Reproducibility checklist

Keep the files and decisions behind your analysis in one research folder.

Save these files

  • Data exports: the CSV files used in your analysis. Include Weighted query terms if you used a weighted dictionary query. See download names and files.
  • Query and settings: the Search configuration JSON, where available. Also record the exact query and any settings the file does not capture.
  • Analysis files: your downstream code or analysis notes, plus the source and version of any external data used.

Save copies outside the platform. Generated dataset results are retained for 90 days. The dataset definition remains available, but a refresh updates the dataset with the latest available calls. Keep the original exports and review notes with your research record.

Add these notes

  • Sentence review: examples of relevant and irrelevant matches, terms added or excluded, and how you handled ambiguous cases.
  • Sample decisions: your intended sample, availability findings, restrictions, and any zero rows excluded with reasons.
  • Analysis decisions: join keys and match checks, unmatched records, and whether you kept raw counts, controlled for length, or normalized them. Record the denominator if you normalized.

Copy a research record

Paste this template into your project notes and fill it in. Mark conditional items Not used when they do not apply. A share link can help locate the dataset; keep the downloaded files alongside the record.

View and copy the template
DATASET AND EXPORT
Dataset name and ID:
Export date:
Exported filenames:
Configuration JSON file (if available):
Corpus statement used for the analysis:

QUERY AND SETTINGS
Exact query:
Date range:
Search engine:
Document and within-document filters:
Scoring settings (including any older context setting):
Selected measures:
Display settings and export options:

REVIEW AND SAMPLE
Relevant and irrelevant match examples:
Terms added or excluded, with reasons:
Ambiguous cases and decisions:
Intended sample and availability findings:
Sample restrictions and reasons:
Zero rows kept or excluded, with reasons:

ANALYSIS
Analysis dataset filename and version:
Raw counts, length control, or normalization:
Denominator (if normalized):
Code or analysis notes location:

IF USED
Join key, identifier checks, unmatched records:
External dataset source and version:
Keyword Tool settings and saved protocol:
Source dataset names/IDs and combination/comparison logic:
Weighted query terms file:
Dataset share link: