Interactive Methods Calculators

How strong must unmeasured confounding be to explain away the association?

Number needed to treat or harm, from risks or from a relative measure.

Overlap-adjusted medication adherence over an observation window.

Bayes' rule in odds form: convert pre-test probability with LR+ and LR−.

Normal approximation (Fleiss). Label as an approximation; verify with exact methods for submissions.

Converts an odds ratio to a risk ratio when the outcome is common. Warns when baseline risk is large.