Limitations in testing for attitudes toward cancer.

A Cancer Attitude And Knowledge Questionnaire was administered to two groups of incoming first year medical students and to one of these groups at the end of their second year. The same questionnaire, except for the questions dealing with cognitive knowledge, was also used to assess the attitudes of practicing oncologists. There was a wide variation in the responses to the attitudinal items in all groups. In view of many inconsistencies in the results, a rigorous analysis of the instrument was carried out. The subscales derived from the Cancer Attitude Survey were found to lack adequate reliability. The only part of the attitudinal test found to be reliable and discriminatory was the semantic differential developed by the authors. It is, therefore, concluded that the currently used techniques are inadequate, and there is a need for a new multidimensional approach to the assessment of attitudes toward cancer.