| Abstract: | Generalizability theory and analysis of variance methods are employed, together with the concept of objective time pressure, to estimate response time distributions and the degree of time pressure in timed tests. By estimating response time variance components due to person, item, and their interaction, and fixed effects due to item types and examinee time pressure, one can predict the distribution (mean and variance) of total response time for a population of examinees and a particular time limit. Furthermore, these variance components and fixed effects can be used in a simulation approach to estimate the distributions of time pressure during the test to help test developers evaluate the appropriateness of specific time limits. I present theoretical considerations and empirical results from two tests. |
