Gemini

Verified against Gemini · 2026-07-31

Critique a research paper's methodology using its actual figures, not just its abstract

A multimodal PDF prompt that reads a paper's figures, plots, and tables directly — sample sizes shown in a caption, error bars, axis scales — as evidence for a methodology critique, instead of a summary built only from the abstract and discussion text.

Gemini (Gemini 3 Pro)Gemini app5 fillable variables

The prompt

Ready to copy — highlighted parts are example details you can swap.

I've uploaded a PDF of a research paper: a 14-page peer-reviewed paper on a new drug's effect on blood pressure, including 4 figures and 2 tables. Read the figures and tables as images, not just the surrounding text — the critique below depends on what the figures actually show, which the abstract's prose claims may or may not accurately represent.

READER
a graduate student who understands basic statistics but is not a specialist in this subfield

CRITIQUE FOCUS
sample size adequacy, whether the reported effect size is visually consistent with the error bars shown, and whether the abstract's conclusion matches what the figures actually show

INSTRUCTIONS
1. For each figure relevant to the critique focus, describe what it actually shows — axis labels, scale, error bars or confidence intervals if present, sample size if stated in the caption — before drawing any conclusion from it.
2. Check whether the abstract's or discussion section's claimed effect is visually consistent with what the figure actually plots. If the prose claims a "significant" or "substantial" effect, look at whether the plotted error bars or confidence intervals actually support that framing, and say plainly if they don't.
3. check whether any figure's y-axis is truncated in a way that visually exaggerates a small difference
4. Note the sample size for each key result, and whether it's large enough to support the strength of the claim being made — don't just report the number, say whether it changes how much weight the result should carry.
5. Distinguish between a methodological weakness (a real problem with how the study was designed or reported) and a stylistic choice (a chart style you'd personally do differently) — only the former belongs in a critique aimed at deciding whether to rely on this paper.

CONTROL GROUPS AND COMPARISONS
If a figure shows an experimental group against a control or comparison group, check whether the two groups are shown on the same scale and axis, and whether the figure makes it easy or hard to actually compare them visually — a comparison presented across two separate panels with different y-axis ranges can obscure a difference that would be obvious if both groups were plotted on one shared scale, and that's a methodology-adjacent presentation choice worth naming even if the underlying data is sound.

WHAT THE TABLES SHOW THAT THE FIGURES DON'T
Read any tables in the paper as carefully as the figures — a table often contains the exact numeric values (means, standard deviations, p-values) that a figure only represents visually, and a precise number in a table can confirm or contradict an impression formed from a figure's visual scale alone.

OUTPUT FORMAT
a structured critique aimed at deciding whether this paper is strong enough to cite in a literature review, not a full rewrite of the paper. Structure it as: a one-paragraph summary of what the paper claims, then a per-figure section noting what it actually shows and whether that supports the claim, then a section on what the tables add or contradict, then a final verdict on whether the methodology issues found are minor, moderate, or serious enough to affect whether this paper should be relied on for the stated purpose.

Customize

Optional — swap in your own details for the highlighted parts above.

Why this works

Reading figures as images lets Gemini check whether the abstract's claimed effect is visually supported by the plotted error bars and confidence intervals, which a text-only reading of the abstract and discussion section cannot verify at all — the abstract's prose claim and the figure's actual plotted data can, and in a non-trivial number of published papers do, tell slightly different stories, and that gap is only visible if the figure itself is actually read rather than taken on faith from the surrounding text. The truncated-axis check targets a well-documented and specific way figures visually overstate a real but small effect: a y-axis that starts at 90 instead of 0 makes a 2-point difference look like a dramatic gap on the page, and this is checkable only by looking at the actual axis on the actual chart, not by reading what the caption or discussion section says the figure shows. Requiring the sample size to be weighed, not just reported, matters because a number alone ('n=24') means very little to a reader without domain expertise unless it's explicitly connected to how much confidence that sample size can actually support for the specific claim being made — the same n might be entirely adequate for one kind of effect and clearly underpowered for another. Separating methodological weaknesses from stylistic disagreements keeps the critique focused on what actually determines whether the paper should be trusted for a given purpose, rather than drifting into an aesthetic review of chart choices that have no bearing on whether the underlying finding holds up. Checking whether an experimental and control group are plotted on the same scale targets a specific and genuinely common presentation choice that can obscure a real methodological weakness: two panels with different y-axis ranges make groups look more similar or more different than a shared-scale comparison would, and this is a fact about the figure's construction that's only visible by actually looking at both axes side by side, not something a text-only read of the results section would surface. Cross-referencing the tables against the figures matters because the two convey different kinds of precision — a figure gives a fast visual impression of scale and spread, while a table often carries the exact numbers (a specific p-value, a precise standard deviation) that either confirm or quietly undercut the confidence a figure's visual presentation seems to project, and a critique that reads only the figures misses whichever of those two sources happens to tell the less flattering part of the story.

Verified against

Gemini Gemini 3 Pro · 2026-07-31

Changelog

  • 2026-07-31 Initial publish, verified against Gemini 3 Pro on a 14-page clinical study PDF with four figures.

Need this built into your business?

If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.

EXPLORE WHAT SCULT BUILDS
Turn a stack of scattered documents into one sourced decision memoA long-context synthesis prompt that loads every related document into the same context window as one connected evidence base and returns a single decision memo with inline citations and flagged disagreements, instead of a chain of disconnected per-document summaries that lose the connections between them.Gemini (Gemini 3 Pro)Gemini app2026-07-20Diff a redlined contract against the original and flag every substantive changeA long-context prompt that loads both the original and the redlined version of a contract into the same window and returns a change-by-change risk analysis, instead of a generic first read that misses a clause quietly weakened somewhere the visible tracked-changes markup doesn't draw the eye.Gemini (Gemini 3 Pro)Gemini app2026-07-21Compare competing sources side by side without smoothing over where they disagreeA long-context prompt that loads several full reports, reviews, or papers at once and produces a structured comparison table instead of a blended, opinion-free summary that quietly hides where the sources actually disagree.Gemini (Gemini 3 Pro)Gemini app2026-07-22Turn an entire small-to-mid codebase into an onboarding doc a new engineer can trustA long-context prompt that reads a whole repository — every file pasted or uploaded, not a handful of representative snippets — and produces an onboarding document that names actual file paths and real architectural seams, instead of generic advice that could describe any Next.js or Python project.Gemini (Gemini 3 Pro)Gemini app2026-07-23
All Gemini prompts

Check your AI visibility

One URL in, a 0–100 score and the exact fixes out.

RUN THE CHECK

Browse all the tools

15 tools across six categories
13 of them never send your data anywhere

Free · No signup · No trial clock

SEE THE DIRECTORY