Most research does not fail at the moment of collection. It fails at the moment of explanation, when a careful finding lands in front of people who did not read the method section and never will.
So here is the direct answer: the reliable way to explain research findings to non-researchers is to name the decision the finding informs, say what changed in ordinary words, give it a comparison point, state what it cannot show, and finish with the next step. Method comes last, or not at all. This is the working version of that, and it holds up in 2026 the same way it did the first time you ran it.
Table of Contents
- What You Need
- Step-by-Step: How to Explain Research Findings to Non-Researchers
- Common Mistakes
- Frequently Asked Questions
- What does it mean to communicate research findings?
- What is a non-technical audience?
- How do I summarize research findings in plain language?
- How do I explain a p-value to someone without a science degree?
- How do I know if the audience understood my explanation?
- What do I do when someone says my research is pseudoscience?
- Conclusion
What You Need
You need six things before you open a blank document, and none of them is a slide template. Most people arrive with the fifth one, the finding, and scramble for the rest an hour before the meeting.
- The audience’s actual question. Not your research question. Someone in the room is thinking “does this affect my team’s budget?” Write that sentence at the top of the page.
- The decision this finding feeds into. Approve, delay, redesign, fund, stop. If you cannot name one, your audience will not know what to do with what you say.
- One finding worth the meeting. Three findings is zero findings. Pick the one that would change something if it is true.
- Two numbers, not forty. The effect size and the comparison. Everything else lives in an appendix you offer, not a slide you push.
- The limits, in one sentence. What this study cannot show, drafted while the data are still fresh.
- One anchor. A single chart, a screenshot, or a concrete scenario from a real participant.
A non-technical audience is any group that does not use your field’s vocabulary to reason about evidence. That covers executives deciding on budget, policymakers weighing a rule, a journalist on deadline, a client who commissioned the work, and your own family at dinner.
They are not one audience, and treating them as one is the most common failure I see. A typical case is a research team that reports weekly to a room that is roughly 80% non-technical. The fix is not more detail, it is a fixed shape per audience.
| Audience | What they came for | Format that works | Length |
|---|---|---|---|
| Executives | A decision they owe someone else | One-pager or three slides, finding first | One page |
| Policymakers | Whether to act, and what it costs | Briefing note with an option list | Two pages |
| Journalists | A quotable line and a usable image | Press release plus one chart | Under 400 words |
| Clients or stakeholders | Proof the work answered their question | Readout deck with the method in an appendix | Ten slides |
| Cross-functional team | What to build or change next | Verbal walkthrough plus annotated examples | 20 minutes |
| Family and friends | Curiosity, and reassurance you are not being odd | Conversation with one analogy | Three minutes |
Step-by-Step: How to Explain Research Findings to Non-Researchers
Start With the Decision, Not the Method
Open with the decision your audience needs to make, not with how you studied the problem. Method-first openings train people to tune out within thirty seconds, because the method only matters once they care about the answer.
A useful structure here is And, But, Therefore: name what everyone already agrees, name the thing that complicates it, then name what follows. It forces your framing into three short sentences and keeps you from burying the news.
Weak: “We ran a mixed-methods study across four markets with 312 participants.”
Stronger: “Most shoppers say they compare prices before buying. But the ones who read a review in the first week returned far more often than the ones who didn’t. So we recommend we move reviews above the fold on the product page and test it this quarter.”
Translate the Finding Into Plain Language
Say the finding as one plain sentence before you say anything technical. If a non-researcher cannot repeat your sentence back to you, it is not finished yet.
The rule is simple: replace a term or define it in the same breath, never silently skip it. Then convert the grammar. Research writing hides the actor in passive voice and hides the claim in abstract nouns; “A significant difference in conversion was observed” becomes “people who saw the review converted better.”
| Research term | Plain-language replacement | How to say it in one sentence |
|---|---|---|
| Statistically significant | Unlikely to be chance | The gap we found is bigger than random noise would usually produce. |
| Effect size | How big the difference actually is | The difference was small enough that we would not change the product for it. |
| Confidence interval | The range of plausible answers | The real number is somewhere between these two, and we cannot pin it down yet. |
| Sample size | How many people | We asked 312 people, which is enough for a rough read but not a last word. |
| Margin of error | The slack around a number | When we say 40 percent, think roughly 36 to 44. |
| Correlation | Two things moved together | They rose at the same time, which does not mean one caused the other. |
| Control group | The comparison group | This is the group that got nothing new, so we can see what changed on its own. |
| Qualitative | What people said in their own words | Rather than a number, we have quotes explaining why they did it. |
Then rewrite the same finding three times and watch how much the meaning shifts. This is the fastest quality check I know, and it exposes over-claiming immediately.
For an executive: “Customers who read a review in week one came back 40 percent more often. Moving reviews higher on the product page is a low-cost test we can run this quarter.”
For a policymaker: “In our sample, people who saw clear labelling at the point of sale were more likely to report the product matched its description. That supports clearer labelling rules, though our sample cannot show whether the effect would hold at national scale.”
For a friend: “I found out that people who actually read the reviews for a product are the ones who keep buying it. So the reviews are doing more work than the advertising.”
The executive version drops the caveat to stay short, which is fine on a one-pager with an appendix attached. The policymaker version keeps it because the decision is larger and the cost of being wrong is higher.
Add Context That Makes the Result Meaningful
A number means nothing until your audience knows what it was compared to. Always give the baseline: last year, the control group, the competitor, or zero.
Three things non-researchers consistently need, and researchers consistently skip:
- The comparison group. “Conversion rose 12 percent” is incomplete. “Conversion rose from 2 percent to 2.4 percent” is a decision.
- The time period. A week and a year are different claims. Say which one.
- The magnitude, not the significance. Significance only tells you the result is probably not chance. Effect size tells you whether anyone should care. “Significant but tiny” is a real and common finding, and reporting it without the size is misleading.
When the claim is about cause, say it out loud: “these two things happened at the same time, and we cannot tell from this design which one drove the other.” That sentence costs you nothing in credibility and prevents the most expensive misunderstanding in the room.
Use an Example or Visual to Anchor the Finding
One clear picture or scenario does more work than three slides of tables. The goal is not decoration; it is giving people something concrete to hold while they process the abstract claim.
For a simple comparison, one horizontal bar chart with the baseline labelled beats almost anything else. For effect size, a single dot on a line with the zero point marked is harder to misread than a bar. Avoid dual axes, 3D anything, pie charts with more than three slices, and truncated y-axes, which manufacture dramatic-looking differences out of trivial ones.
The example has to come from real material. A quote from a participant, anonymised and unedited, teaches more than any chart, because it carries the reason behind the number. The one rule: never smooth a quote into a cleaner sentence than the person actually said.
State the Limits and Confidence Clearly
Say what the study cannot show, in one plain sentence, before someone else does it for you. Volunteering the limit does not weaken a finding. Getting caught by the limit you hid does.
Two templates work. The first separates what you can say from what you cannot: “We can say the label changed how people described the product. We cannot yet say whether it changed how often they buy it.” The second sets confidence as a level: “This is a strong read on behaviour and a weak read on motive, because the interviews were short.”
The failure mode is hedging until the finding sounds worthless. Fifteen “may suggest potentially” clauses do not express rigour. One precise sentence about the limit does more, and it lets the rest of your claim land with confidence.
Close With the Implication and Next Step
End with what changes, who does it, and by when. A finding without an implication is trivia, and a finding without an owner will be politely forgotten.
Try a two-part close: the implication in one sentence, then a named request. “If we move reviews above the fold we expect to see more week-two returns. Can we get 15 minutes next Thursday to design the test with marketing?”
Give people the vocabulary they need to repeat you later, too. If they leave with your exact phrase, your finding travels through the building after you have gone home.
Check Whether the Explanation Works
If you never check, you are guessing. Two minutes of verification saves a quarter of misunderstanding.
The best test is teach-back. Ask one person to say the finding back to you in their own words, then listen to what they leave out. The gap between your sentence and theirs is the jargon you did not remove. You will also hear the wrong conclusion they drew, which is more useful than any reaction you could have asked for.
Three questions cover most of it:
- What is the main thing we learned?
- What is the main thing we still do not know?
- What are we doing about it?
If someone answers the second one with your limitation and the third with your next step, the explanation worked. If they answer the first with your sample size, you led with the wrong thing.
A useful deadline test too: if the finding will be dead in six months, does the explanation depend on language that only your field will still recognise then?
Common Mistakes
Most communication problems are not writing problems. They are the same six errors, repeated, and each one has a specific fix.
| Common mistake | What the audience hears | The fix |
|---|---|---|
| Leading with method | This will take forty minutes and change nothing | Method goes in an appendix; the decision goes first |
| Presenting correlation as cause | They proved A causes B | Say the design, then say what it rules in and out |
| Hiding the uncertainty | Why does this feel like a sales pitch | One plain sentence on limits, said first |
| Unexplained charts | I cannot read the axis, so I stop looking | Title the chart with the finding, label the baseline |
| Too many findings | None of these stick | One finding per meeting, others offered not pushed |
| No ask | Interesting, nothing to do with it | Close with the implication and a named request |
Skeptical or dismissive audiences need their own handling. When someone calls the work pseudoscience simply because it is unfamiliar, arguing about rigour loses, because you are now defending your credentials instead of their question.
Three moves work better. Name the shared standard they already accept, such as published methods, replication, or a funder’s review, and put your work inside it. Then ask what would convince them, and answer honestly, including when the answer is that this study would not.
People who get teased for their field often carry the defensive version of this problem in their daily lives. Acknowledging how little of the process outsiders can see does more for trust than another citation.
One more line is worth keeping in your back pocket for public and media settings: the simplification line. Say that you have removed detail, not certainty, and name the one thing you removed that a critic would care about. It draws the criticism to the place you chose rather than the place you did not.
Frequently Asked Questions
What does it mean to communicate research findings?
It means translating what a study found, how confident you can be in it, and why it matters into language your specific audience can act on. The goal is plain language without jargon, and without stripping away the uncertainty that makes the finding trustworthy. In practice it means writing one take-home sentence, giving every number a comparison point, and naming the limit of the claim.
What is a non-technical audience?
A non-technical audience is any group that does not use your field’s vocabulary to reason about evidence. That includes executives making funding decisions, policymakers weighing a rule, journalists on a deadline, clients who commissioned the work, and your own family. The label tells you nothing useful about how much detail to give, so segment by the decision each group needs to make instead.
How do I summarize research findings in plain language?
Write the decision first, then the finding in one active sentence with a subject and a verb. Give the number a baseline and a time period, replace or define any technical term, and separate what you observed from what you think it means. Finish with the limit of the claim and the next step. Five sentences is a realistic target.
How do I explain a p-value to someone without a science degree?
Say it as: if there were no real effect here, how often would we expect to see a difference this big by luck alone. A low number means the result is hard to explain by chance. Then add the part people actually need, which is the size of the difference, because a result can be reliable and still too small to act on.
How do I know if the audience understood my explanation?
Ask one person to restate the finding in their own words and listen to what they drop and what they add. Ask three questions: what is the main thing we learned, what are we still unsure about, and what are we doing next. If the first answer is your sample size, you opened in the wrong place, and if the third is missing, close with a named request.
What do I do when someone says my research is pseudoscience?
Do not argue rigour, because that turns the conversation into a debate about your credentials. Point to a standard the person already accepts, such as published methods or independent review, and show where your work sits inside it. Then ask what evidence would change their mind and answer honestly. If nothing would, say so and stop.
Conclusion
How to explain research findings to non-researchers is a five-sentence discipline, not a writing talent. Write the decision, the finding in plain words, the number with its baseline, the one limit, and the next step with a name against it.
Start there today, before the meeting: put those five sentences in a document and read them out loud to a colleague who does not work in your field. If they cannot repeat the finding and the ask back to you, the rest of the deck will not rescue it.