Uncertainty is the lack of certainty about a value, outcome, or conclusion. Probability measures it by considering possible outcomes and how strongly each is supported, rather than forcing one deterministic answer. section 3.0
Data rarely captures every variable affecting reality. Probability distributions make that complexity manageable by representing multiple outcomes and degrees of confidence, encouraging calculated approximation instead of overconfident precision. section 3.0
Statistical practice expresses uncertainty in several connected ways:
- Samples estimate population characteristics, but representative samples support ranges rather than absolute certainty. section 5.0
- Confidence intervals combine an estimate with a range of plausible values, such as a 95% interval for mean walking speed. section 5.0
- Statistical analysis quantifies uncertainty and explainability, while machine learning often emphasizes predictive performance and generalization. section 1.5
- Uncertainty-aware reasoning helps audit claims for cherry-picking, bias, ignored incentives, and unsupported confidence. section 1.8
A range beats a single guess
Instead of reporting that average walking speed is exactly 1.31 meters per second, report that a 95% confidence interval places the population mean between 1.27 and 1.35 meters per second. The interval communicates both the estimate and its uncertainty. section 5.0
Use the figure to visualize how sample means and confidence intervals turn uncertain estimates into interpretable ranges. section 5.0
