1. Staging algorithms and self-categorizations are conceptually more consistent with the assumption of discrete stages than are multi-dimensional questionnaires and continuous measures such as contemplation ladders and readiness rulers. Although staging algorithms and self-categorizations may be less reliable than multi-dimensional questionnaires, they should be the method of choice for researchers and practitioners who test stage theories.
2. Researchers who plan to use the TTM may wish to use the algorithms developed by the Rhode Island group for comparability with previous research. However, they should also consider including alternative stage algorithms that avoid some of the problems identified with the standard measures.
3. In developing new stage algorithms or self-categorizations, pre-action stages should not be defined in terms of past behavior so as to avoid logical problems. Arbitrary time periods should also be avoided. Stages should be defined in terms of events, for example making a decision to change one’s behavior or setting a quit date. This may involve modifying the original theory or developing a new theory, so that the stage measures accurately reflect the theoretical stage definitions.
4. Stage measures should incorporate a clear and precise definition of the behavior.
5. In testing predictions from stage theories (and thus at the same time testing the construct validity of stage measures), researchers should use strong research designs, in particular longitudinal prediction of stage transitions and experimental studies of matched and mismatched interventions.