Wald inference tools

What question are you trying to answer?

Choose by inferential question. Observed-data tools evaluate candidate effects against a reported estimate and confidence interval; design tools evaluate repeated-study behavior under assumed true effects.

All calculations happen in the linked client-side apps. This catalog performs no statistical calculation and sends no user values anywhere.

“Software validated” means numerical and release conformance to the declared formulas and tests. It does not mean clinical validation or suitability for clinical decision support. See the portfolio validation report.

Choose a path

Match the tool to the question

Show tools All Observed data Design

Loading the checked-in tool manifest…

JavaScript is required only to render the checked-in catalog manifest. No calculations or user data are involved. The same metadata is available in data/tools.json.

Keep the conditioning straight

Tool comparison

Tool	Question	Conditions on	X-axis means	Core outputs	Assumed truth?	Selection rule?	Primary limitation

One distinction prevents many errors

Observed reconstruction is not assumed truth

Observed-data conditioning

Compatibility and relative-likelihood displays ask how candidate effect values relate to one reported estimate and confidence interval under a Wald reconstruction.

Design conditioning

Detectability, Type S/M, and precision planning ask what repeated future statistics would do if specified effects were true under a chosen claim rule.