How long does an Unjournal evaluation take?

Evaluator-reported time: median 8 hours, mean 9.6 hours, based on 33 evaluation records. Includes the range, quartiles, and missing-data limits.
evaluators
process
survey
Hybrid
Author

The Unjournal

Published

November 12, 2025

Modified

November 2, 2026

Production: Hybrid. The original article credited David Reinstein with Claude Code assistance. The updated hours analysis and wording were prepared with Codex under human direction.

Evaluator-reported hours per evaluation

Evaluators reported a median of 8 hours per evaluation and a mean of 9.6 hours, across 33 evaluation records with a usable hours value. The middle half took 4–12 hours; the full range was 3–32 hours.

Measure Hours
Median 8
Mean 9.6
25th percentile 4
75th percentile 12
Minimum 3
Maximum 32

The public export contains 215 evaluation records: 33 (15.3%) have a usable hours value and 182 do not. These statistics describe the records with reported hours. They may not represent evaluations without reported hours, and multiple records may come from the same evaluator.

Public export checked 2026-11-02. This is the date of this calculation, not a claim that the underlying survey responses are recent.

What the figures measure

Evaluators report the time they spent on an evaluation. The published numeric field, hours_spent_manual_impute, includes manual conversion or imputation of reported answers. These figures are estimates of working time, rather than elapsed days from invitation to publication.

The unit is an evaluation record, rather than a paper or a unique evaluator. The approved export removes paper and evaluator identifiers, so it supports aggregate time statistics without linking hours to individuals. Missing hours are excluded from the calculations, rather than counted as zero.

Data, calculation, and dashboard

The interactive dashboard presents self-reported time under Statistics → Time Spent on Evaluations, alongside the mean, median, quartiles, range, and a histogram.

The table above uses the approved public hours export. Its calculation script verifies that the file matches the public export manifest, excludes missing values, and computes the quartiles by linear interpolation (equivalent to R’s default quantile method). Run it from the repository with:

python3 code/summarize_evaluation_hours.py --as-of YYYY-MM-DD

Browse other data posts.