CALCULATE WITH CONFIDENCE
Editorial category
Learn how to summarize data, reason about probability, quantify uncertainty, and interpret statistical results without overstating what the evidence proves.
Descriptive statistics
Build a truthful first description of numerical data by combining location, variability, shape, and observation-level context.
Descriptive statistics
Choose a measure of center from the variable, distribution, and decision instead of defaulting automatically to the arithmetic mean.
Descriptive statistics
Understand deviations, squared distance, sample and population denominators, and why standard deviation retains the data’s unit.
Descriptive statistics
Read relative position and middle-half spread while keeping quantile conventions, interpolation, and reference populations explicit.
Probability
Understand why learning that one event occurred changes the denominator and how to read intersections, tables, and tree branches consistently.
Probability
Separate two probability ideas that sound similar but imply opposite overlap behavior for nonzero events.
Probability
Build, multiply, and recombine sequential branches while keeping conditional labels and complete probability mass visible.
Probability
Understand why prior prevalence and false positives can dominate posterior probability even when evidence appears accurate.
Probability
Interpret probability-weighted averages alongside variability, downside outcomes, time horizon, and repeated-decision assumptions.
Probability
See how returning or retaining sampled items changes the population, dependence, path probabilities, and appropriate distribution model.
Probability distributions
Choose a candidate model from the process, support, dependence, and parameter meaning—and then check fit instead of matching by shape alone.
Probability distributions
Connect location, scale, standardized distance, density, cumulative probability, and model assumptions without treating normality as automatic.
Probability distributions
Choose among three common count models by examining trials, exposure, replacement, finite populations, and event-probability assumptions.
Probability distributions
See how three degrees-of-freedom families arise from normal-model estimation and why their shapes and supports serve different inferential roles.
Probability distributions
Compare positive time-to-event models through process assumptions, survival shape, hazard behavior, and parameter meaning.
Probability distributions
Understand why ordinary averages and normal tails can fail for extreme processes and why tail modeling demands carefully defined data and thresholds.
Sampling and confidence
Separate variation among observations from variation among sample estimates and understand what the central limit theorem does—and does not—supply.
Sampling and confidence
Interpret confidence as long-run procedure coverage rather than a probability assigned to one fixed parameter after the interval is calculated.
Sampling and confidence
Plan precision by connecting confidence, variability, design, finite populations, and rounding without treating one formula as a universal sample plan.
Sampling and confidence
Compare interval behavior near boundaries and understand why the simple Wald interval can misrepresent uncertainty for small samples or extreme proportions.
Sampling and confidence
Understand why survey and grouped designs change estimator construction, effective information, and variance beyond ordinary simple-random formulas.
Sampling and confidence
Separate compatibility with a null model from effect magnitude, uncertainty, costs, benefits, and domain-specific consequences.
Hypothesis testing
Connect evidence, decision thresholds, long-run error rates, effect assumptions, and sample size without turning a p-value into the probability a hypothesis is true.
Hypothesis testing
Choose a candidate method from the estimand, outcome scale, design, pairing, group count, assumptions, and reporting goal.
Hypothesis testing
Match analysis to the observation relationship and understand why unequal-variance Welch inference is a strong default for independent means.
Hypothesis testing
Choose a categorical comparison from table structure, pairing, fixed margins, sparsity, and the exact null question.
Hypothesis testing
Choose among one-way, repeated-measures, factorial, rank-based, and follow-up procedures from the design and target comparison.
Hypothesis testing
Interpret association, fitted prediction, uncertainty, nonlinearity, outliers, confounding, and study design without turning a coefficient into a causal claim.
Descriptive statistics
Connect the box, median, whiskers, and flagged observations to the raw data without inventing density or sample-size information.
Descriptive statistics
Use bins and numerical shape summaries together without mistaking display choices for stable properties of the population.
