Services / Results chapter review
A statistician checks the results chapter you wrote
You have written it. Before it goes to a supervisor or an examiner, someone should check that the tests were right, the numbers in the text match the tables, and no claim is stronger than the data supports. That is this.
- Line-by-line statistical comments
- Every figure cross-checked
- Reporting corrected to APA
- Likely examiner questions flagged
A results chapter review checks that the statistical tests suit the design, that assumptions were tested and reported, that every figure in the text matches the tables, that reporting follows APA convention, and that no claim is stronger than the analysis supports.
What is included
What you get with results chapter review
Test selection reviewed
Whether the analysis you ran suits the design and the data — and what to do if it does not.
Assumptions and reporting
Whether the required assumption checks were run and reported, and whether the conclusions survive them.
Numbers cross-checked
Every figure in the prose checked against the tables. Mismatches between text and table are among the most common corrections examiners issue.
APA and house style
Statistics reported to the exact convention your department requires, including decimals, italics and exact p-values.
Claims kept proportionate
Where the writing over-reaches what the analysis shows, we mark it and suggest wording that still says something.
Examiner questions flagged
The specific things a viva panel is most likely to press on, so you can prepare an answer.
Who it is for
Two kinds of client, one standard of work
- A chapter written and nothing checked
- Uncertainty about APA reporting
- A supervisor who queried the statistics
- A viva approaching
- A results section before journal submission
- Reviewer comments about statistical reporting
- Co-authored work needing an independent check
- Turning a completed chapter into a journal article
The review
What gets checked, in order
A results chapter is judged on whether the analysis suits the design, whether the reporting is complete, and whether the claims match the evidence. We check all three, in that order, because a reporting problem in an inappropriate analysis is not worth fixing.
Test selection
Whether the analysis you ran matches the design described in your methods, your variable types, and your sample. Mismatches between methods and results are an immediate flag.
Assumptions
Whether the required checks were run, whether they were reported, and whether the conclusions survive them. The most common single omission in submitted chapters.
Numbers and tables
Every figure in the prose cross-checked against the tables. Mismatches are among the most frequently issued corrections and are entirely avoidable.
Reporting format
Exact p-values, effect sizes, confidence intervals, degrees of freedom, italicisation and decimal places, to APA or your department's convention.
Proportionate claims
Where the writing says caused, improved or proved and the design supports only associated with, we mark it and suggest wording that still says something.
Examiner questions
The specific points a viva panel is most likely to press on, listed so you can prepare answers.
Common findings
The corrections we issue most often
| What we find | Why it matters |
|---|---|
| Assumption checks not reported | The reader cannot tell whether the test was appropriate. Often the quickest correction to make and the most common to omit. |
| p-values reported as p = .000 | No p-value is zero. It should read p < .001 — a small point that signals unfamiliarity with the convention. |
| Effect sizes missing | Significance without magnitude leaves the reader unable to judge importance, and it is the question an examiner asks next. |
| Text and table disagree | Usually the residue of a late reanalysis that updated the tables but not the paragraphs written around them. |
| Causal language from correlational data | The most substantive problem we find, and the one most likely to cost marks. |
| Every test reported, significant or not, with no structure | A results chapter is an argument, not a log of everything the software produced. |
| Tables that duplicate the text | Either the table or the sentence should carry the detail, not both. |
The boundary
What we will not do
Not this
- Rewriting the chapter for you
- Producing text for you to submit as your own
- Improving the prose style or proofreading
- Reanalysing quietly and presenting new results as yours
This
- Comments on what you wrote, with reasons
- Suggested wording where a claim over-reaches
- Corrections to statistical reporting format
- Telling you plainly if reanalysis is needed, and quoting for it separately
The distinction is practical, not just principled: a chapter written in someone else's voice is difficult to defend at a viva, and examiners are experienced at spotting the seam between a student's writing and someone else's.
Bad news, delivered early
If we find something serious
Occasionally the review finds that the analysis itself will not hold — the wrong test for the design, an assumption violated in a way that changes the conclusion, or a dataset that cannot support the claims made of it.
We tell you plainly, explain what it means for the chapter, and set out the options with the cost of each. Reanalysis is quoted separately so you can decide rather than being committed to it.
Finding it now costs a delay. Finding it in a viva costs a resubmission — and by then the fix is the same amount of work, undertaken under considerably more pressure.
What we need from you
The chapter, and the dataset and output if you have them. With the data we can verify the numbers and re-run checks; without it we can still review reporting, internal consistency and interpretation, which catches the majority of issues.
Getting started
What we need from you
Without the dataset we can still check reporting, internal consistency, formatting and interpretation, which catches the majority of problems. With it, we can also confirm the numbers are what the analysis actually produced — which is where the more serious findings tend to come from.
Practicalities
Turnaround, cost and revisions
Priced by chapter length and returned in three to five working days. The quote is fixed before you commit, and names the deadline.
The chapter you send to a supervisor sets their expectations. It is considerably easier to fix a reporting problem before that impression forms than to correct it afterwards.
APA conventions
How each test should be reported
Most reporting corrections we issue are variations on the same few conventions. This is what correct looks like for the tests that appear in the large majority of chapters.
| Test | Report as | Also required |
|---|---|---|
| Independent t-test | t(df) = 2.41, p = .018 | Cohen’s d, group means and SDs |
| Paired t-test | t(df) = 3.02, p = .004 | Cohen’s d, mean difference with CI |
| One-way ANOVA | F(2, 87) = 4.55, p = .013 | η² or partial η², plus post-hoc tests |
| Chi-square | χ²(1, N = 240) = 6.14, p = .013 | Cramér’s V, and expected counts if any are small |
| Correlation | r(118) = .34, p < .001 | 95% CI; note whether Pearson or Spearman |
| Linear regression | β = 0.42, t = 3.31, p = .001 | R², adjusted R², F for the model, CIs |
| Logistic regression | OR = 1.84, 95% CI [1.21, 2.79], p = .004 | Model fit, and events per predictor |
p-values to three decimals, except below .001 where p < .001 is correctp = .000 — no p-value is zerop = .018, not 0.018)After the review
Working with supervisor feedback
Our comments and your supervisor’s will occasionally disagree. That is normal, and it is usually a difference of convention rather than of correctness — departments have house preferences on effect size reporting, on whether to report non-significant results in full, and on how much detail belongs in tables.
Where the disagreement is substantive rather than stylistic, we will explain the statistical reasoning so you can take it back to your supervisor and have the conversation on the evidence. What we will not do is tell you to overrule them: they are marking the work and they know your department’s expectations.
The most useful outcome is usually that the review surfaces a question your supervisor had not had time to look at closely, and the two of you resolve it before submission rather than at a viva.
How it works
Three steps, entirely in writing
Send what you have
Attach the data, or just describe the project. No account, no mandatory call, and a five-field form rather than a fifteen-field one.
Approve a fixed quote
Usually within one working day, naming the statistician assigned, the deliverables and the deadline.
Receive the work
Output, syntax and a plain-English explanation — with follow-up questions answered at no extra cost.
Client feedback
What clients say about this work
“Excellent attention to detail. They corrected inconsistencies and ensured my statistical findings were presented professionally.”
Proof
This work, on real projects
Clients are not named and no identifying detail is published, so what is described is the statistical problem — which is the part that shows whether a consultancy knows what it is doing.
Random Effects Meta-analysis of Forty-Five Published Studies
Combining heterogeneous evidence into a defensible overall estimate, with publication bias investigated rather than assumed away.
Education · Multilevel modellingMultilevel Modelling of Student Performance Across Twenty Departments
Accounting for students clustered within courses and departments, where ignoring the hierarchy would have manufactured significance.
Life sciences · Linear mixed effects modelsLinear Mixed Effects Models for an Intervention Measured Over Six Occasions
Recovering a defensible effect estimate from repeated measures with missing follow-up, where repeated measures ANOVA could not be used.
Your statistician
Who does this work
Member of the Royal Statistical Society · International Society for Clinical Biostatistics
Clinical trials, epidemiology and evidence synthesis, with a focus on longitudinal and multicentre studies.
This page was reviewed for statistical accuracy by Sidra Khan on 14 August 2026.
Common questions
Answers before you ask
Will you rewrite the chapter?
No — we comment on what you wrote and suggest wording where a claim over-reaches. The writing stays yours, which is the point of having it checked.
What do you need from me?
The chapter, the dataset if you have it, and the output. With the data we can verify the numbers; without it we can still check reporting, consistency and interpretation.
What if you find a serious problem?
We tell you plainly and explain the options, including reanalysis. Finding it now is considerably better than an examiner finding it later.
Can you check the whole thesis?
We review the statistical content wherever it appears — usually methods and results. We do not proofread or edit prose.
How long does it take?
Three to five working days depending on length. Tell us your deadline before you commit.
Related
Often needed alongside this
Dissertation statistics help
Analysis of the data you collected, explained test by test so you can defend it.
From £95SPSS data analysis
The right tests, annotated output, reusable syntax and APA-ready tables.
From £95Academic CV review
Line-by-line feedback on an academic CV for PhD applications and research posts.
£95Send the data. Get a fixed quote.
Attach your dataset or just describe the project. A named statistician replies with a price and a deadline, usually within one working day.