Peterson, J. (1980) — Relatives of the impaired elderly: correlates of feelings of burden — The Gerontologist