PAPER-PENCIL LABORATORY TESTS AND THEIR RELATIONSHIP TO VARIOUS ACHIEVEMENT MEASURES IN PHYSICS
Abstract
Characteristics of the tests were investigated, and relationships with other achievement measures were explored. The paper-pencil tests to measure specific instructional objectives had a normal distribution over a much narrower range of scores than the laboratory performance tests. The Davis discrimination indices for the items were highest with the total score on each test as a criterion, but the average index was low for every test. The point biserial correlations with the performance items tended to cluster about zero. In general, the Hoyt reliabilities were on the low side and were comparable in value to the reliabilities of the performance tests. The items had fairly high face validities as based on the average of 4 expert ratings. With the exception of the identification and function, pre-test intercorrelations were low. These 2 pre-tests were fair predictors for the corresponding post-tests in mechanics. All the pre-tests in electricity were fair predictors for the corresponding post-tests. The correlations between the paper-pencil tests and other criteria of achievement in elementary physics were either nonsignificantly different from zero, to low. On 3 of the 4 tests in mechanics, boys scored significantly higher than the girls. On all the post-tests the means were significantly higher than for the pre-tests, irrespective of sex or course taken. The investigation lends support to the hypothesis that paper-pencil laboratory tests contain elements other than those evaluated by performance tests and the conventional measures of achievement in elementary physics.(See also AD19185)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1953
- Accession Number
- AD0022612
Entities
People
- Haym Kruglak
Organizations
- University of Minnesota