Journal Article Writing — Section by Section, From Title to Conclusion

Narrow rather than survey. Open at the field, move to the specific problem, then state precisely what prior work has not established — and stop there, handing over to your aim as the answer to that gap. Around a sixth of the paper is the usual proportion. If your introduction is longer than that, it has become a literature review, which is the commonest structural fault in papers built from a thesis chapter.

Results report; discussion explains. A results section gives values, direction and significance and stops, while the discussion places each finding against prior work and argues what the field now knows. Interpretation appearing in results is marked twice — once for arriving early and once for leaving the discussion with nothing to say.

That is the point of sending what you have written. The section is calibrated to your register rather than to a house style, because a paper where paragraph forty reads differently from paragraph ten raises exactly the question you were trying to avoid. Co-authored papers are handled the same way — send whichever co-author's writing the new part must sit beside.

Yes, and reproducibility is the standard we write to rather than word count. Equipment named with specifications, sampling stated, instruments cited with their reliability, procedure given step by step and analysis named test by test. The test is whether a stranger in your field could rerun the study without emailing you.

Yes, at RM 380 for a post-review rewrite. Overreach is usually a handful of verbs rather than a broken argument — "proves" where the design supports "suggests", recommendations extending past a single-sector sample. We recalibrate the claims to what your data can carry and flag anywhere the finding genuinely cannot support the point being attempted.

Yes, at RM 1,100. Journals expect substantial extension over a conference version — additional data, deeper analysis, or a fuller discussion and literature positioning — and they expect the earlier paper declared in your cover letter. Submitting the same work twice without declaring it is what causes real trouble, not the reuse itself.

Yours. A discussion argued against literature you have not read is a viva you cannot survive and a claim you cannot defend to a reviewer. Where a point genuinely needs a source you do not have, it comes back as a margin comment rather than as a reference we located to fill the gap for you.

It writes the shape of one, and each section fails differently. An introduction comes back as background with no gap. A methods section reads plausibly and omits the parameters that make replication possible. A discussion cites studies you have never read. The one that catches people is results — generated prose reports values that do not appear in your own output, and that is checkable in a minute.

Work at the section's job rather than its wording. An introduction is fixed by naming the gap from your own citations; a discussion by replacing general commentary with named studies you actually reviewed; a methods section by adding the specifics only you know. Doing that removes most generated text and fixes the weakness that made the section thin, which a paraphrasing tool does neither of.

Introductions and discussions, because both are continuous prose written in an even academic register — the pattern detectors were trained on. Methods and results usually score lower simply because they are dense with numbers, instrument names and test names. A paper flagged only in its introduction is worth checking before assuming the whole thing needs work.

Completely. Your data, drafts and sections are used only for your order, never reused, never shared beyond the writer assigned to you, and deleted once the order closes. Unpublished findings are the most sensitive thing a researcher owns, and priority is not something you can recover once it is lost.

Put your own reading into it. Generated introductions are flagged because they are general — field, importance, aim, with nothing that could only have come from you. Name the specific studies you actually read, state what those studies left unresolved in your own words, and let the gap emerge from your citations rather than from a claim that research is limited. Specificity fixes the score and the section at once.

Sometimes, and it depends on how much of the flag that section was carrying. Detectors score the document, so if your introduction was generated and everything else was yours, rebuilding it can move the number considerably. If the whole paper was generated, one rebuilt section will not rescue the other six — and you should know that before paying for one.

This is the assumption that catches people. A methods section reads plausibly when generated and quietly omits the parameters that make replication possible — the calibration step, the exact instrument model, the sampling interval. Referees test methods by trying to picture running the study, so the omissions surface at exactly the point you cannot afford them.

Raise it before submission, not after. Every co-author signs off on the whole paper, so an undeclared generated section becomes a problem for all of you rather than for the person who wrote it. Check the citations in that section first — fabricated references are the usual damage — then agree together how the use is declared under your target journal's policy.

Run each section separately rather than the whole document. Most tools give a per-passage breakdown, and pasting sections in one at a time isolates the problem in a few minutes. In practice it is almost always the introduction or the discussion — both are continuous prose in an even academic register, while methods and results score lower because they are dense with numbers and named tests.