No abstract

This study applies theory linking students' self-perceived competence with academic achievement, using data from the Third International Mathematics and Science Study (TIMSS). Within-country data generally show a positive relationship between student achievement and self-perceived competence in mathematics and science. However, when one examines this relationship between countries, the opposite relationship occurs, i.e. countries with higher student self-evaluations usually performed poorly on the TIMSS tests and vice versa. Similarly, countries with a high proportion of students perceiving the subjects as being easy performed poorly on the TIMSS tests and vice versa. This pattern exists for both mathematics and science at grades 3, 4, 7 and 8. The authors suggest that this pattern may re ect low academic expectations and standards in low performing countries and high academic expectations and standards in high performing countries.

Ideally, test performance is unrelated to the mode in which the test is administered. This study investigated the relationships between test mode (paper and pencil vs. computerized with editorial control and computerized without editorial control) and computer familiarity (lower, moderate, and higher) with test performance on the Graduate Record Exam (GRE). The GRE was administered to 222 undergraduates stratified by gender and randomly assigned to the three test mode groups. With self-reported grade point average as a covariate in a MANCOVA, the authors found that examinees in the paper-and-pencil group outperformed the computerized-without-editorial-control group on all subtests. The computerized-with-editorial-control group outperformed the computerized-without-editorial-control group on the Analytical subtest only. The authors also found a significant main effect for computer familiarity on the Analytical and Quantitative subtests. A significant interaction between computer familiarity and test mode on the Quantitative subtest confounded the main effect for that subtest. The subtests were dramatically more speeded in the computerized forms. The results emphasize the importance of evaluating time constraints when converting exams from paper-and-pencil to computer-delivery mode.

No abstract

Teachers in Ireland (n = 170) rated students (n = 2,617) on IQ, mathematics, and English, as well as on 12 social and academic classroom behaviors. Factor analysis oflQ, mathematics, and English standardized test scores, together with the 15 teacher ratings, showed that there is overlap between ratings and test results but that the information obtained is not redundant. Three factors were identified: one was comprised primarily of the social behaviors; a second was comprised of the academic classroom behaviors and teacher ratings on IQ, mathematics, and English; and the third was comprised of the test scores in IQ, mathematics, and English together with the corresponding teacher ratings.The question under investigation, the relationship of teacher ratings to standardized tests, goes back to Binet. However, opportunity for new illumination into this area was provided by a nationwide, longitudinal, societal experiment conducted in the Republic of Ireland to assess the impact of introducing standardized testing on various populations, particularly students and teachers (Airasian, Kellaghan, & Madaus, 1971). For the present study, 170 fifth-grade teachers were asked to provide ratings for each of their students on a variety of behaviors using a 5-point rating scale. Just prior to making their ratings, teachers administered standardized tests of intelligence, Portions of this paper were presented at the annual conference of the National Council on Measurement in Education, Toronto, 1978.