The Effects of Self-Imposed and Externally Imposed Rewards of Children's Intrinsic Motivation

.... ACKNOWLEDGEMENTS. .PREFACE LIST OF TABLES. LIST OF FIGURES Introduction . . . Overview of the study Hypotheses and predictions. Method. Results Discussion. References. Appendix A: Literature review. Appendix B: Pilot studies. . • . Appendix C: Analysis of variance summary tables. Appendix D: Post-experimental questionnaire. Appendix E: Raw data . Appendi x F: Bibliography


Results
Discussion. References.
Appendix A: Literature review.
Appendix B: Pilot studies.    Staw, 1975;Deci, 1971Deci, , 1972Lepper, Greene, & Nisbett, 1973;Lepper & Greene, 1975;Ross, 1975 For most educators, an important goal of schooling is for the student to become exposed to an interested in the subject matter such that the student will continue to develo~ skills and expand knowledge when outside the immediate confines of the school building.
Educators strive to develop a love of learning in general, and positive feelings and behavior towards as much of the specific subject matter as possible.
That is, tnere appears to be a concern to maintain or increase a student's intrinsic motivation (i.~., motivation to perform an activity for its own sake, rather than for extrinsic conse- Following Skinner's (1953) conceptualization, the notion of self-reinforcement has been operationalized in various ways and has been investigated as a process and as a dimension of personality 3 with both children and adults (Bandura, 1971;Masters & Mokros, 1974).
Self-reward procedures as behavior change tools are gaining in popularity (Thoresen & Mahoney, 1974)     F (1, 84) = .22. Thus, the assumption that the music task was of higher interest than the voice task was supported for boys but not for girls, for whom the voice scores in the control condition were nonsignificantly higher than the music scores.

Self-reward scores
The means and standard deviations for the number of pe_ nnies taken for each self-reward trial are presented in Table 2. An F test conmax Insert            Table 9.

64
Finally, the temptation paradigm has used self-reward procedures to investigate cheating behavior. Thus, for example, Kanfer (1966) found that younger children and children who were ranked lower in the class by their teacher, took more "undeserved" self-rewards. In this experiment, children were required to guess a n. umber from one to one hundred and to self-reward with points if they believed their guess to be correct. Since the probability of a correct guess is near zero, any self-reward was defined as undeserved.
It is, therefore, important to distinguish between self-reward and self-punishment as the two components of self-reinforcement.
Observation sessions have either been built into the treatment free-time session (Deci, 1971(Deci, , 1972Ross, 1975) or have occurred several days after treatment (Lepper et al, 1973;Lepper & Greene, 1975 Mean number of seconds spent by subjects on puzzles in the 8 minute free choice period (Deci, 1972). free choice period (Deci, 1972). In regard to reward timing (the only significant factor), Deci employed inequity theory (Adams, 1963(Adams, , 1965  Children were asked to rank order tasks both within and between the categories. The format for al1. rankings was "The best is" and "The worst is". The middle choice is, of course, inferred.

Procedure.
The investigator spent about half an hour in each of the two classrooms. The within category rankings occurred first.
Rartking sheets were distributed t4 the class and their use was explained.
The three music selections were played and students rank ordered their preferences. The same procedure was followed for the voice selection. Magazines were ranked after being visually displayed for approximately one minute each. Children were then asked to complete the betw . een category rankings.
Results. The percentages of children ranking each activity as either best or worst are summarized in Table 12.
Insert Table 12 about here Iri the between category rankings music and magazines were the preferred activities, while listening to voice was ranked worst by a large propoition of the children.
In the music category, none of the selections ~eceived a gre~t deal of support. Children appeared to prefer the sound of conventional rock, but more than a third thought this selection was the worst. The story about vegetables was a good deal more preferred than either spelling or the 2x table, and The music, voice and magazine were presented to the children, after which they were asked to pick the best and worst catego~{es and to tate the music and voice as Very Good, G6od, Fair, Bad, or Very Bad. Table 13 shows the proportion of children selecting each category as either best or worst. The percentage of children rating music and voice from very good to very bad is presented in Table 14.
Insert Tables 13 and 14 about here As predicted, the children almost unanimously preferred the new music selection, while slightly fewer children ranked the 2x table as worst.
Within category ratings showed that 85% of the group rated music as either very good or good and 75% rated the voice selection as either bad or very bad.
Based on the above data it was decided to make Joy to the World and four similar tunes from the same album the tentative music selections and the 2x table the voice selection. Based on the opinions of U. S. News and World Report, the general category of newsmagazine was to be used for the distractor in the main study.
In summary, the two pilot studies enabled task cho . ices to be made that were based on questionnaire data rather than on merely the intuitions of the experimenter. There is empirical support for the contention that the particular music selections are more interesting to the children of this age group than the particular voice selection.

Best
Worst