Free scoping read. A written assessment of your manuscript in two business days, at no cost.

Start one
Contents

Academy · Research Writing · Data Analysis in Manuscript Writing

Interpreting Results: The Role of Data Analysis in Manuscript Writing

The SAMPL guidelines exist because, according to their authors, most statistical errors reported in published manuscripts involve basic rather than advanced methods; the manner in which a results section is written is therefore as important as the choice of statistical test.

Pubrica AcademyWritten by a scientific writing consultant Reviewed bya biostatistics reviewer Published10 October 2026 Last checked10 October 2026 Reading time8 minutes Sources12 peer-reviewed

Data analysis and manuscript writing are often treated as two separate stages, run the statistics, then write them up, but how results are interpreted and reported is itself a writing task with its own documented failure modes. The Statistical Analyses and Methods in the Published Literature (SAMPL) guidelines, developed by Thomas Lang and Douglas Altman specifically to address this, were created on the observation that most statistical reporting errors in published biomedical research involve basic, common methods rather than exotic ones, and that most journals say surprisingly little in their author instructions about how results should actually be reported.

Learning objectives

  • Why the SAMPL guidelines were created, and what gap in standard journal instructions they were meant to fill
  • The documented difference between reporting a standard deviation and a standard error, and why conflating them misleads readers
  • What the American Statistical Association's own guidance says about correctly interpreting a p-value in a results section

Common misconceptions

  • “Provided that the statistics were run correctly, the way they are described in the text is a minor stylistic choice.” The SAMPL guidelines were developed specifically because, in Lang and Altman’s own account, most statistical reporting errors found in the published biomedical literature involve basic methods; these errors are therefore not failures of the underlying analysis but failures in how that analysis is subsequently described, labeled and contextualized in the manuscript text.

Purpose of the SAMPL guidelines and the gap they address

Thomas Lang and Douglas Altman developed the SAMPL guidelines, published through the European Association of Science Editors, to give authors, editors and reviewers a shared reference for reporting basic statistical methods and their results, after observing that most of the statistical errors found in the published biomedical literature involve basic methods rather than advanced or unusual ones1. A central part of their rationale was that most journals’ Instructions for Authors say very little about statistical reporting, leaving authors to infer conventions rather than follow an explicit standard1. Unlike many reporting guidelines, SAMPL was not produced through a formal multi-stakeholder consensus process; instead, it draws on existing published guidance and on the authors’ observation that there is already near-universal agreement among statisticians on how the most common methods should be reported, so that the guidelines largely consolidate settled practice rather than propose new standards1. The guidelines are explicitly intended to complement, not replace, study-design-specific reporting standards such as CONSORT for randomized trials and STROBE for observational studies, which already contain statistics-related items but do not address basic reporting in the same detail23. This complementary relationship has practical consequences for the preparation of a results section. A study-design-specific checklist such as CONSORT confirms that the appropriate categories of information were reported, whereas SAMPL addresses whether the specific numbers within each category (a mean, a dispersion statistic or a test result) were reported with sufficient precision and correct labeling for a reader to interpret them. Authors are best served by consulting both kinds of guidance together rather than treating either as sufficient alone: compliance with a structural checklist such as CONSORT does not guarantee that the individual statistics within each section are labeled or interpreted correctly, and adherence to SAMPL’s basic reporting conventions does not by itself guarantee that the full set of required outcomes and analyses was reported. Consulting the two together before drafting the results section, rather than only at the final proofreading stage, is the practice that most directly reflects the recommendations of this body of guidance.

Confusion between standard deviation and standard error

One of the most frequently documented errors in results reporting is the confusion of the standard deviation, which describes the spread of individual data points, with the standard error, which describes the precision of an estimated mean. Douglas Altman and J. Martin Bland addressed this confusion directly in a short, widely cited explainer, noting that the two serve entirely different interpretive purposes: the standard deviation describes variability in the sample itself, whereas the standard error describes the uncertainty in an estimate of the population mean, and reporting one when the reader needs the other materially changes what a results section communicates4. Douglas Curran-Everett made a related pedagogical case in a dedicated teaching article for physiology researchers, demonstrating how a standard error decreases as sample size increases even when the underlying biological variability, captured by the standard deviation, remains constant; the distinction is simple to state but is routinely blurred in published results sections5.

Interpretation of p-values according to the American Statistical Association statement

The American Statistical Association issued a formal statement on p-values in 2016, directly addressing widespread misinterpretation in the reporting of results and stating explicitly that a p-value does not measure the probability that the studied hypothesis is true, nor the probability that the observed data were produced by random chance alone6. A companion methodological guide by Sander Greenland and colleagues catalogued the most common misinterpretations found in published results sections, including treating a non-significant result as proof of “no effect” and treating a statistically significant result as necessarily clinically or practically important, both of which the authors found recurring across the literature they reviewed7. A separate practical comparison of confidence intervals and p-values by du Prel, Hommel, Röhrig, and Blettner argued that reporting a confidence interval alongside, or instead of, a bare p-value provides more interpretive information, because the interval communicates both the estimated effect size and its precision in a single range rather than reducing that information to a binary significant or non-significant judgment8.

Effect size as an essential reporting element

A result can be statistically significant while being too small to matter in practice, or non-significant while reflecting a meaningful effect that a small sample lacked the power to detect, and effect size is the quantity that distinguishes between these possibilities in a way that a p-value alone cannot. Gail Sullivan and Richard Feinn, writing to address the overreliance on p-values in published medical education research, argued that effect size answers the more clinically relevant question, “how much,” whereas a p-value answers only “whether,” and that a results section reporting one without the other leaves the reader unable to judge practical importance11. Jacob Cohen’s foundational work on statistical power analysis, which remains the reference point for standard effect-size conventions such as Cohen’s d for mean differences, established the benchmarks (small, medium and large) that allow readers to calibrate what a reported value represents in practical terms, rather than evaluating it only against an arbitrary significance threshold12. Reporting an effect size alongside its confidence interval, consistent with the emphasis of the SAMPL guidelines, gives a reader both the estimated magnitude of an effect and an indication of the precision with which that magnitude was measured, a combination of information that a bare significance statement cannot provide.

Structured reporting guidelines and the link between results and methods

The CONSORT statement for randomized trials and the STROBE statement for observational studies both require a close, verifiable correspondence between what a manuscript’s methods section states was planned and what its results section reports, specifically to prevent selective or inconsistent reporting of statistical outcomes23. Geoff Cumming’s broader argument for what he termed “the new statistics” draws these threads together, proposing that manuscripts should foreground effect sizes, confidence intervals and meta-analytic thinking throughout the results narrative rather than treating a single significance threshold as the principal finding; this shift changes how a results section is written, not only which numbers are calculated9. The implication for manuscript writing is that the correct interpretation of results in prose, stating what an effect size and its interval mean rather than only whether a threshold was crossed, is where many of the most commonly documented reporting errors in the literature occur10. None of this requires different statistical analyses; it requires the same statistical output to be described more completely and more carefully in prose. A manuscript’s underlying analysis can be entirely correct while its results section still misrepresents the findings, through imprecise language, an unlabeled dispersion statistic, or a p-value discussed as though it resolved a question it could not answer, which is precisely the gap that the guidelines and standards reviewed here were designed to close.

Key points

Key points

Most documented statistical reporting errors in published manuscripts involve basic methods described incorrectly in the text rather than flawed underlying analysis; careful, standards-aligned writing of the results section is therefore a distinct and necessary skill, separate from running the statistics themselves.

Practical application

Practical application

  1. Whether a reported value is a standard deviation or a standard error should be stated explicitly. The abbreviation alone should not be relied upon to convey this distinction; it should be labeled in the text and in table and figure captions, since the two convey fundamentally different information to a reader.
  2. Confidence intervals should be reported alongside, or instead of, bare significance statements. An interval communicates effect size and precision together, which a significant/non-significant label alone does not.
  3. The results section should be checked against the relevant structured reporting standard before submission. CONSORT for trials or STROBE for observational studies should be used as a line-by-line check that the results deliver what the methods section specified.

Frequently asked questions

Lang and Altman, the authors of the SAMPL guidelines, attribute this partly to the limited specific guidance on basic statistical reporting in most journals’ author instructions, which leaves authors to infer conventions for even the most common methods.

Current guidance, including the ASA’s statement, recommends against reliance on a p-value alone, since it does not by itself communicate the size or precision of an effect, both of which are necessary to interpret a result’s practical importance.

No. Greenland and colleagues specifically identify this as one of the most common misinterpretations in published results sections; a non-significant result can reflect insufficient statistical power rather than a true absence of effect.

Many major journals require or strongly recommend reporting guidelines such as CONSORT and STROBE for the relevant study designs, whereas SAMPL functions as a reference resource on which authors and editors can draw for basic statistical reporting, since it was not developed as a formal consensus requirement in the same way.

References

  1. Lang TA, Altman DG. Basic statistical reporting for articles published in biomedical journals: the SAMPL guidelines. In: Smart P, Maisonneuve H, Polderman A, eds. Science Editors' Handbook. European Association of Science Editors; 2013. read 10 October 2026. View on EQUATOR Network
  2. Moher D, Hopewell S, Schulz KF, et al. CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. BMJ. 2010;340:c869. read 10 October 2026. View on BMJ
  3. Vandenbroucke JP, von Elm E, Altman DG, et al. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLOS Medicine. 2007;4(10):e297. read 10 October 2026. View on PLOS Medicine
  4. Altman DG, Bland JM. Standard deviations and standard errors. BMJ. 2005;331(7521):903. read 10 October 2026. View on BMJ
  5. Curran-Everett D. Explorations in statistics: standard deviations and standard errors. Advances in Physiology Education. 2008;32(3):203-208. read 10 October 2026. View on Advances in Physiology Education
  6. Wasserstein RL, Lazar NA. The ASA Statement on p-Values: Context, Process, and Purpose. The American Statistician. 2016;70(2):129-133. read 10 October 2026. View on The American Statistician
  7. Greenland S, Senn SJ, Rothman KJ, et al. Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations. European Journal of Epidemiology. 2016;31(4):337-350. read 10 October 2026. View on European Journal of Epidemiology
  8. du Prel JB, Hommel G, Röhrig B, Blettner M. Confidence interval or p-value? Part 4 of a series on evaluation of scientific publications. Deutsches Ärzteblatt International. 2009;106(19):335-339. read 10 October 2026. View on Deutsches Ärzteblatt International
  9. Cumming G. The new statistics: why and how. Psychological Science. 2014;25(1):7-29. read 10 October 2026. View on Psychological Science
  10. Moher D, Schulz KF, Altman DG, for the CONSORT Group. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomised trials. Lancet. 2001;357(9263):1191-1194. read 10 October 2026. View on The Lancet
  11. Sullivan GM, Feinn R. Using effect size—or why the P value is not enough. Journal of Graduate Medical Education. 2012;4(3):279-282. read 10 October 2026. View on Journal of Graduate Medical Education
  12. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Lawrence Erlbaum Associates; 1988. read 10 October 2026. View publisher record

Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.

How to cite this article

Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.

Pubrica Academy. (2026, October 10). Interpreting Results: The Role of Data Analysis in Manuscript Writing. Pubrica. https://pubrica.com/academy/concepts-definitions/role-of-data-analysis-in-manuscript-writing/

This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.

Results Section Statistical Reporting Checklist

This checklist should be applied to the results section before submission.

[ ] Labeled every reported dispersion statistic clearly as SD or SE, never ambiguous
[ ] Reported confidence intervals alongside or instead of bare p-values
[ ] Avoided describing a non-significant result as proof of 'no effect'
[ ] Avoided describing a significant result as automatically 'important' without discussing effect size
[ ] Checked the results section against CONSORT or STROBE items relevant to your study design
[ ] Confirmed every number in the results section matches what the methods section said would be analyzed

Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.

How Pubrica can help

Pubrica’s manuscript writing team reviews results sections against SAMPL, CONSORT and STROBE reporting standards, correcting common errors in the reporting of standard deviation (SD) and standard error (SE) and in p-value interpretation before submission.

Find something else

Search the Pubrica Academy

Access is free and requires no account. Enter a topic, or open the full library to filter by category, sub-topic or resource type.

Correct statistical reporting in the results section

A scoping assessment is free and takes two business days. It provides a written assessment of what the work requires and which specialist would undertake it.

Get a free quote

Regulations

Check your project against this rule

  1. 1Your work
  2. 2Your details
  3. 3Send
What do you need?

Material you send is held under signed confidentiality, with access limited to the assigned team and encrypted transfer. It is not used to train anything and is not passed to third-party tools. Ask for the agreement before you send anything you would rather cover first.