Lay summary examples, before and after

Two complete examples of a plain-language summary written from a researcher's own abstract: one for a journal's 150-word box and one for a 200-word public-outcomes report. Each shows the abstract, the summary, the technical terms that were replaced, and the list of claims the author must still check.

How these were made, honestly: both abstracts are invented for this page (no real study, no real data). The summaries were written by Claude, an AI model made by Anthropic, using the brief behind our Lay Summary Draft service, in one run each, and are shown exactly as returned. Your own result will differ, and an AI draft must be read against your paper before you submit anything.

Example 1: journal abstract to a 150-word summary

Where it was going: Journal plain-language summary, max 150 words.

The abstract the author supplied (invented for this page)

Urban green space is thought to support ground-dwelling invertebrates, but the effect of small, isolated lawns and verges is rarely quantified. We sampled ground beetles (Carabidae) with pitfall traps at 24 sites in a mid-sized city between May and August 2024: 12 mown lawns and 12 verges left unmown for the season. Unmown verges held more beetle species per trap than mown lawns (mean 9.1 versus 5.4; mixed-effects model, p < 0.01), and total beetle abundance did not differ between treatments. Community composition differed between treatments, with several open-habitat specialists recorded only on unmown verges. Site-level soil moisture explained part of the variation in species richness, but we could not separate it from mowing regime because treatments were not randomised across soil types. Results suggest that reduced mowing may increase local beetle diversity, but our single-season, single-city design limits generalisation. Longer monitoring across more cities is needed to test whether these gains persist.

The plain-language summary the service wrote (147 words)

City green spaces are thought to support small ground-living animals. But researchers rarely measure how much small, isolated lawns and roadside grass strips help. We caught ground beetles in buried cup traps at 24 city sites from May to August 2024. The sites were 12 mown lawns and 12 roadside strips left unmown all season. On average, unmown strips had more beetle species per trap than mown lawns: 9.1 versus 5.4. Total beetle numbers did not differ. But the mix of species differed, and several species of open habitats appeared only on unmown strips. Soil moisture explained some differences in species numbers. We could not separate this from mowing, as mown and unmown sites were not spread randomly across soil types. The results suggest less mowing may increase local beetle variety. This single-season, single-city study limits how far results apply. Longer monitoring in more cities is needed.

What it told the author about the changes

Your abstract clearly gives the design, the main result with figures, the soil-moisture problem and the limits of one season in one city. To stay under 150 words I dropped the statistical detail (mixed-effects model, p < 0.01), the word "mid-sized" for the city, and the aim of testing "whether these gains persist." The biggest gap for a lay reader is that the abstract does not say why beetle variety matters to people or who could use these findings.

Terms it replaced

• "ground-dwelling invertebrates" -> small ground-living animals
• "pitfall traps" -> buried cup traps
• "verges" -> roadside grass strips
• "species richness" -> species numbers
• "Community composition" -> the mix of species
• "open-habitat specialists" -> species of open habitats
• "total beetle abundance" -> total beetle numbers
• "mixed-effects model, p < 0.01" -> left out (statistical detail)

Claims it asked the author to double-check

• I described "verges" as roadside grass strips; confirm all 12 verges were beside roads.
• "Species numbers" stands in for species richness per trap; confirm readers won't take it as the total species count per site.
• "Total beetle numbers did not differ" leaves out that this was a statistical test result; confirm the plain wording is fair.
• "Soil moisture explained some differences" simplifies "site-level soil moisture explained part of the variation"; confirm that "some" isn't too strong or too weak.
• "May increase local beetle variety" keeps your hedges "suggest" and "may"; confirm "variety" is an acceptable plain word for diversity here.
• The p-value and the aim of testing whether gains persist were cut for length; confirm the journal doesn't require them.

Gaps it left for the author to fill

• None in the summary itself.
• Why this matters: the abstract does not say why beetle variety is important (for example, its role in city ecosystems). Consider adding one line if your journal allows.
• Who could use this: the abstract does not name who might act on the findings, such as people who manage city grass. Add this only if your paper supports it.

Example 2: a randomised trial to a 200-word public-outcomes summary

Where it was going: NSF Public Outcomes lay summary, 200 words, general public.

The abstract the author supplied (invented for this page)

Students from groups that are underrepresented in science often earn lower grades in gateway courses, and brief psychological interventions have been proposed to narrow these gaps. We tested a values-affirmation writing exercise in a randomised, double-blind field experiment across 6 introductory biology courses at 3 public universities. In total, 1,214 first-year students were randomly assigned to write for 15 minutes about a value that mattered to them (n = 607) or about a neutral topic (n = 607), in the second week of the semester and again in week 8. The primary outcome was final course grade on a 4-point scale. Students in the affirmation condition earned a mean final grade of 2.94, compared with 2.81 in the control condition (difference 0.13, 95% CI 0.04 to 0.22). The benefit was larger among first-generation students (0.21) than among continuing-generation students (0.06). Effects on course withdrawal were small and not statistically distinguishable from zero. Because students knew they were taking part in a writing study, and because we did not measure sense of belonging directly, we cannot say through which mechanism the exercise worked. Our findings suggest that a short, low-cost exercise may modestly improve grades for some students, but they do not show that the effect would hold at institutions unlike these three.

The plain-language summary the service wrote (196 words)

Students from groups that are underrepresented in science often earn lower grades in early gateway courses. Short psychological exercises have been proposed to narrow these gaps. We tested one: a short writing task about a personal value. The study ran in 6 introductory biology courses at 3 public universities. We randomly assigned 1,214 first-year students to one of two groups. In one group, 607 students wrote for 15 minutes about such a value. The other 607 wrote about a neutral topic. Students wrote in the second week of the semester and again in week 8. Neither students nor researchers knew who was in which group. The values group averaged a final grade of 2.94 on a 4-point scale, versus 2.81. The gain was larger for first-generation students, the first in their family to attend college. Effects on dropping the course were small and unclear. Students knew they were in a writing study, and we did not measure sense of belonging directly. So we cannot say how the exercise worked. Our findings suggest this short, low-cost exercise may modestly improve grades for some students. They do not show the effect would hold at institutions unlike these three.

What it told the author about the changes

Your abstract clearly covers the problem, design, sample, main result and limits, so nearly all of it fits the 200-word limit. To stay under it, I left out the exact grade difference (0.13), its confidence interval, and the subgroup figures (0.21 and 0.06). The biggest gap for an NSF Public Outcomes report is who benefits beyond these courses, which your abstract does not say.

Terms it replaced

• "brief psychological interventions" -> short psychological exercises
• "values-affirmation writing exercise" -> a short writing task about a personal value
• "randomised, double-blind field experiment" -> we randomly assigned students to groups, and neither students nor researchers knew who was in which group
• "affirmation condition" -> the values group
• "course withdrawal" -> dropping the course
• "not statistically distinguishable from zero" -> small and unclear

Claims it asked the author to double-check

• "Neither students nor researchers knew who was in which group" is my plain reading of "double-blind." Confirm this matches who was actually blinded in your design.
• "The gain was larger for first-generation students" drops the comparison figures. Check that this wording does not overstate the subgroup difference.
• "Effects on dropping the course were small and unclear" stands in for "not statistically distinguishable from zero." Confirm that this reading is acceptable.
• "Early gateway courses" adds the word "early" to explain "gateway." Check that this fits your use of the term.
• The hedges "suggest" and "may modestly improve grades for some students" are kept exactly. Do not strengthen them in later edits.

Gaps it left for the author to fill

• There are no bracketed blanks in the summary.
• NSF Public Outcomes reports usually say who benefits and why the work matters to the public. Your abstract says only "some students." Consider adding a sentence on wider value, such as for instructors or universities, if your results support it.
• Consider whether the 0.13-point difference or the subgroup figures (0.21 vs 0.06) should replace a less central sentence.

What to copy from these examples

• Every figure survives, or is left out on purpose and said so. The first summary keeps 24 sites, 12 and 12, and 9.1 versus 5.4, and states that it dropped the p-value. A summary that quietly rounds or adds a number is the first thing a careful reviewer notices.
• The hedges stay. "Suggest" and "may" in the abstract are still "suggest" and "may" in the summary. Plain language is not permission to claim more.
• Jargon becomes a short phrase, once. "Pitfall traps" becomes "buried cup traps" and "species richness" becomes "species numbers".
• Missing impact is flagged, not invented. Neither abstract says who benefits beyond the study, so neither summary makes that up. The author is told to add it only if the results support it.

Check your own

Paste your draft into the free lay summary checker to see its length against your limit, reading grade, long sentences, jargon and unexplained acronyms, in your browser. For seven rules with a before-and-after sentence example, read how to write a lay summary. If you want a first draft from your own abstract, the Lay Summary Draft service costs $7 and is AI-written.