nulhypotese

The null hypothesis, or nulhypotese in several languages, is a default statement in statistical hypothesis testing about a population parameter. It typically asserts that there is no effect, no difference, or no association. Common forms include H0: μ = μ0 or H0: p = p0. The alternative hypothesis, Ha or H1, expresses what the researcher seeks to demonstrate, such as μ ≠ μ0, μ > μ0, or p ≠ p0. The null hypothesis serves as the baseline against which evidence from data is evaluated.

In a formal test, researchers select a significance level, usually denoted alpha (often 0.05). A test statistic is computed from the sample data, and a p-value is obtained—the probability, assuming H0 is true, of observing data as extreme or more extreme than what was observed. If the p-value is at most alpha, H0 is rejected in favor of Ha; otherwise, H0 is not rejected. It is important to distinguish that a failure to reject H0 does not prove it true. Errors are possible: a Type I error occurs when a true H0 is incorrectly rejected, while a Type II error happens when a false H0 is not rejected.

Interpreting results requires context. A small p-value indicates data are inconsistent with H0 under the assumed model, but does not establish the truth of H0. Limitations of null hypothesis significance testing include reliance on model assumptions and frequent misinterpretations of p-values. Alternatives and complements include confidence intervals, Bayesian methods, equivalence testing, and estimation-focused approaches that emphasize effect sizes and uncertainty. The concept of a nulhypotese remains central across scientific disciplines as a framework for evaluating evidence.