Abstract:
|
In a 3-group, gender-matched, ability controlled, randomized experiment, we evaluated the effect of learner control on performance in an online statistics class for healthcare executives. We advised students to follow a weekly outline, but their actual control over learning pace and practice distribution varied: Groups 1, 2, and 3 received weekly, monthly, and course-end deadlines. A pretest controlled for prior knowledge. Univariate analysis of variance showed differences in mean performance, F(3,48) = 6.35, p< .01, Eta^2 = .284, group means: 98% Group 1, 89% Group 2, and 76% Group 3. Group effect disappeared when we included external locus of control and performance goal in a logistic regression to predict class failure, -2LL = 37.60, p< .01, Nagelkerke R^2 = .40. Less learner control over pace improves performance, but this effect diminishes when psychological factors are understood.
|