01
P — Probability
Think in likelihoods, not certainties. Understand probability as belief calibrated by evidence. Differentiate base rates from conditional probabilities.
BayesianUncertainty bands
02
R — Risk
A decision is incomplete without understanding what can go wrong. Think in expected loss, worst-case loss, and tail risk. Severity × likelihood × reversibility.
DownsideTail risk
03
I — Inference
Disciplined interpretation — connect research questions to data patterns. Distinguish association from causation. Apply the correct inferential framework.
CausationStatistical inference
04
S — Significance
Separate signal from noise. Effect size, confidence intervals, power, and practical significance — not just p-values. Context shapes what "significant" means.
Effect sizep-values
05
M — Modeling
Choose the right model for the right question. Understand assumptions, validate them, and explain what the model is — and what it is not — doing.
Model selectionAssumptions