| Abstract: | The theory of planned behavior presents clear operational definitions of attitudes, subjective norm, perceived behavioral control, and each of their corresponding belief-based measures. Theoretically, the direct and indirect measures of a given construct must be closely correlated. Empirical results, however, indicate that this is not always the case. In the present study, 2 of the factors that could be responsible for this situation-namely, the scaling of the variables defining each belief-based construct and the adequacy of using an expectancy-value model within the belief-based measures-were verified among a data set of 16 studies concerned with the application of the theory of planned behavior to the field of health. The results indicate that the scaling method used affected the correlation coefficients between indirect and direct measures. However, the face validity of these scaling methods must be demonstrated. The results also support the idea that, in most cases, using the expectancy-value model is no better than using only one arm of the belief-based measure. |
