Measuring attitudes by error-choice; an indirect method.