Study design, program evaluation, statistical analysis, AI evaluation, and working with statisticians.
Ask two public benchmarks how often an AI model hallucinates and you can get two very different answers for the same model. They are not measuring the same thing. A case study in why a metric's name is not its definition.
A hospital AI sepsis alert was followed by a 17% drop in sepsis deaths. That sounds enormous until you ask what it is 17% of. The same result is 1.9 percentage points, and the study was a before-and-after comparison. A case study in relative versus absolute risk.
This week's AI-evaluation research asked what a good score actually measures: tuning small judges toward human labels, safety scores that hide fragility, a judge checked against curated data, and an interval that leaves out model uncertainty.
Write a clinical trial statistical analysis plan that specifies the estimand, primary model, missing-data rules, safety analyses, and sign-off.
A step-by-step guide to reporting a multiple regression in APA 7: what to say in the text, what goes in the coefficient table, how to format the numbers, and which assumption checks to mention. With a worked example and a free formatter.
The chi-square test of independence and the chi-square test of homogeneity use the same table, the same formula and the same p-value. They differ in how the data were collected and what you can conclude. How to tell which one you ran, and how to word and report it.
{{ p.excerpt }}
{{ noResultsMsg }}
One email whenever we publish something new. No spam, unsubscribe anytime.
{{ subscribeErrorMsg }}