Undermining learning: the impact of rewards on learning behavior
Date
2016
Authors
Wehe, Hillary, author
Seger, Carol, advisor
Rhodes, Matthew, committee member
Conner, Bradley, committee member
Hoke, Kim, committee member
Journal Title
Journal ISSN
Volume Title
Abstract
The undermining effect suggests that external rewards can decrease levels of internal motivation. Research exploring student motivation shows that internally motivated students appear to engage longer and in more challenging tasks compared to students focused on external rewards or performance feedback. The current study tested variables that may decrease susceptibility to motivational undermining for learning behaviors. In all studies, students were assigned to either a reward or non-reward condition and completed a word-learning task followed by a final test. Subjects were given the option to choose to re-study the words at two times during the task—pre- (while reward is still achievable) and post-test (after reward is given and no further extrinsic reward is achievable). Across all studies, an undermining effect was expected: Non-reward subjects would spend a greater amount of time reviewing the words during the post-test interval compared to the reward group. Study 1 directly tested the hypothesis by observing whether or not the reward groups behaved differentially at the pre- and post-test choice. Reward subjects spent significantly less time engaging in the task during the post-test review phase, supporting the presence of the undermining effect (t (1,60)=2.06, p = .02, 1-tailed) but a 2 (group: reward x non-reward) x 2 (study time: pre-test x post-test) repeated measures ANOVA comparing the mean study times for the reward and non-reward subjects’ pre-test study and post-test review time revealed that the interaction between group and study time did not reach significance (F (1,60) = 3.52, p = .065). Study 2 was identical to the first study but with the addition of a surprise, 24-hour delayed memory test to examine whether the extra post-test study had beneficial effects on long-term retrieval. Non-reward subjects were hypothesized to recall more items on a delayed memory test compared to reward subjects due to increased study time. A 2 (group: reward x non-reward) x 2 (study time: pre-test x post-test) repeated measures ANOVA was conducted to compare the mean study times for the reward and non-reward subjects’ pre-test study and post-test review times. The interaction between group and time spent on task was significant (F (1,241) = 4.24, p < .05) but there was not a significant main effect for the between subjects variable of reward on the amount of time spent engaging in the task during the pre- and post-test phases (F (1)= .63, p = .44). A 2 (group: reward vs. non-reward group) x 2 (test performance: immediate x delayed) repeated measures ANOVA was conducted to compare the average accuracy between groups on the delayed memory test. There was not a main effect of group on performance (F (1, 110) = .82, p = .38) and the interaction between reward group and immediate or delayed test was not significant (F (1,156) = .201, p = .65). Study 3 was similar to the first study but subjects were allowed to choose the material they were learning (i.e., Swahili or Lithuanian words). The element of choice was expected to increase the degree of control and internal motivation students experienced and consequently decrease the effect of undermining between the reward and non-reward group. Specifically, study times between the reward and non-reward group were hypothesized to be equal between groups and higher than then a forced choice condition. A 2 (group: reward x non-reward) x 2 (choice: self-determined x forced-choice) x 2 (study time: pre-test x post-test) x 2 (language: Swahili x Lithuanian) repeated measures ANOVA was conducted. The main effect of choice condition was not significant (F (1,60) = .140, p = .71). The main effect of reward was also not significant (F (1,60) = .920, p = .34) but the interaction between choice and reward on time spent on task was significant (F (1,60) =4.11, p < .05). A 2 (group: reward x non-reward) x 2 (choice: self-determined x forced-choice) repeated measures ANOVA was conducted to compare performance on an immediate memory test for the self-determined and forced choice group but the effect was non-significant (F (1,60) = .67, p = .16); in addition, there was not a significant main effect of reward (p =.32) nor was there an interaction (p = .16).