065 again there was no significant interaction effect

This preview shows page 11 - 12 out of 16 pages.

.065. Again, there was no significant interaction effect,F(l,65) = 0.07.Finally, we also examined the selected level of difficultyfor each game. A one-way ANOVA revealed that there weresignificant differences among the groups, F(l, 65) = 3.67,MSE= 0.93,p <.01. Specifically, students in the no-fantasy control condition chose the less challenging versionof the activity significantly more often than those studentsin the fantasy conditions, F(l, 65) = 7.36,p <.01.For thismeasure, however, the main effects for personalization andchoice were not significant, F(l, 65) = 1.60,p >.21, andF(l,65) =2.71,p >.10, respectively.In general, then, participants in the contextualization,personalization, and choice conditions exhibited signifi-cantly higher levels of task involvement, as evidenced bytheir preference for more challenging versions of the game,their greater use of complex operations, and their emphasison strategic play relative to participants in the no-fantasy,nonpersonalized, and no-choice conditions.Effects on LearningGiven these differences in the motivational appeal of thevarious versions of the educational program and in thedegree of task-involvement, it is of central importance toask how students' learning was affected by our experimen-tal manipulations. Thus, we examined the students' perfor-mance on the mathematics test that had been administeredduring the posttest session.To examine the effects of personalization and choice onstudents' learning, we conducted a one-way analysis ofcovariance (ANCOVA), using the pretest score as a covari-ate,on children's scores on the 20-item math test that hadbeen administered during the posttest sessions 1 week later,in the absence of the computer. This analysis yielded asignificant effect of experimental treatments on the learningmeasure, F(4, 65) = 13.80,MSE= 0.86,p <.0001.Therelevant means and standard deviations for this measure arepresented in Table 3.Once again, a series of orthogonal contrasts was per-formed to test the specific predictions of the study. First,students for whom the educational activity had been pre-sented in an unembellished, no-fantasy context exhibitedsignificantly lower levels of learning than students forwhom the activity had been embedded in a fantasy context,?(65) = 5.00,p <.0001.Second, children in the personal-ized fantasy conditions showed significantly higher levelsof learning than those in the generic fantasy conditions,f(65) = 4.89,p <.0001.Last, the provision of some degreeof choice over various instructionally irrelevant aspects ofthe game also had beneficial effects on students' learning, aschildren in the choice conditions scored significantly higheron the math test than those in the no-choice conditions,t(65) =2.02,p <.05. Again, no interaction effect wasfound, r(65) = 1.11,p> .27.Effects on Perceived Competence andLevel of AspirationFinally, we examined the possibility that there might besubsequent motivational effects due to our experimentalmanipulations. To this end, we examined the students'responses to two additional measures. The first of thesemeasures focused on children's perceptions of their ability

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture