frequentistinen

Frequentistinen (frequentist) statistics is an approach to statistics in which probability is understood as the long-run relative frequency of events in repeated sampling. Inference treats model parameters as fixed quantities, while randomness is attributed to the observed data. The central idea is to study the properties of procedures—such as estimators or tests—over hypothetical repetitions rather than to assign probabilities to specific parameter values.

Core concepts include sampling distributions, estimators with properties like unbiasedness, consistency, and efficiency, and procedures with proven long-run performance. Confidence intervals are designed to contain the true parameter a stated proportion of the time in repeated sampling. Hypothesis testing uses null and alternative hypotheses, significance levels, p-values, and decisions that control error rates in the long run. The Neyman–Pearson framework provides rules for choosing tests to maximize power at fixed error levels, while maximum likelihood estimation is a common method for deriving point estimates based on observed data.

Historical strands include Fisher’s emphasis on p-values and significance tests and the Neyman–Pearson approach to decision rules. Bayesian statistics, by contrast, interprets probability as a degree of belief and incorporates prior information, whereas frequentists rely on sampling properties without prior distributions.

Critiques of the frequentist paradigm concern the misinterpretation of p-values, dependence on large-sample approximations, and sensitivity to model misspecification. Nonetheless, frequentist methods remain foundational in many scientific disciplines, informing experimental design, estimation, and hypothesis testing through guarantees about long-run performance rather than beliefs about particular parameter values.