Skattningen

Skattningen is the Swedish term for estimation, the process of inferring an unknown quantity from observed data. In statistics, skattningen aims to determine a parameter of a population or a model based on a sample, producing a numerical value or a range that reflects what is plausible given the data and assumptions.

There are two broad categories: point estimation and interval estimation. Point estimation yields a single best guess for the parameter, such as the sample mean estimating the population mean. Interval estimation provides a range, like a confidence interval or a Bayesian credible interval, that quantifies uncertainty about the parameter.

Estimators are functions of the data used to compute estimates. Desirable properties include unbiasedness (the average of the estimator equals the true parameter), efficiency (lowest variance among unbiased estimators), and consistency (estimates converge to the true value as sample size grows). Other considerations include robustness and sufficiency.

Common estimation methods include maximum likelihood estimation, which selects parameters that maximize the likelihood of the observed data, and the method of moments, which aligns sample moments with population moments. Bayesian estimation uses prior information along with the data to form a posterior distribution for the parameter.

Estimation is central to many fields, including science, engineering, economics and policy. Typical tasks include estimating a population mean or variance, regression coefficients, or distributional characteristics. Practical challenges include data quality, measurement error, sampling bias, and model misspecification, all of which affect bias and precision. Uncertainty about the estimate is communicated via intervals or posterior distributions.

Skattningen can also refer to non-statistical assessments, such as tax assessments or asset valuations, but in a statistical context it denotes the inference process described above.