Results

Report your results so they can't be misread

The results section has the simplest job in the paper and is easy to get wrong: say what you found, in order, with the numbers. No explanation, no repeating every cell of a table, and no claim the statistics don't support.

Try it freeSee the steps

Free plan includes 100 AI Autocomplete / Ghost Text · 5,000 AI Writing Words · 100 AI Rewrites.

AI Edit

Results section · 8 suggestions
Conciseness

As can be clearly seen in Table 2, it is evident that

Table 2 shows that

AcceptReject
Consistency

p<0.05 / p < .05 / P = 0.03

p < .05, p = .03

AcceptReject
Academic tone

Group A did way better

Group A scored higher

AcceptReject
The problem

What reviewers flag in results sections

Most of it comes down to putting things in the wrong place.

Interpretation creeps in

"This suggests that..." belongs in the discussion. The results section says what was observed.

Tables repeated in prose

A paragraph that reads out every value in the table beneath it. The text should point to the pattern; the table holds the numbers.

No visible order

Analyses reported in the sequence they were run, not the sequence of the research questions.

Incomplete statistics

A p value with no test statistic, effect size or sample size. Or "significant" used to mean "large".

Step by step

How to do it in LitraMind

Finish the analysis and the tables first. This section is written from them.

  1. List your research questions as subheadings

    In your document, add one subheading per question or hypothesis, in the order the introduction gave them. Start with a short description of the sample.

    About Write New Paper

  2. Decide what goes in a table or figure

    Anything with more than three or four numbers belongs in a table. Trends over time or comparisons across groups often work better as a figure. Build them in your analysis software.

  3. Write one finding per paragraph

    For each, state the result in words, give the key statistic, and point to the table or figure. Type the numbers yourself from your output. Ghost Text can help with the wording around them.

    About AI Autocomplete

  4. See how your target journal reports similar analyses

    Upload a recent paper from the journal that uses the same method and ask how it reports the statistics. Follow the page reference and match the format.

    About Chat with Files

  5. Remove interpretation and filler

    Search for "suggests", "because" and "interestingly" and move those sentences to the discussion. Use Trim on wordy lead-ins such as "it can be seen that".

    About AI Rewrite

  6. Check consistency, then check every number

    Run AI Edit for consistent statistical notation and terminology. Then compare every figure in the text against your output and your tables, one by one.

    About AI Edit
Start step 1 free
The guide

How to report results

Conventions vary by field and style guide, so check your target journal. The principles are stable.

Text, table or figure?

Each does a different job. Choose by what the reader needs to do with the information.

Where to put a result
UseWhenAvoid
TextOne to three numbers; the main finding in wordsListing many values in a sentence
TableExact values the reader may want to look up or compareRepeating the same values in the text
FigurePatterns, trends and comparisons at a glanceDecorative charts that show two numbers

Reporting statistics

Give the test, the test statistic with its degrees of freedom, the exact p value, and an effect size with a confidence interval where your field expects it. Report descriptive statistics, such as means and standard deviations, alongside the inferential test.

"Significant" should mean statistically significant and nothing else. A small effect can be significant in a large sample, so report the size of the effect and let the discussion judge its importance.

Qualitative results

Organize by theme. For each theme, define it, describe the range of views within it, and support it with a small number of well-chosen quotations, identified by a participant code. Say how common a view was without implying statistical precision.

Many qualitative traditions combine findings and interpretation in one section. Follow the convention of your approach and your journal.

Report everything you said you would

Include results that didn't support your hypothesis and analyses that found nothing. Leaving them out distorts the record, and reviewers who compare your results with your methods will notice the gap.

Label any analysis you added after seeing the data as exploratory.

Tense

Past tense for what you found ("scores were higher"). Present tense when pointing to a table or figure ("Table 2 shows").

Written by the LitraMind team. Last reviewed .

In their words

What researchers and students say

“It helped me present the main findings in a much more logical sequence without interpreting them prematurely.”

James W•••PhD Candidate, Biomedical SciencesN••• University, United Kingdom

“My results section contained too much commentary. Separating what the data showed from what it meant made the manuscript much cleaner.”

Natalie C•••Doctoral Researcher, PsychologyW••• University, United States

“It was especially useful for checking that the text complemented the tables rather than simply repeating every number.”

Akash P•••Research Scholar, Public HealthN••• Institute of H••• R•••, India

“The sequence of analyses became easier to follow once the results were organised according to the research questions.”

Eva S•••PhD Researcher, EconomicsR••• University, Germany

“It helped me decide which findings belonged in the main text and which could be left to tables or supplementary material.”

Antoine M•••Doctoral Candidate, NeuroscienceP••• Institute of N•••, France

“The final section was shorter but actually easier to understand because the main findings were much more visible.”

Michelle L•••PhD Candidate, EngineeringP••• Institute of T•••, Singapore
The tools

What this workflow uses

All of them live in the same editor, so you don't move text between apps.

AI Autocomplete

Ghost Text suggests the next sentence as you type. Press Tab to keep it, keep typing to ignore it.

About AI Autocomplete

Chat with Files

Ask questions of a PDF or DOCX and get answers with page references you can check.

About Chat with Files

AI Rewrite

Seven rewrite modes for selected text: Paraphrase, Make Academic, Improve Fluency, Simplify, Trim, Change Tone and Synonyms.

About AI Rewrite

AI Edit

Grammar, tone, clarity and consistency suggestions that you accept or reject one at a time.

About AI Edit
Straight answer

What this won't do

This is the section where an error does the most damage.

  • It doesn't run your analysis or make tables and figures. Those come from your statistics software.
  • It can't verify your numbers. A rewrite could alter a figure, so compare every value with your output after editing.
  • It doesn't choose the right test. That's a methodological decision for you, your supervisor or a statistician.
Questions

Questions people ask

Short answers before you start. Plan limits are on the pricing page.

See plans and limits
Can AI write my results section?
It can help you phrase findings and keep notation consistent. It has no access to your data, so the numbers must be entered by you and checked against your output. Never accept a suggested figure.
Should I interpret my results in the results section?
In most quantitative papers, no: report in the results and interpret in the discussion. Some journals and many qualitative papers use a combined section. Check the format you're writing to.
How should I report p values?
Most style guides ask for the exact value, such as p = .03, with p < .001 for very small values. APA style omits the leading zero; other styles keep it. Be consistent.
Do I have to report non-significant results?
Yes, for every analysis you planned. Selective reporting misleads readers and is considered a questionable research practice.
How long should the results section be?
As long as the findings need and no longer. Well-designed tables and figures shorten it considerably.
Keep going

Related use cases

All use cases

Discussion section

Interpret your results against the literature, be honest about limitations, and stop short of claims the data can't carry.

Read the guide

Methodology section

Describe what you did precisely enough to repeat, say why you did it that way, and keep terms and tense consistent.

Read the guide

Lab report writing

Turn bench notes into a clear report: methods in the past tense, results without interpretation, and a discussion that explains your errors.

Read the guide

Start with your research questions as subheadings

Add them to your document in LitraMind and write one finding under each.

Create a free accountView pricing