| The purpose of this study was to determine whether alternative placement of discrepant events would impact affective and cognitive outcomes of ninth-grade physical science students grouped into intact classes and classified as either "high" or "low" in prior academic achievement. Although researchers have found discrepant events to be effective in terms of cognition and recall, their chronological placement within science lessons had not been empirically researched. In this study, discrepant events were presented before, during, and after specific science lessons involving thermodynamics and heat. Discrepant events were withheld from the control group.;To measure affective outcomes, the "enjoyment" and "motivation" scales taken from Sandman's (1973) Attitudes Towards Science Inventory (ATSI) were used to index subjects' global feelings about studying science, while a 20-item set of Semantic Differential (SD) scales was employed to determine their attitudes regarding the specific subject matter taught. To measure cognitive outcomes, a 20-item, selected response test was constructed by the researcher, with 6 items intended to assess subjects' knowledge of unit materials, and 14 items designed to query their understanding of unit concepts. Each subject (N = 131) was administered identical forms of each test in both pre-and post-test formats, both before and after the four-week study.;Analyzed using a 4 x 2 mixed Analysis of Variance (ANOVA) model, data pertinent to the ATSI suggested neither between- nor within-group differences in subjects' global attitudes about studying science, although data pertinent to the SD scales indicated generally improved attitudes about studying thermodynamics and heat (F (1,122) = 2.759, p < .10). On the cognitive pretests and posttests, significant two-way interactions were observed for the overall test and experimental condition (F (3,121) = 4.068, p < .01), as well as for the overall test and higher prior achievement in physical science (F (1,121) = 7.059,p < .01). As contrasted with negligible changes in the control group's scores, robust mean-difference effect sizes were observed for all three treatment groups---"beginning" (d = 1.24), "during" (d = 0.70), and "after" ( d = 0.78)---but particularly for the "beginning" group. Subsequent analysis revealed that the apparent advantage of the "beginning" group was largely attributable to a particularly strong showing on the six test items concerned with knowledge (d = 2.06). |
