Guide 03 · Weighing it for your family

How to read, weigh, and act on autism research.

Two perspectives, the skill of reading science, and what your family can do right now.

A note before you read This is general education about how to make good decisions — not advice about any one child, and not a reason to start, stop, or change any therapy on your own. Bring what you read here to your child's care team.

The big picture The “science side” and the “autistic community side” are usually described as opponents. On the choices you actually face this week, they mostly agree. Learning to read evidence is a skill you can build — with clear limits on where to lean on experts.

What the research and clinical world asks of you

Experts have built lists of “evidence-based practices” (techniques that cleared a bar of supporting studies). The current national list names 28 such practices, with free how-to guides. A list is a starting map, not a prescription: it says which approaches have some credible support for some goals, and the right choice still depends on your child.

Leading researchers say recommendations should shift as studies improve, and that much of what is “commonly recommended” rests on weak evidence. Even the official lists have published critiques. So the ask is: prefer tested approaches, and stay open to updates.

A testimonial or a website is not the same as a study, and newly-diagnosed families are the most targeted. Watch for: cure or recovery promises, testimonials instead of data, “no side effects,” secrecy or “secret ingredients,” very high cost, a lone provider against the mainstream, and anything that claims to treat many unrelated conditions at once.

Chelation, hyperbaric oxygen chambers, chlorine dioxide (“Miracle Mineral Solution,” which is bleach), detox clay baths, and essential-oil “cures” have been formally warned against — and some have killed children. (The U.S. FDA page listing these was taken down in late 2025; the warnings themselves remain valid, so lean on independent groups like ASAT and the Autism Science Foundation.)

One study rarely settles anything. You want the weight of good studies pulled together, read alongside your clinician's judgment and your family's values.

What the autistic community asks of you

Autistic adults ask to be part of decisions about autism — including a parent's. In practice: seek out autistic adults' perspectives, not only professionals', when deciding how to support your child.

The neurodiversity view treats autism as a natural variation, not a disease to cure. That shifts the aim from making a child seem less autistic toward supporting communication, easing distress, building skills the child values, accommodating differences, and protecting autonomy and wellbeing. Be cautious of any program whose main measure of success is “looks more typical.”

Especially anything built around compliance or forced eye contact rather than the child's comfort and communication. This is partly an evidence question and partly a values question — keep them separate. A useful idea here is the “double empathy problem”: misunderstanding between autistic and non-autistic people runs both ways, so social difficulty is not simply a flaw inside the child.

Quality of life, mental health, self-determination — not only symptom counts.

Where the two sides actually agree

Both reject cure-quackery and dangerous “detox” treatments — autistic advocates have been as loud as regulators about the dangers of chelation and chlorine dioxide.
Both distrust testimonials and marketing, and favor real evidence plus lived experience.
Both want autistic people and families helping set what gets researched.
Both put the child's wellbeing and quality of life at the center.

The real disagreement is narrow: it is mostly about the goal of intervention and about specific methods. That is a values conversation, and values are not settled by evidence grades — which both sides would agree on.

How to actually read a study

The core skill. The evidence ladder, weakest to strongest, for “does it work” questions:

Expert opinion / testimonial

A hunch at best; cannot show a treatment works.
Case report or small series

No comparison; good for spotting harms, cannot prove benefit.
Observational study (cohort, case-control)

Shows associations, not causes. “Correlation is not causation”: two things moving together does not mean one caused the other.
Randomized controlled trial (RCT)

Randomly assigns children to treatment vs comparison; this is what lets you say the treatment caused the change.
Systematic review / meta-analysis

Pulls together all qualifying studies by an explicit method; usually the strongest, because it is the whole picture, not one lucky study.

Three honest cautions: the ladder is not absolute (a large, well-run RCT can beat a weak review); a review is only as good as the studies inside it; and reviews can be out of date.

What makes a treatment study trustworthy

A fair comparison group — or you cannot tell if the child would have progressed anyway.
Blinding and who rated the outcome — an unblinded parent or teacher who knows the child got the therapy tips results optimistic; blinded independent measures are stronger. (Unblinded parent-report is the most common soft spot in autism research.)
Sample size — a handful of children can produce a dramatic number by chance.
Effect size vs “significant” — “significant” only means “probably not chance”; ask how big the change was and whether it matters in daily life.
Follow-up length — short-term gains often fade; lasting change is the harder, better bar.
Who funded and ran it — conflicts of interest are common and under-reported here; a study by the people who sell the therapy deserves extra scrutiny.
Replication — one result is a lead; repeated results are a finding.

Fast credibility checks anyone can do

Peer-reviewed in a reputable journal, or just posted online? (Anyone can post a “study.”)
Is it a preprint (not yet peer-reviewed)? Treat as provisional.
Retracted or flagged? Retraction Watch and the journal will say.
Is there a plain-language summary from a trustworthy synthesizer (Cochrane, ASAT's “Is There Science Behind That?”, evidence maps)? Start there.

How much can a parent interpret alone?

The honest boundary.

You can reasonably learn to

Tell a testimonial from a study
Rank study types
Ask the trustworthiness questions
Spot red-flag claims
Read a plain-language summary

Generally specialist work

The technical risk-of-bias appraisal of a trial
Judging whether a meta-analysis pooled correctly
Weighing conflicting studies against each other

What you can learn is genuinely valuable. Two traps to avoid:

“No evidence yet” is not “evidence it fails,” and neither is “evidence it works.” There are four different states — not studied / studied but too little to judge / studied and shown ineffective / shown dangerous — and collapsing them loses the most useful distinction you can make.
Do not over-act on a single study or preprint. Bring it to your care team instead of changing course on it alone.

The healthy stance is a division of labor: you learn enough to ask sharp questions, detect nonsense, and know your child; you lean on synthesizers and clinicians for the technical appraisal; and you combine their read of the evidence with your values and your specific child. That combination is what “evidence-based” actually means. It was never “the study decides.”

Healthy practices you can use right now

Run any proposed therapy through this question set:

What is the evidence, and what kind of study is it — a testimonial, one small study, or a systematic review?
What exactly improved, and how was it measured — by a blinded observer, or by someone who knew the child got the treatment?
What are the possible harms and side effects? (Be suspicious of “none.”)
Who is likely to benefit, and is my child like the children who were studied?
What does it cost in money, time, and opportunity — what are we giving up to do it?
What happens if we wait, or do nothing, for now?
Who delivers it and what training do they need? Are the claims “quick fix / miracle / secret ingredient”? If so, walk away.

Watch two traps: sunk cost (“we've spent so much we can't stop”) and opportunity cost (time and money on an unproven therapy is time and money not spent on a tested one). Naming them helps.

Set measurable, individual goals with your team

Track your child's own response — decide in advance what “working” looks like, in observable terms. Its limit: your own before/after impression is unblinded, so treat it as a signal to discuss, not proof.

Practice shared decision-making and get second opinions

Ask directly: “How strong is the evidence, and what would you do?” A good provider welcomes it.

Learn from autistic adults, not only professionals

They often see which “improvements” actually matter to the person.

Separate acceptance from intervention, and do both

Accepting your child as they are and pursuing support for specific challenges are not opposites.

Protect your own wellbeing

Caregiver stress affects your capacity to decide well; your support is part of the plan, not a luxury.

The choice stays with your family.

The single healthiest habit is a fact about method: decide by combining graded evidence, clinical judgment, autistic-informed values, and your specific child — and distrust anything that asks you to skip that combination. That is not being told what to choose. It is knowing how to choose.

Where this comes from

Guidance from research and clinical bodies, autistic-led organizations, and the standard guides to weighing evidence.

Steinbrenner JR, et al. Evidence-Based Practices for Children, Youth, and Young Adults with Autism. National Clearinghouse on Autism Evidence and Practice (NCAEP), UNC Frank Porter Graham Child Development Institute. 2020. NCAEP
AFIRM modules (how-to guides for the evidence-based practices). AFIRM
National Autism Center. National Standards Project (2009; 2015). May Institute.
Leaf JB, et al. The evidence-based practices report: concerns and critiques. Behavioral Interventions. 2021. Journal
Sandbank M, Bottema-Beutel K, Woynaroski T. Intervention recommendations in light of a changing evidence base. JAMA Pediatrics. 2021;175(4):341–342.
Bottema-Beutel K, et al. Conflicts of interest in autism early intervention research. Journal of Child Psychology and Psychiatry. 2021;62(1):5–15. PubMed
Association for Science in Autism Treatment. Becoming a Savvy Consumer. ASAT
Autism Science Foundation. Beware of non-evidence-based treatments. ASF
On the removed FDA consumer warning about dangerous autism “cures”: Undark, February 2026. Undark
Autistic Self Advocacy Network. What We Believe. ASAN
Milton D. The double empathy problem. Disability & Society. 2012;27(6):883–887.
Pellicano E, Dinsmore A, Charman T. What should autism research focus upon? Autism. 2014;18:756–770.
Greenhalgh T. How to Read a Paper. Wiley; and university evidence-based practice guides. UC Davis levels of evidence
Cochrane plain-language summaries; Retraction Watch. Cochrane · Retraction Watch