Plain-Language Evidence Notes

Research should help people make better decisions. It should not require a reader to decode a journal article, memorize statistical vocabulary, or accept a confident-sounding conclusion without knowing how strong the evidence actually is.

A plain-language evidence note is a bridge between a source and a real decision. It does more than shorten an abstract. It identifies the question being studied, explains what researchers found, shows important limits and uncertainty, describes who or what was actually studied, and turns the result into practical questions without pretending the research can make an individual decision by itself.

This matters in healthcare, education, disability services, public policy, employment, benefits, community planning, and public-safety work. It matters especially when people are under time pressure, overloaded, unfamiliar with technical language, or trying to compare several competing claims at once.

The four questions every evidence note should answer

If a public research summary leaves out one of those questions, readers may still have information but not enough context to use it safely.

Plain language should preserve the science

Clear communication does not mean replacing every technical term with a vague everyday phrase. Some technical terms are necessary. The useful approach is to define them, show why they matter, and keep the sentence around them understandable.

The second version is longer, but it gives the reader more usable meaning. It explains the statistical statement without pretending that “statistically significant” automatically means “large,” “important,” or “right for everyone.”

A study result is not the same as a universal rule

Research describes what was observed in a particular study or body of studies. Real-world decisions involve additional questions: Is the population similar to the person in front of us? Is the setting comparable? Were disabled people, autistic adults, people with intellectual disability, rural participants, people with complex health needs, or other relevant groups represented? Were outcomes measured in a way that actually matters to the people affected?

A useful evidence note makes those boundaries visible instead of hiding them in fine print.

What kind of evidence are you looking at?

There is no single study design that is “best” for every question. A randomized trial may be useful for estimating the effects of an intervention. A cohort study can help examine outcomes over time. Qualitative research can answer questions about experience, implementation, barriers, and meaning that a numerical outcome alone cannot answer. Surveillance data can describe patterns in a population. A systematic review can bring together multiple studies when those studies are sufficiently relevant to the same question.

The important issue is whether the method fits the question—and whether the conclusion stays inside what that method can reasonably support.

Certainty matters as much as the headline

Two summaries can report the same estimated effect but deserve very different levels of confidence. Certainty can be reduced by problems such as small samples, inconsistent findings, missing data, high risk of bias, indirect evidence, imprecise estimates, or important differences between the study population and the people who will use the result.

Evidence frameworks such as GRADE are designed to make certainty explicit rather than treating every published result as equally dependable.

The exact wording should follow the source and the evidence assessment. The purpose is not to attach a decorative confidence label. It is to tell the reader how much weight the finding can reasonably carry.

Association is not automatically causation

If two things occur together, one may cause the other, the direction may run the other way, a third factor may influence both, or several mechanisms may be involved. Good research methods can strengthen or weaken a causal explanation, but a summary should not silently convert “associated with” into “caused by.”

Numbers need a denominator

Percentages can sound dramatic when the starting numbers are missing. Readers should be able to see the size of the effect in a form that can be understood.

When possible, an evidence note should provide natural frequencies—such as “8 out of 100” rather than only a percentage—and should identify the comparison group and time period.

“Statistically significant” is not the finish line

Statistical significance is one part of interpreting a result. It does not by itself tell a reader whether the effect is large, meaningful, clinically important, educationally useful, affordable, accessible, safe, or relevant to an individual.

A practical evidence note should also ask:

• How large was the difference?
• What does that difference mean in everyday terms?
• How precise is the estimate?
• Were there harms or tradeoffs?
• Did the study measure outcomes that matter to the people affected?
• How long did the study follow participants?
• Would the effect matter in the setting where the decision is being made?

Who was actually studied?

This is one of the most important questions for ANCHOR. Evidence may be presented as though it applies to “autistic people,” “children,” “patients,” or “students” even when the actual sample was much narrower.

A study may include only young children, only people without intellectual disability, only people who speak fluently, only participants able to tolerate a research environment, only people living near a major academic center, or only people whose families had the time and resources to participate.

Those results may still be useful. The limitation is in assuming the same estimate automatically applies to people who were not represented.

Outcomes can be technically measurable and still miss what matters

Researchers need measurable outcomes, but the choice of outcome shapes the story the research tells. A program might change a test score without changing daily functioning. An intervention may reduce a behavior that observers dislike while increasing distress for the person. A service may look successful by appointment-completion numbers while remaining exhausting, inaccessible, or unaffordable.

Evidence notes should identify whose outcome is being measured and why it matters.

Lived experience and research evidence answer different questions

Personal experience should not be presented as proof that an intervention works for everyone. Research findings should not be used to erase what a person reports about their own body, communication, environment, safety, or quality of life.

Lived experience can identify outcomes researchers failed to measure, reveal implementation problems, show why a technically available service is not practically accessible, and help determine whether an evidence-based option fits the person and setting.

The useful question is not whether research or lived experience “wins.” It is which type of information is needed for the decision being made.

Research about autism needs an extra applicability check

Autism research spans genetics, development, communication, mental health, education, services, sensory processing, co-occurring conditions, employment, aging, public health, and many other domains. Findings from one domain should not be stretched into claims about another.

The autistic population is also heterogeneous. Support needs, communication methods, sensory profiles, intellectual ability, health conditions, age, environment, and access to services vary substantially.

ANCHOR evidence notes should therefore avoid language that turns a group average into a description of every autistic person.

Local context matters in Arkansas

Evidence developed in a university clinic, large metropolitan hospital, or well-funded school district may not transfer directly to a rural community with long travel distances, provider shortages, limited public transportation, inconsistent broadband access, or different staffing.

That does not make the research irrelevant. It means implementation is part of the evidence question.

A useful Arkansas evidence note can separate what the research suggests from what would be required to make the approach usable here.

One paper is not “the research”

A newly published study can be important without representing the full state of knowledge. When possible, look for systematic reviews, evidence-based guidelines, major evidence syntheses, and whether independent research teams have found similar results.

Also check the date. Some questions change quickly as new evidence accumulates, while other findings remain stable for years. A polished web page can be outdated even when it still ranks highly in search results.

Source quality is more than the domain name

A government, university, journal, professional association, advocacy organization, company, nonprofit, or news outlet can each publish useful information. The organization name alone does not tell you whether a specific claim is well supported.

For the actual claim, ask:

• Is the original research or evidence review identified?
• Can the reader tell what kind of study it was?
• Are limitations discussed?
• Are benefits and harms both addressed when relevant?
• Are financial interests or funding relationships disclosed?
• Is the information current enough for the question?
• Does the source distinguish evidence from opinion, recommendation, or organizational policy?

Preprints, press releases, news stories, and social posts

These can be useful for discovering new research, but they are not interchangeable with the underlying evidence.

For a high-stakes decision, move toward the original source and then toward the larger body of evidence around it.

Conflict and disagreement are not automatically signs that science failed

Studies can disagree for legitimate reasons. They may enroll different populations, use different definitions, measure different outcomes, follow people for different lengths of time, use different statistical models, or evaluate interventions that sound similar but are not actually the same.

A good evidence note does not hide disagreement. It explains the most important reasons the evidence may differ and identifies what would resolve the uncertainty.

What “no evidence” actually means

“No evidence of benefit” can mean several different things: high-quality studies found little or no benefit; studies were too small or imprecise to know; the question has barely been studied; or the available evidence does not match the population or outcome being considered.

Those situations should not be collapsed into the same sentence.

From research finding to practical question

Evidence notes become especially useful when they give the reader language to carry into a real interaction.

Teach-back works for evidence, too

In healthcare, teach-back is used to check whether information was communicated clearly. The same principle can be used in research communication.

Instead of asking, “Do you understand?”, ask the reader or participant to describe the decision in their own words: What do we know? What are we still unsure about? What are the options? What happens next?

If the explanation cannot be repeated accurately, the problem may be the explanation—not the person.

An ANCHOR evidence note should be layered

Not every reader needs the same amount of detail. A useful public page can make the essential answer visible first and let readers go deeper without forcing everyone through a wall of text.

A reusable evidence-note template

Terms that should not be left unexplained

For appointments

A reader should be able to leave an ANCHOR evidence page with a short set of questions instead of a pile of vocabulary.

For schools and service teams

Evidence should support individualized planning rather than become a shortcut for saying, “the research says we always do this.”

Useful team questions include:

• What problem are we actually trying to solve?
• What outcome matters to the student or person receiving support?
• What evidence supports this specific strategy?
• Was the strategy studied with people who have similar communication and support needs?
• How will we know if it is helping rather than merely changing what adults observe?
• What accommodations are required for the strategy itself to be accessible?
• When will we review the plan?

For public policy and program planning

Public decisions require more than asking whether an intervention worked somewhere. Decision-makers also need to know how large the effect was, what implementation requires, how benefits and burdens are distributed, what the evidence leaves uncertain, and whether the program can be delivered under Arkansas conditions.

A policy evidence note should separate at least three things:

Mixing these together can make a policy preference look like a research conclusion or make a research finding look like it automatically dictates one policy.

For journalists, advocates, and public educators

The headline should not be stronger than the study. If the article says “linked to,” the headline should not say “causes.” If the sample was 42 adults at one clinic, the headline should not imply that the finding describes every autistic person. If the result is preliminary, the reader should know that before the final paragraph.

Plain language is an accessibility practice

Dense information can become harder to process during illness, pain, sensory overload, crisis, fatigue, stress, or an unfamiliar appointment. Clear structure benefits people across literacy levels and is especially important when the consequences of misunderstanding are high.

Useful presentation choices include short sections, descriptive headings, everyday words, defined technical terms, clear numbers, meaningful examples, enough white space, printable summaries, text-to-speech compatibility, and Spanish-language access.

Accessibility also means not forcing every reader through the same amount of information. The central conclusion should be easy to locate, while deeper methods and source material remain available to readers who need them.

The ANCHOR standard for an evidence note

The goal is not to make every reader into a statistician. The goal is to give people enough information to recognize what a claim can support, what it cannot support, and what they should ask next.