a specific value assigned (as opposed to estimated) by a researcher when testing a statistical model. For example, in structural equation modeling, researchers may use a fixed parameter of 1.0 for one of the factor loadings or variances of each latent variable in a model. Compare free parameter.