Four Measures of Topical Knowledge: A Study of Construct Validity. Technical Report No. 501

A study was conducted to establish the construct validity of measures designed to assess topical knowledge. Thirty-one ninth-grade students and 44 third/fourth-grade students were interviewed to ascertain their knowledge about 4 topics; 2 weeks later they responded to 3 tests of topical knowledge. Results were analyzed quantitatively and qualitatively. Correlations between the interview scores and the paper-and-pencil measures did not reveal a strong and clear relationship between the students' performance on the interviews and their performance on the tests of topical knowledge. Conditional probability analysis revealed that the interview and the paper-and-pencil tests provided different information regarding an individual's knowledge of a topic. These differences seem to be related to the difference between a recall and a recognition task. Results indicated that the interview approach captured individual differences more clearly and more dramatically than the paper-and-pencil tests of topical knowledge. (Fourteen tables of data and four figures are included. Forty-eight references are attached.) (MG)