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

Start one
Home·Academy·Research impact·Abstract writing formats

Research Abstract Formats Explained: A Guide for Academic Writing

Most descriptions of abstract formats are lists of headings. The more useful question is how well published abstracts represent the papers beneath them, and the answer is measured. An audit of six general medical journals found between 18 and 68 per cent of abstracts deficient. Their data were inconsistent with the article body, or absent from it. This article sets out the formats, the checklists that govern them, and that evidence.

What you will learn
  • Which abstract formats exist, and which body actually determines the one you must use
  • The reporting checklists that apply to an abstract, and how they are selected
  • How often published abstracts disagree with their own papers, and in which direction
What to unlearn
  • “Abstract length follows a general standard of 150 to 300 words.” Limits are set by the individual journal or conference and vary widely. No general standard governs them.
  • “A conference abstract is a step towards a paper.” Across 307,028 abstracts, 37.3 per cent were subsequently published in full, and positive results were published more often than negative ones.

What this article covers

This article treats the abstract as a reported document rather than as a writing exercise, and two things follow from that. The first is that specific reporting checklists apply to it. They are selected by the design of the study rather than by the discipline. The second is that its fidelity to the paper has been audited, and those audits found substantial discrepancy.

The formats themselves are covered first, because the terminology is used loosely. The distinctions also matter less than is usually claimed. What determines the format in practice is the journal or conference, which publishes its requirement. Discipline-level generalisations about which format is used where are weaker than they appear. None is asserted here without a source.

The remainder concerns evidence, and three studies are drawn on for it. The first is an audit of abstract accuracy across six journals. The second analyses how nonsignificant results are described in abstracts. The third is a Cochrane methodology review of what becomes of conference abstracts. Each reports figures that bear directly on how an abstract should be written.

Where this sits among the related articles

The structure of the paper the abstract summarises is covered in how to structure a research article. The checklist governing trial reports, including its abstract items, is set out in CONSORT 2025. The metadata entered alongside an abstract at submission is covered in the journal submission process.

Common Research Abstract Formats
Common Research Abstract Formats

The formats and their origins

Two formats are commonly distinguished, and the distinction between them is real but narrow. One carries explicit headings and the other does not. Everything else described as a separate format is a variation on those two. The variation is in length, in audience or in the headings chosen.

The structured abstract

A structured abstract divides the summary under labelled headings. The labels vary, and the common set follows the shape of the paper itself. Background, methods, results and conclusions is the usual minimum. Longer variants add objectives, design, setting, participants, interventions, main outcome measures and limitations. The purpose of the labels is to force each element to be present rather than to improve the prose.

For randomised trials the requirement is specific rather than stylistic. The current trial reporting statement carries items covering the title and abstract, and places them first in its checklist.1 Those items sit within a checklist of thirty items in six sections.1 An abstract for a trial is therefore not written to a journal's house preference alone. It is written against a published checklist that the journal may also require to be submitted.

The unstructured abstract

An unstructured abstract runs as continuous prose without headings. It is common in the humanities, in some physical sciences, and in journals that prefer a narrative summary. The absence of headings does not remove the obligation to cover the same elements. It removes only the scaffolding that makes an omission visible. Unstructured abstracts are therefore easier to write, and easier to write badly.

Length and the question of a standard

Abstract length is frequently presented as a general convention of around 150 to 300 words. No body publishes such a standard, and the figure describes a common range rather than a rule. The binding limit is the one set by the journal or conference, and it is published in the instructions for authors. Limits differ substantially, and a limit expressed as 250 words is not satisfied by 260.

The same applies to conference abstracts, which are often described as shorter than journal abstracts. That is frequently true and is not a standard. The submission system for the specific meeting states the limit, and it is usually enforced automatically. A general figure quoted without a source describes practice. It should not be treated as a requirement.

The format is set by the venue

The practical rule is simpler than the taxonomy suggests. The journal or conference determines the format, the headings and the length, and publishes all three. Discipline-level generalisations are a poor substitute for reading that page. A reporting checklist also applies for trials and systematic reviews. There the checklist determines the content, and the venue determines the container.

Figures: evidence on abstract accuracy

Figure 1 summarises how often abstracts in six general medical journals contained data that could not be verified in the article. Figure 2 shows how frequently the abstracts of trials with nonsignificant results contained spin. They illustrate why an abstract should be checked line by line against the main text before submission.

Figure 1 · Abstracts whose data could not be verified in the article

Six general medical journals, 44 original research articles sampled from each. Markers are point estimates and bars are 95 per cent confidence intervals.

Per cent of sampled abstracts found deficientJournal A18%Journal C30%Journal E32%Journal B43%Journal D45%Journal F68%020406080100Per cent of abstracts

What to look for. No journal was free of the problem. The lowest estimate still places roughly one abstract in six outside what the paper supports. The intervals are wide because each journal contributed 44 articles. The ordering between adjacent journals should not be read as meaningful. The spread between the extremes is the finding, and the authors report it as statistically significant.

Journals are labelled A to F as in the source table. The published article does not map those letters to the six named journals, so no mapping is attempted here. Positions are to scale on a linear axis.

Figure 2 · Spin in the abstracts of trials with nonsignificant results

72 trial reports appraised, all with a statistically nonsignificant primary outcome.

SPIN PRESENT — 72 abstracts appraisedIn at least one section68.1%In the conclusions58.3%In the results37.5%0255075100Per cent of abstractsTHE FORMS SPIN TOOKIN THE CONCLUSIONSOnly treatment effectiveness · 23.6%Claiming equivalence · 13.9%Other significant results · 8.3%Efficacy despite the primary · 5.6%IN THE RESULTSWithin-group comparisons · 11.1%Subgroup analyses · 8.3%Secondary outcomes · 4.2%

What to look for. The conclusions section carried spin more often than the results section. That is the reverse of where an author might expect scrutiny to fall. Every trial in this sample had a nonsignificant primary outcome. Each abstract therefore had the same unfavourable finding to report. The listed forms are the devices by which that finding was made to read more favourably.

Percentages for the individual forms are of all 72 abstracts and do not sum to the section totals. An abstract could carry more than one form, or none.

The checklists that govern abstracts

An abstract reporting a study of a recognised design is covered by a published checklist. The checklist is selected by the design, not by the journal and not by the discipline. This is the single most useful fact about abstract writing. It is absent from most general advice on the subject.

The library and the selection rule

Reporting guidelines are collected by the EQUATOR Network, whose library listed 709 entries at the time of writing.2 Separate guidelines cover randomised trials, systematic reviews, observational studies and case reports.2 Others cover qualitative studies, diagnostic accuracy, economic evaluations and study protocols.2 Several of these have extensions written specifically for abstracts, and the library records them alongside the main statements.2

Trials and the abstract items

The current trial reporting statement places title and abstract first among its thirty items.1 Those items are numbered 1a and 1b, and they open a checklist arranged in six sections.1 The statement supersedes its predecessor, which its authors say should no longer be used.1 An abstract prepared against the older version omits items the current checklist requires.1

Systematic reviews and the twelve-item checklist

Systematic reviews have an abstract checklist of their own, carrying twelve items and covering both journal and conference abstracts.3 The original version was published in 2013 as a standalone statement.4 The 2020 revision of the main systematic review guideline published an updated abstract checklist within it, and that update replaces the 2013 statement.3

A pattern worth noticing across guidelines

Two of the checklists described here have superseded earlier versions.13 In both cases the earlier version is explicitly withdrawn from use. An author may be working from a template, a supervisor's notes or a previously published paper. Each of those carries whichever version was current when it was made. Checking the date of the checklist costs very little. It is the most common avoidable error in this area.

The scope and limits of a checklist

A checklist names the items an abstract must contain, and nothing beyond that. It does not determine the headings, the word limit or the order, all of which remain with the venue. Nor does it make the abstract accurate, which is a separate property examined in the next section. The checklist guarantees only that an item was addressed somewhere, and an item can be addressed misleadingly.

The accuracy of published abstracts

An abstract is the part of a paper most people read and the part fewest people check. Whether it agrees with the paper beneath it has been audited directly. The answer is the most useful thing an author can know before drafting one.

The cross-journal audit

A study published in 1999 examined whether data in abstracts could be verified in the article body.5 Random samples of 44 original research articles were drawn from each of five journals.5 A consecutive sample of 44 was taken from a sixth.5 The articles were published between July 1996 and August 1997.5 An abstract was counted deficient if its data were inconsistent with the body or absent from it.5

The proportions found were substantial at every journal in the sample. The lowest was 18 per cent, with an interval from 6 to 30.5 The highest was 68 per cent, with an interval from 54 to 82.5 The four intermediate figures were 30, 32, 43 and 45 per cent.5 The difference among journals was statistically significant.5 The authors concluded that data in the abstract inconsistent with or absent from the body “are common”, and noted that this held even in large-circulation general medical journals.5

The two failure modes

The definition used covers two distinct errors, and they have different causes. Data inconsistent with the body usually arise from revision during review. A figure changes, the body is updated, and the abstract is not. Data absent from the body usually arise from compression. A number is calculated for the abstract and never appears in the paper. The reader is then unable to see how it was derived.

Both failure modes are avoidable by the same procedure. The abstract is written last, from the final tables. Every number in it is then located in the paper before submission. This is mechanical work rather than editorial judgement. That is why it is often skipped, and why skipping it is detectable.

The age of this evidence

The audit sampled articles published in 1996 and 1997, and reporting practice has changed since. Abstract reporting checklists did not then exist for trials or systematic reviews, and structured abstracts were less widely required. The figures establish that the problem is real and was once large. They should not be read as a current rate. No comparably designed audit across the same six journals has been identified for a recent period.

Spin and the conclusions section

Accuracy concerns whether the numbers in an abstract match the paper. A separate question is how those numbers are characterised, and that has also been measured. The relevant study examined trials whose primary outcome was not statistically significant. That is the circumstance in which the temptation to recharacterise is strongest.

The analysis and its definition

A 2010 study screened 616 published trial reports and appraised the 72 that were eligible.6 Every trial in that set had a statistically nonsignificant result for its primary outcome.6 The authors defined spin as “specific reporting strategies, whatever their motive, to highlight that the experimental treatment is beneficial”.6 The definition is deliberately silent on intent, which makes it applicable without attributing motive.

Spin appeared in at least one section of 49 abstracts, or 68.1 per cent.6 The interval ran from 56.0 to 78.6.6 The conclusions section carried it in 42 abstracts, or 58.3 per cent.6 The results section carried it in 27, or 37.5 per cent.6 The authors concluded that the reporting and interpretation of findings “was frequently inconsistent with the results”.6

The forms it took

The devices recorded in the conclusions section are specific and recognisable. Seventeen abstracts, or 23.6 per cent, focused only on treatment effectiveness.6 Ten, or 13.9 per cent, claimed equivalence on the basis of a nonsignificant result.6 Six emphasised other significant results, and four claimed efficacy without engaging the nonsignificant primary outcome.6

The results section used different devices again. Eight abstracts, or 11.1 per cent, foregrounded statistically significant within-group comparisons.6 Six turned to subgroup analyses, and three to secondary outcomes.6 Each of these substitutes a comparison that reached significance for the comparison the trial was designed to make.

The equivalence claim in particular

One form deserves separate attention, because it is a logical error rather than a matter of emphasis. A nonsignificant difference is not evidence of equivalence. It is compatible with equivalence and also with a difference the study was too small to detect. Establishing equivalence requires a trial designed for that purpose, with a prespecified margin. Describing a null result as showing the treatments comparable asserts something the data cannot support. It appeared in roughly one abstract in seven.6

Conference abstracts and full publication

A conference abstract is often treated as a preliminary version of a paper. The evidence does not support that description for most of them. A Cochrane methodology review examined what becomes of results first presented in abstract form. The proportion reaching full publication is lower than the assumption implies.

The methodology review

The review included 425 reports covering 307,028 abstracts.7 Across varying lengths of follow-up, 37.3 per cent were subsequently published in full.7 The interval ran from 35.3 to 39.3.7 A survival analysis estimated the proportion at ten years as 46.4 per cent for all studies.7 Randomised and controlled trials fared better at 68.7 per cent, and other study types reached 44.9 per cent.7

The asymmetry by result

Publication was not independent of what the abstract reported. Where positive was defined as any statistically significant result, the risk ratio for full publication was 1.31.7 Its interval ran from 1.23 to 1.40.7 That estimate rests on 45 reports and 15,783 abstracts.7 Where positive was defined as favouring the experimental treatment, the ratio was 1.17.7 That interval ran from 1.07 to 1.28.7

The same pattern held within trials specifically. For randomised and controlled trials the ratio was 1.21 on the significance definition and 1.17 on the direction definition.7 The authors stated the conclusion directly: publication bias is present, in that positive results were more frequently published than results that were not positive.7 They also reported that the problem does not appear to be decreasing over time.7

What this means for the abstract itself

Two consequences follow from that. The first concerns the author, since more than half of results presented as abstracts were never published in full.7 Almost a third of randomised trial results were not either.7 For those, the conference abstract is the only published record, and it should be written to stand as one. The second concerns the reader, because a body of conference abstracts is not a representative sample of the work presented.7 What reached full publication was selected partly on result.7

The reporting standard that still applies

A conference abstract reporting a systematic review falls under the same twelve-item checklist as a journal abstract, which covers both settings explicitly.3 The shorter length does not remove the reporting obligation. It makes the obligation harder to meet. That is an argument for drafting against the checklist rather than against the word count.

The constraints that bind an abstract

Several constraints are stated as rules in general writing advice without a source. Some of them are genuine and traceable to a requirement, while others are conventions that vary. Separating the two prevents an author from treating a convention as binding and a requirement as optional.

The constraints that are genuine

The word limit is genuine and is set by the venue. It is published in the instructions for authors and is often enforced by the submission system. The reporting checklist for the study design is genuine. For trials and systematic reviews it specifies what the abstract must contain.13 The requirement that the abstract be consistent with the paper is genuine. The audit described above shows how often it is unmet.5

The standalone requirement

An abstract is indexed, distributed and read separately from its paper. Bibliographic databases carry the abstract without the full text, and many readers see nothing else. For the substantial fraction of conference abstracts never published in full, the abstract is the entire published record.7 It therefore has to be comprehensible without the paper, which is a design constraint rather than a stylistic preference.

Two practical consequences follow from that. Abbreviations introduced in the paper are not available to a reader of the abstract alone. Each must therefore be defined where it first appears, or avoided. References cannot be resolved from an abstract. That is why most journals exclude citations from it. Both rules are usually presented as conventions, and both follow from the way the abstract circulates.

The constraints that are conventions

Several widely repeated rules describe practice rather than any requirement. A range of 150 to 300 words is a description of common limits, and not a standard. Structured abstracts are often said to belong to the sciences and unstructured ones to the humanities. That generalisation has exceptions in both directions. The assertion that a particular format improves acceptance is supported by no source identified for this article, and it is not made here.

The order of drafting

One piece of conventional advice is supported by the evidence above, although it is usually offered without it. Writing the abstract last is recommended because the paper changes during drafting and revision. The audit of abstract accuracy found data inconsistent with the body in a substantial proportion of papers.5 That is the signature of an abstract written early and not reconciled afterwards. The recommendation is sound, and the reason for it is measurable.

Take-home message

Abstract formats reduce to structured and unstructured. The venue decides which applies, along with the headings and the word limit. Where the study has a recognised design, a reporting checklist also applies. It is selected by the design rather than by the journal. Two of those checklists have superseded earlier versions that should no longer be used. The evidence on how published abstracts actually behave is the part most often omitted from general advice. An audit of six general medical journals found 18 to 68 per cent of abstracts deficient. Their data were inconsistent with the article body, or absent from it. Among 72 trials with a nonsignificant primary outcome, 68.1 per cent of abstracts contained spin, most of it in the conclusions. Of 307,028 conference abstracts, 37.3 per cent reached full publication. Significant results reached it more often than nonsignificant ones.

How to apply this to your work

In order, from finished paper to submitted abstract
  1. Read the venue's published requirement first. It sets the format, the headings and the word limit that will actually be enforced at submission.
  2. Locate the reporting checklist for the study design. Confirm you are working from its current version rather than a superseded one.
  3. Draft the abstract last, from the final tables rather than from an earlier draft of the paper.
  4. Trace every number in the abstract to a place in the paper. Inconsistent and absent data were the two failure modes the audit identified.
  5. State the primary outcome result plainly, including when it did not reach significance.
  6. Do not describe a nonsignificant difference as equivalence, which requires a trial designed and powered for that conclusion.
  7. Check what the results section foregrounds. Within-group, subgroup and secondary results must not stand in for the comparison the study was designed to make.
  8. Define every abbreviation or remove it, and expect the abstract to be read with no access to the paper.

Common questions

How long should a research abstract be

As long as the venue permits, and no general standard governs it. The commonly quoted range of 150 to 300 words describes practice rather than a published requirement. The binding limit appears in the journal's instructions for authors or the conference submission system, and it varies substantially between them.

Which abstract format should be used

The one the venue specifies, and no other consideration overrides that. The real distinction is between structured abstracts, which carry labelled headings, and unstructured abstracts, which run as continuous prose. Where a reporting checklist applies, it determines the content that must appear; the venue determines the headings, the order and the length.1

Is there a reporting checklist for abstracts

For several designs there is. The current trial reporting statement places title and abstract first among its thirty items, numbered 1a and 1b.1 Systematic reviews have a twelve-item abstract checklist covering both journal and conference abstracts.3 The EQUATOR library, which listed 709 guidelines at the time of writing, records the extensions written for abstracts alongside the main statements.2

How often do abstracts disagree with their own papers

Often enough to make the checking worthwhile. An audit sampled 44 articles from each of six general medical journals.5 An abstract counted as deficient if its data were inconsistent with the body or absent from it.5 The proportions ran from 18 to 68 per cent, and the difference among journals was statistically significant.5 Those articles were published in 1996 and 1997. The figures establish that the problem is real rather than giving a current rate.

What counts as spin in an abstract

Reporting strategies that present the experimental treatment as beneficial when the primary outcome did not show it.6 In 72 trials with a nonsignificant primary outcome, 68.1 per cent of abstracts carried spin in at least one section.6 The conclusions section carried it in 58.3 per cent and the results section in 37.5 per cent.6

Can a nonsignificant result be described as showing equivalence

No. A nonsignificant difference is compatible with equivalence. It is equally compatible with a real difference the study was too small to detect. Establishing equivalence requires a study designed for it, with a prespecified margin. In the analysis cited, 13.9 per cent of abstracts claimed equivalence on the basis of a nonsignificant result.6

How many conference abstracts become full papers

A minority. Across 425 reports covering 307,028 abstracts, 37.3 per cent were published in full, with an interval from 35.3 to 39.3.7 A survival analysis estimated 46.4 per cent at ten years for all studies.7 Randomised and controlled trials reached 68.7 per cent.7 Statistically significant results were published more often, at a risk ratio of 1.31.7

Why are citations usually excluded from abstracts

Because a reference cannot be resolved from an abstract read on its own. Abstracts are indexed and distributed separately from the full text, and many readers see nothing else. The same reasoning governs abbreviations, each of which must be defined at first appearance or avoided. The paper's own definitions are unavailable to that reader.

Declared interest: Pubrica provides medical writing and publication support commercially. The accuracy audit cited here sampled articles published in 1996 and 1997. Abstract reporting checklists did not then exist for trials or systematic reviews. Its figures are not presented as a current rate. The spin analysis covers 72 trials with nonsignificant primary outcomes and does not describe trials generally. Nothing here should be read as a claim that a particular format improves the likelihood of acceptance. No source supporting such a claim was identified.

References and sources

  1. Hopewell S, Chan A, Collins GS, et al. CONSORT 2025 statement: updated guideline for reporting randomized trials. JAMA. 2025;333(22):1998–2005. doi:10.1001/jama.2025.4347 read 21 October 2026 at the journal, and the explanation and elaboration document read in full from the University of Glasgow repository. Cited here for the checklist's structure and its title and abstract items, not for the content of those items in detail.View the full text
  2. EQUATOR Network. Library of reporting guidelines, searchable by study type, clinical area and section of report. read 21 October 2026; the listing stated 709 reporting guidelines, and the portion returned covered roughly the first nine-tenths of the page.View the library
  3. PRISMA. Abstracts: the PRISMA 2020 extension for abstracts, a twelve-item checklist for reporting systematic reviews in journal and conference abstracts. read 22 October 2026. The page states the checklist carries twelve items and that the 2020 extension replaces the 2013 statement. It does not reproduce the items, which are held in a separate file, so no item is quoted here.View the page
  4. Beller E, Glasziou P, Altman D, Hopewell S, Bastian H, Chalmers I, Gøtzsche P, Lasserson T, Tovey D. PRISMA for Abstracts: reporting systematic reviews in journal and conference abstracts. PLoS Medicine. 2013;10(4):e1001419. doi:10.1371/journal.pmed.1001419 bibliographic record read 22 October 2026 from the Oxford University Research Archive. The full text was not retrieved, so this reference is cited only as the superseded original, and none of its content is quoted.View the record
  5. Pitkin RM, Branagan MA, Burmeister LF. Accuracy of data in abstracts of published research articles. JAMA. 1999;281(12):1110–1111. doi:10.1001/jama.281.12.1110 read 22 October 2026, in full at the journal. The source table labels the six journals A to F without mapping those letters to the named journals, so no mapping is given on this page.View the full text
  6. Boutron I, Dutton S, Ravaud P, Altman DG. Reporting and interpretation of randomized controlled trials with statistically nonsignificant results for primary outcomes. JAMA. 2010;303(20):2058–2064. doi:10.1001/jama.2010.651 read 22 October 2026, in full at the journal.View the full text
  7. Scherer RW, Meerpohl JJ, Pfeifer N, Schmucker C, Schwarzer G, von Elm E. Full publication of results initially presented in abstracts. Cochrane Database of Systematic Reviews. 2018, Issue 11. Art. No.: MR000005. doi:10.1002/14651858.MR000005.pub4 read 22 October 2026 from the Cochrane plain-language record and from an independent summary of the same review; the two agreed on every figure cited here. The publisher's own full-text page could not be retrieved.View the review record

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 22). Research Abstract Writing Formats. Pubrica. https://pubrica.com/academy/research-impactt/research-abstract-writing-formats/
Pubrica Academy. “Research Abstract Writing Formats.” Pubrica, 22 Oct. 2026, pubrica.com/academy/research-impactt/research-abstract-writing-formats/.
Pubrica Academy (2026) Research Abstract Writing Formats. Available at: https://pubrica.com/academy/research-impactt/research-abstract-writing-formats/ (Accessed: 22 October 2026).
Pubrica Academy. “Research Abstract Writing Formats.” Pubrica. October 22, 2026. https://pubrica.com/academy/research-impactt/research-abstract-writing-formats/.
Pubrica Academy, ‘Research Abstract Writing Formats’ (Pubrica, 22 October 2026) <https://pubrica.com/academy/research-impactt/research-abstract-writing-formats/> accessed 22 October 2026.
Pubrica Academy. Research Abstract Writing Formats [Internet]. Pubrica; 2026 Oct 22 [cited 2026 Oct 22]. Available from: https://pubrica.com/academy/research-impactt/research-abstract-writing-formats/
[1] Pubrica Academy, “Research Abstract Writing Formats,” Pubrica, Oct. 22, 2026. [Online]. Available: https://pubrica.com/academy/research-impactt/research-abstract-writing-formats/
@misc{pubrica_abs_2026, author = {{Pubrica Academy}}, title = {Research Abstract Writing Formats}, year = {2026}, month = {oct}, day = {22}, howpublished = {\url{https://pubrica.com/academy/research-impactt/research-abstract-writing-formats/}}, note = {Accessed: 2026-10-22} }

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.

Abstract reconciliation checklist

Work through these with the final paper open beside the abstract.

BEFORE DRAFTING
[ ] Venue's published abstract requirement located and read
[ ] Word limit noted as a number, not as a general range
[ ] Required headings noted, or confirmed that none are required
[ ] Reporting checklist for the study design located
[ ] Checklist confirmed to be the current version, not a superseded one
[ ] Abstract extension for that design identified, where one exists
[ ] Paper finalised, so the abstract is drafted from final tables

CONTENT
[ ] Objective stated, matching the objective in the paper
[ ] Design named, using the term the paper uses
[ ] Setting and participants described
[ ] Primary outcome named explicitly
[ ] Primary outcome result given, whatever its direction
[ ] Effect estimate given with its precision, not a bare p-value
[ ] Harms reported where the design collected them
[ ] Conclusion limited to what the primary outcome supports
[ ] Registration or review identifier included where the venue asks

NUMERICAL RECONCILIATION
[ ] Every number in the abstract located in the paper
[ ] Denominators in the abstract match the denominators in the tables
[ ] Percentages recomputed from the final counts
[ ] Confidence intervals copied, not retyped from memory
[ ] No number appears in the abstract that appears nowhere in the paper
[ ] Abstract re-checked after the last revision to the results

LANGUAGE AND HONESTY
[ ] Nonsignificant primary outcome stated as nonsignificant
[ ] No claim of equivalence resting on a nonsignificant difference
[ ] No within-group comparison standing in for the between-group one
[ ] No subgroup result standing in for the primary outcome
[ ] No secondary outcome foregrounded over the primary
[ ] Conclusion would read the same to a reader who disliked the result

STANDALONE READING
[ ] Every abbreviation defined at first use, or removed
[ ] No citations, unless the venue explicitly permits them
[ ] No reference to figures, tables or sections of the paper
[ ] Abstract read once in isolation, by someone who has not read the paper
[ ] Abstract comprehensible to that reader without further explanation
PA
Pubrica Academy

Written and maintained by the subject matter experts of Pubrica who do this work. Articles are revised when the standard underneath them moves, and each one carries the date it was last checked against its sources.

How Pubrica helps

We draft abstracts against the reporting checklist for the design. We reconcile every figure against the final tables, and check that the conclusion states what the primary outcome supports.

Explore

Related reading in the Academy.

Picked from the same sub-topic and category as this piece, newest first.

Find something else

Search the rest of the Academy.

Open with no account. Type what you are working on, or open the full library to filter by category, sub-topic or resource type.

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.