Follow the order set in your Methods
ICMJE advises presenting results in logical sequence in the text, tables and figures, giving the main or most important findings first, and providing data on all primary and secondary outcomes identified in the Methods. For observational studies, STROBE items 13–17 set out what to report, from participant flow to other analyses.[1][2]
On small screens, scroll the table horizontally to compare every column.
| Order | What to report | Common problem |
|---|---|---|
| Participant flow (STROBE 13) | Numbers at each stage and reasons for exclusion; a flow diagram where useful | The difference between the starting number and the analysed number is unexplained. Item 13 explained → |
| Baseline characteristics (14) | Characteristics by group, missing data per variable and, for a cohort, follow-up time | Every table seems to share one denominator, but missing data per variable are not shown. |
| Outcome data (15) | Event numbers or summary measures suited to the design, with denominators | Only percentages, without denominators, event numbers or periods. |
| Main results (16) | Unadjusted estimates and, where relevant, adjusted estimates with precision and the variables adjusted for | Only P values, without estimates or confidence intervals. Item 16 explained → |
| Other analyses (17) | Subgroup, interaction and sensitivity analyses, and whether they were prespecified | An analysis described in the Methods has no result, or an exploratory finding reads like the conclusion. |
A worked revision: taking interpretation out of a Results paragraph
This is the same fictional teaching study as in the Discussion guide. A single-centre retrospective cohort included one index appointment for each of 200 adults. Attendance within 30 days was 82/100 in the reminder group and 70/100 in the comparison group. Reminder receipt was not randomized, and baseline differences were not adjusted for. This is not a real study, patient dataset or product-performance example.
- “Significantly” was removed. It is unclear whether it is a statistical term or everyday emphasis. Show statistical findings through estimates and their uncertainty.
- “12%” became “12 percentage points.” The difference between 82% and 70% is 12 percentage points. If you report a relative difference, define it (here 12/70, about 17%).
- “Effective” was moved out of the Results. It interprets an unrandomized, unadjusted comparison as an effect. Any interpretation belongs in the Discussion, with its limitations.
A main-result paragraph in English
Wording to improve
The reminder group had significantly better attendance than the comparison group (82% vs 70%, p < 0.05), an increase of 12%. This suggests that reminders are an effective way to improve attendance. Detailed results are shown in Table 2.Revision or template
Of [number] adults assessed for eligibility, [number] were excluded ([reasons]; Figure 1), and 200 were included: 100 who had received a reminder and 100 who had not. Baseline characteristics are shown in Table 1. Attendance within 30 days was 82/100 (82%) in the reminder group and 70/100 (70%) in the comparison group, an unadjusted difference of 12 percentage points ([95% CI from the analysis]; Table 2).The revision reports numerators and denominators, the difference in percentage points, participant flow and the locations of tables and figures, and moves the claim of effectiveness to the Discussion. The confidence interval is left as a placeholder for the value from your actual analysis; no P value, adjusted analysis or new number is invented. Fill in screening and exclusion numbers and reasons from your records.
Reporting numbers: give denominators, differences and uncertainty
ICMJE advises giving numeric results not only as derivatives such as percentages but also as the absolute numbers from which they were calculated; quantifying findings with indicators of uncertainty such as confidence intervals where possible; and not relying solely on hypothesis tests such as P values. It also advises avoiding nontechnical uses of statistical terms such as “random,” “normal,” “significant,” “correlations” and “sample.”[1]
- Keep decimals and units consistent. Use the same precision and units for a measure across the text, tables and abstract. If rounding stops percentages adding up to 100%, say so in a table footnote.
On small screens, scroll the table horizontally to compare every column.
| Check | Wording to improve | Revision |
|---|---|---|
| Numerator and denominator | 82% | 82/100 (82%) |
| Percentage points vs relative difference | 12% higher | 12 percentage points higher |
| Estimate and precision | p < 0.05 | [estimate] (95% CI [lower]–[upper]); the exact P value if reported |
| Everyday use of statistical terms | significantly better | Describe the difference with the estimate and confidence interval |
| Denominators that change by analysis | Mean distance was [value] km. | Among the [n] participants with complete data, … |
Divide the work between text, tables and figures
ICMJE advises against repeating all the data in tables or figures in the text; emphasize or summarize only the most important observations. Restrict tables and figures to those needed to explain the argument and assess supporting data; graphs can replace tables with many entries, but do not duplicate data in graphs and tables. Separate reporting by demographic variables such as age and sex should be routine unless there are compelling reasons not to stratify, which should be explained.[1]
- Text Summarize the main findings that answer the research question, with their numbers, and point readers to the table or figure with the details.
- Tables Let readers compare exact values across groups and variables. State denominators and units in column headings and show missing data per variable.
- Figures Use them when a pattern matters, such as change over time or a distribution. Do not repeat the same values in a table.
Leave interpretation for the Discussion
The Results report what was observed; the Discussion explains what it means. When a Results sentence starts to say “this may be because…,” “consistent with previous studies” or “effective,” move it to your Discussion draft.
- Keep in the Results Who was included and analysed; observed estimates, direction and uncertainty; which analysis produced each value; whether it was prespecified.
- Move to the Discussion Mechanisms and causes, comparison with previous studies, clinical meaning, and how limitations affect interpretation.
- Verbs to watch in observational studies Instead of causal verbs such as “improved,” “reduced” or “led to,” use wording that matches the analysis, such as “was higher” or “was associated with.”
A Results scaffold to adapt
Replace brackets with your actual analysis results and records. Report every outcome and analysis named in the Methods, even when the estimate is uncertain or unexpected. Do not fill in analyses you did not perform or numbers you have not verified.
Participants
Of [number] [patients or records] assessed for eligibility, [number] were excluded ([reasons]; Figure 1). [Number] were included in the analysis: [number] in [group A] and [number] in [group B].
Baseline characteristics
Table 1 shows [characteristics] by group. [Variable] was missing for [number] participants.
[For a cohort: follow-up was [summary measure] ([measure of spread]).]
Main outcome
[Outcome] occurred in [n/N] ([%]) of [group A] and [n/N] ([%]) of [group B]. The [unadjusted/adjusted] [effect measure] was [estimate] (95% CI [lower]–[upper]) (Table 2).
[If adjusted: adjusted for [variables actually included]; the reasons are given in Methods.]
Secondary outcomes and other analyses
[Report every secondary outcome named in Methods, including uncertain or null estimates.]
[Label subgroup and sensitivity analyses as prespecified or exploratory, as recorded.]Review the Results alongside the tables and abstract
Prepare a Word or PDF manuscript containing Methods, Results and tables, then select “Start my manuscript review.” Sign in, open the manuscript and run Health Check. In the Results, look first at whether numbers and denominators agree across text, tables and abstract, whether every outcome in the Methods is reported, and whether interpretation has crept into result sentences.
Request to adapt: “Separate numbers or denominators in the Results that differ from the tables or abstract, outcomes named in the Methods but not reported, sentences with a P value but no estimate, and sentences that interpret. Identify the supporting table or text locations. Do not invent numbers or analyses. List anything the manuscript cannot establish as a question for the authors.”
- Before revising Check flagged numbers against the original analysis output. Matching numbers only to each other inside the manuscript can leave a wrong value in place.
- After revising Reconcile numbers and denominators across the abstract, Results, tables and figure legends. AI review does not replace raw-data reanalysis or judge whether the statistical methods were appropriate.
Frequently asked questions
Should I avoid P values in the Results?
No. ICMJE advises not relying solely on hypothesis tests such as P values and, where possible, quantifying findings with indicators of uncertainty such as confidence intervals. If you report P values, use the values from your actual analysis and follow the journal’s style.
How do I report a nonsignificant result?
If the outcome was defined in the Methods, report it with its estimate and confidence interval rather than leaving it out. Do not conclude “no difference” from a large P value alone; explain what the result means, with its limitations, in the Discussion.
Can I repeat numbers from a table in the text?
Yes, for the key numbers that answer the research question. ICMJE advises against repeating all the data from tables or figures in the text; emphasize or summarize only the most important observations.
Do trials and systematic reviews follow the same order?
The logic—flow, characteristics, main results, other analyses—is similar, but randomized trials follow CONSORT and systematic reviews follow PRISMA items and flow diagrams. See the writing-order guide for where each design starts.
Sources and how we checked them
ICMJE guidance on Results and statistics and the original STROBE checklist were checked on November 18, 2026. The order table, examples and checklists are educational suggestions created by Paper Buddy, not official forms.