Logo Utrecht University

Informative Hypotheses


Welcome on the informative hypotheses website.

Informative Hypotheses are hypotheses constructed using (in)equality constraints among the parameters of interest. A simple example is three ordered means, Hi: μ1 > μ2 > μ3. Often informative hypotheses paint a more realistic picture of the state of affairs in a population expected by a researcher than the traditional null and alternative hypotheses.

For more information, we refer to several published papers on informative hypotheses that you can find on this website. A distinction is made between papers introducing methodological new developments, more didactical/tutorial papers and applied papers where informative hypothesis methods are used for the primary data analysis. Furthermore, we provide several software packages with which informative hypotheses can be evaluated. Most current are Bain (which uses the Bayes factor) and GORIC (which uses an AIC-type information criterion). Both packages can be found in the menu in the left hand margin of this webpage.