Testing a Model of Mammography Intention1


  • 1

    The authors would like to thank Robin Bechhofer and Gregg Pfeifer, for their assistance in the preparation of the data, and Caroline Clark, Project Manager, for her contributions in the development of this project. This research was supported by an American Cancer Society Psychosocial and Behavioral Research Grant PBR-51, Dr. Richard R. Love, P. I.

Correspondence concerning this article should be addressed to Dr. Linda J. Baumann, 600 Highland Avenue, K6/364, Madison, WI 53792.


This paper presents the results of a study testing components of the Triandis model of choice to predict intention to obtain screening mammography in the next 2 years. A questionnaire was developed to measure the theoretical components of this model. Analysis was based on 2,521 women, 50 to 65 years old, who received primary care from 1 of 28 participating primary care group practices. Based on the four theoretical components of intention described in the Triandis model-social influence, affect, consequence, and barriers-a set of structural equation models was explored. The sample was split into two equal subsamples: one sample used for model exploration and one used for model confirmation. Results showed that the original intention model was not supported. Instead, the model that best fit the data, and upheld it in the subsample used for model confirmation, was a model in which habit and intention were strongly related. The strongest predictors of habit were barriers and consequences; the strongest predictors of intention were social influences and barriers.