One construct throughout
Confidence is not relabelled performance, and a 0.7-point difference is not presented as a percentage improvement in research ability.
Academic editing · Demonstration study
A result that says exactly what the numbers support
An editorial demonstration linking the denominator, measured outcome, results sentence and response to a reviewer.
The brief
What needs to change?
Edit a fictional workshop manuscript so that its numerical reporting and conclusion agree. All numbers are synthetic. The presentation study uses the same fixture, not a second result.
The challenge
Confident prose can overstate a modest descriptive comparison. The editor must preserve the actual construct, identify missing follow-up and separate a mean change from statistical significance or effectiveness.
Working excerpt · Illustrative material
Results excerpt, data note and author queries
| Measure | Before | After |
|---|---|---|
| Paired respondents | 60 | 60 |
| Mean confidence (1–5) | 3.1 | 3.8 |
+0.7 scale points among completers. Not an effectiveness estimate.
80 baseline responses; 60 paired (75%); 20 missing follow-up (25%). No comparison group. No significance test or confidence interval has been calculated.
| Query | Reason | Author action | Status |
|---|---|---|---|
| AQ1 · Outcome | Confidence is a self-report measure, not writing performance. | Confirm the instrument wording and its response anchors. | Open |
| AQ2 · Denominator | 80 entered at baseline; only 60 have paired responses. | Confirm the pairing process and explain the 20 missing follow-ups. | Open |
| AQ3 · Inference | The aggregate frequencies do not preserve individual pairings. | Supply paired records and an analysis plan before adding inferential statistics. | Not calculated |
| AQ4 · Scope | A before–after difference does not isolate a workshop effect. | Retain the descriptive conclusion; do not substitute “effective”. | Reflected in the revised excerpt |
| Confidence score | Before: count | After: count |
|---|---|---|
| 1 | 2 | 0 |
| 2 | 8 | 4 |
| 3 | 36 | 14 |
| 4 | 10 | 32 |
| 5 | 4 | 10 |
| Total | 60 | 60 |
| Weighted total | 186 | 228 |
| Mean | 3.1 | 3.8 |
01 / Define the analysis population
Synthetic demonstration—not observations or customer results: 80 people provide a baseline questionnaire; 60 also provide follow-up; 20 do not. Both reported means refer to the same 60 paired respondents. The measure is self-reported confidence on a scale from 1 to 5, not research performance. No comparison group is included.
02 / Make the arithmetic reproducible
Synthetic frequencies for scores 1, 2, 3, 4, 5: baseline among paired respondents = 2, 8, 36, 10, 4; follow-up = 0, 4, 14, 32, 10. Each totals 60. Weighted sums are 186 and 228; 186/60 = 3.1 and 228/60 = 3.8; difference = 0.7 scale points. Completion = 60/80 = 75%; missing follow-up = 20/80 = 25%.
03 / Revise the substantive claim
Original demonstration sentence: “The workshop significantly improved postgraduate research performance.” Revised: “Among the 60 respondents who completed both questionnaires, mean self-reported confidence increased from 3.1 to 3.8 on a five-point scale. The study included no comparison group.” The revision names the analysed group and measured construct while removing unsupported significance and causality.
04 / Leave unresolved questions visible
AQ-01: Confirm the item wording and identical scale anchors. AQ-02: Explain response matching and missing-follow-up handling. AQ-03: Supply respondent-level paired data and the proposed analysis before adding an inferential result. AQ-04: Clarify recruitment and whether non-completers differed at baseline. These questions have not been answered by this demonstration.
05 / Connect the response to the revision
Fictional reviewer query—not a real journal decision: “Please justify the effectiveness language.” Demonstration response: “The revised wording describes the mean difference in self-reported confidence among paired respondents. Causal and statistical-significance claims have been removed. The discussion identifies the absent comparison group and incomplete follow-up.” Real manuscript locations would be inserted only after a revision file exists.
Read alongside the reasoning and limitations below. These excerpts are not a complete commissioned deliverable.
Editorial & methodological judgement
Why these decisions?
The value is in the choices that connect the brief to the result.
No invented precision
The synthetic marginal distributions make the means reproducible but do not specify individual pairings. They are not enough to derive a paired-change confidence interval or p-value from the mean difference alone.
Queries are part of the work
Missing information and substantive research decisions remain visible to the author. Language editing does not silently repair an absent analysis.
Review checklist
What to check.
- Abstract, table and discussion use the same 60-person paired denominator.
- All synthetic frequencies, means and completion proportions reconcile.
- Descriptive change is not called causation or statistical significance.
- No journal-specific audit or acceptance outcome is implied.
A possible commissioned handover
What to specify.
- Before/after results excerpt
- Synthetic denominator and frequency table
- Author-query log
- Fictional reviewer-response example
The actual deliverable list, format and review stages are agreed in your quote.
Bring your own brief
Make the next step specific.
Explore this service See what your quote includes Discuss your projectAnother kind of work