Judging drugs: patients' conceptions of therapeutic efficacy in the treatment of arthritis.

Abstract

a large number of items, a minimal ranking system was used. Ten cards, each with the name of one job factor, were laid out on a table. Each respondent was asked to select the first, and then the second, most important factor and to hand the interviewer the corresponding cards. From the remaining cards, the respondent chose the next two most important factors, as well as the two following those. The remaining four factors were not ranked at all. Subsequent rank-order calculations were based on assigning a rank of 1 for a first choice, 2 for a second choice, 3.5 for a third and fourth choice, 5.5 for a fifth and sixth choice, and 8.5 for an unranked item. Table 1 shows the mean ranks attained by the job factors using this procedure. ' Each respondent was shown a picture of a six-rung ladder and asked to indicate extent of agreement with each statement by reference to how high up the ladder, starting from ground level, their agreement went. I found during the pretest that this visual, rather than numerical or verbal scale, was far easier to administer. This item was administered and scored using the procedure described in note 3 for the item dealing with coworker behavior.