a statistical procedure to determine whether means for an outcome or dependent variable differ across a specific set of conditions (i.e., levels of the independent variable) that have been selected by a researcher. This is contrasted with a random-effects analysis of variance, in which the conditions have been randomly selected from a wide range of possible choices before any mean differences are examined.