A batch of paper abstracts into a sourced reference table

When to use

Twenty or thirty abstracts to organise into a table you can cite while writing a review. The fear is wrong years, claims attached to the wrong paper, or reviews mistaken for primary studies.

Preparation

Host version

To fill (record the exact rc version when testing)

Model

To fill

Mode / permissions

Standard mode + modify-within-workspace is enough to start

Plugins required

None (add name & version after testing if used)

Input materials

Keep in a separate workspace; back up originals first

Full instruction

Read the abstract files under papers/ and build a reference table at refs.md.

[Output]
Columns: id | authors (first author et al.) | year | study type | sample or data source | main finding (one sentence) | relation to my topic | verbatim excerpt | source file:line

[Hard constraints]
1. 'Study type' must be one of exactly four: randomised controlled trial, observational study, review/meta-analysis, other (specify). The classification must come from what the abstract states, not your inference. If unstated, write 'not stated in abstract'.
2. 'Sample or data source' comes from the abstract. If it only says 'large-scale data' with no figure, write 'large-scale data (no figure given)'.
3. Year and authors follow the abstract file. If the year is absent, write 'not stated'; never fill it from a version you remember.
4. 'Main finding' must be the abstract's conclusion, not your inference. If absent, write 'no conclusion in abstract'.
5. Every row needs a verbatim excerpt of one or two sentences from the abstract that supports the finding column.

Report at the end: total papers, count per study type, how many lack a year, how many lack a conclusion, any files that could not be read.

Do the first five so I can check the format.

Copy full instruction

Run log

Output

After a successful run, list files and screenshots here, and note whether originals changed.

Acceptance checklist

Tap to tick each item off as you verify.

• Study type uses only the four allowed values
• Every row has a verbatim excerpt supporting its finding
• Missing year or conclusion is stated as missing, not filled
• No review is counted as primary research — the most common error in such tables
• The four type counts plus missing items equal the total

Failures & boundaries

• Models remember papers and will fill years and conclusions from memory. State 'missing means missing' explicitly, then spot-check two rows against the source
• Counting a review as primary research is the most insidious error: review conclusions are stronger and easier to cite, hence more tempting. The type column must follow the abstract's own self-description
• Abstracts routinely omit sample size. Do not expect this table to replace reading the paper; it is for screening
• Blocked on: Needs real paper abstracts. Synthetic ones miss the two features that matter: how different study types describe themselves, and the genuine absence of sample sizes — exactly the errors this recipe guards against.

Claim this recipe Back to recipes