Abstract
The standard two-wave multiple-indicator model (2WMIM) commonly used to analyze test-retest data provides information at both the group and item level. Furthermore, when applied to binary and graded item responses, it is related to well-known item response theory (IRT) models. In this article the IRT-2WMIM relations are used to obtain additional information for each individual respondent. Procedures are proposed for (a) obtaining individual estimates of trait levels and amount of change, and (b) assessing whether the main assumptions on which these estimates are based are met. The procedures are organized in a comprehensive approach that can be used with binary, graded, and continuous responses. The relevance of the proposal is discussed and guidelines are given on how to use the approach in applied research. Finally, the approach is illustrated with an empirical data set. It worked well and provided meaningful information.