Skip to content

Mysteries Department

Sign in Request access

Methods Epidemiology / statisticsCluster base-rate test: listed deaths against expected suicides in the source workforce

Method · epidemiology / statistics

Cluster base-rate test: listed deaths against expected suicides in the source workforce

To tell whether a reported cluster of deaths in an occupation is larger than chance before looking for a common cause.

Epidemiology / statisticsVersion 1.0Licence MIT1 runUsed on 1 mystery

Questions it answers

Were there more deaths among British defence scientists and engineers in 1982–1990 than ordinary death rates would produce?

Code Python

Download module (.py)base_rate.pyView source

Runs and results 1 run

Run	Mystery	Measured	Result	Against	Date
RUN0DGGDGRHYP78	Marconi	Observed/expected ratio with exact Poisson 95% interval	0.25	1 (as many listed deaths as ordinary rates predict)

Applied to

Marconi scientists' deaths, 1982–1990

1 datasetThe Marconi deaths case table, 1982–1990

References

Download module

Category: Epidemiology / statistics
Version: 1.0 · 3 Nov 2026
Licence: MIT
Language: Python
Inputs: A case table with each death's date, employer group and the kind of death the list presents; age-specific suicide rates (ONS definition, including undetermined intent) by year and sex; workforce sizes from official sources.
Output: For each workforce: expected count, observed count, observed/expected ratio with exact Poisson 95% interval, one-sided Poisson probabilities, and the workforce size at which the observed count would equal the expected one.
Cite: BibTeX · APA

Develop this method

Methods are versioned like code. Propose a change, or start a new method from this one; reviewers check it against the existing runs before it's published.

Start a new method from this Propose a change Run on a dataset

Authors

apolloWrote version 1.0