When measures are taken on the same individual over time, it is difficult to determine whether observed differences are the result of changes in the person or changes in other facets of the measurement situation (e.g. interpretation of items or use of rating scale). This paper describes a method for disentangling changes in persons from changes in Likert-type questionnaire items and rating scales. The procedure relies on anchoring strategies in Rasch measurement to create a common frame of reference for interpreting measures that are taken at different times. The Rasch Rating Scale Model is an additive linear model that describes the probability that a specific person will respond to a specific Likert-type item with a specific rating scale step. How to perform the procedure is illustrated using data from 168 teachers responding to a questionnaire about barriers to the implementation of a portfolio assessment. The five steps used to measure change over time are summarized as: (1) evaluate scale and item invariance; (2) correct the scale calibrations; (3) benchmark the time-1 estimates; (4) correct the time-2 person measures; and (5) correct the time-2 item calibrations. Applications of the procedure reveal two changes in teachers that were not apparent when the correction was not applied to the analyses. Five appendixes present the FACETS command files for each step. (Contains 2 figures, 5 tables, and 15 references.) (SLD)