UBC Theses and Dissertations
The evaluation of a less structured form of interest test item Smith, Robin Nelson
The investigation might be summed up as follows. A new type of item for use in interest inventories was conceived to be one stated in relatively general terms. It was felt that the use of this type of item, coupled with assistance to the subject in making the comparisons by the use of the paired-comparison form might result in increased reliability and counselling validity in interest inventories. Parallel tests composed of matched items were administered, under as nearly as possible identical conditions, to the same group of approximately eighty high school students, and the results compared. The first of these tests was composed of less structured items matched to certain items of the second test, "The Kuder Preference Record" form BB, as exemplifying the use of the usual specific or structured item. There appeared to be no significant difference in the reliabilities of the two tests. Since there were approximately seven times as many items in the Kuder as there were in the experimental test it was concluded that the evidence warranted giving credit for superior reliability to the new type, or less structured, item. Very significant coefficients of correlation were obtained between the total scores for the corresponding areas of the two tests. This appeared to provide substantial evidence that the two types of item were measuring essentially the same thing, at least in major respects. A test of internal consistency was devised. The items of the experimental test and the matching items of the Kuder were compared on this basis. The less structured items appeared to be manifestly superior in regard to this property. An effort was made to indicate the extent to which each test distinguished clearly between the subjects most preferred area and least preferred area. Here also the experimental test and hence the less structured item appeared to demonstrate a distinct superiority. In the last section of the report various incidental observations and impressions were reported as such. In conclusion it seems safe to say that the new type of item shows considerable promise for use in interest inventories, and that verification and further investigation would be well warranted.
Item Citations and Data