Interaction Between Locus of Control and Three Pacing Procedures in a Personalized System of Instruction Course Robert A. Reiser
Robert A. Reiser is associate professor and research associate with the Center for Educational Technology at The Florida State University, Tallahassee, FL 32306. The author thanks Richard L. Tate for his helpful comments on earlier drafts of this article. Thanks are also extended to Richard R. Lee, Frances S. Holley, Jane C. Wager, and Dee H. Andrews for their assistance in conducting the study reported here.
This study examined the interaction effects of three pacing procedures and perception of locus of control on student final examination performance in a PSI course. College students identified by the Rotter Internal-External Locus of Control Scale as expecting external reinforcement performed better under the reward and control conditions, and those identified as expecting internal reinforcement did belter under the penairy condition. Student anxiely was postulated as having been a factor affecting performance.
ECTJ, VOL. 28, NO. 3, PAGES 194-202 ISSN 0148-5806
Numerous studies have shown that the Personalized System of Instruction (Keller, 1968) is an effective instructional method. Students in Personalized System of Instruction (PSI) courses often learn more and have more positive attitudes than students in traditional lecture-discussion courses (Kulik, Kulik, & Cohen, 1979). However, the specific procedures employed in a PSI course vary and, as some studies have indicated, student performance in a PSI course can be greatly influenced by the particular variation employed (Kulik, Jaksa, & Kulik, 1978). One factor that often varies across PSI courses is the type of student pacing procedure employed. Keller has consistently advocated an approach involving few, if any, pacing contingencies (Keller, 1968, 1974). Others have used procedures in which students are rewarded for maintaining a particular pace (Cheney & Powers, 1971; Lloyd, 1971; Powers, Edwards, & Hoehle, 1973) or penalized for failing to do so (Malott & Svinicki, 1969; Miller, Weaver, & Semb, 1974; Reiser & Sullivan, 1977). Most studies that have compared various pacing procedures have found no significant difference between groups in mean final examination performance (Calhoun, 1976; Fernald, Chiseri, Lawson, Scroggs, & Riddell, 1975; Reiser & Sullivan, 1977). H o w e v e r , no studies have explored whether various pacing procedures might
INTERACTION BETWEEN LOCUS OF CONTROL AND PACING PROCEDURES
be differentially effective for various types of students. This study was designed to examine the interaction effects of various paring procedures and perception of locus of control on final examination performance in a PSI course. Perception of locus of control of reinforcement, the aptitude examined in this study, was m e a s u r e d by the Rotter Internal-External Locus of Control Scale, often referred to as the I-E scale (Rotter, 1966). The I-E scale measures individual differences in generalized expectancy for internal vs. external control of reinforcement. The higher an individual's score on the scale, the more external the individual. External individuals are those who usually perceive reinforcement as being externally controlled. Internal individuals are those who usually consider reinforcement to be contingent upon their own behavior. Three p a d n g procedures were employed in this study. One involved rewarding students (by adding points to their test scores) if they mastered unit tests on or before suggested deadlines. Another involved penalizing students (by subtracting points from their test scores) if they failed to pass unit tests on or before suggested deadlines. The third procedure, a control condition, i n v o l v e d i n f o r m i n g s t u d e n t s of the FIGURE 1
Expected Relationship Between Final Examination Performance and Score on the I-E Scale
195
suggested deadlines, but did not involve any rewards or penalties. In this study, it was expected that there would be an ordinal interaction between the three treatment procedures and students' scores on the I-E scale. Figure 1 depicts the expected relationship. It was expected that regardless of treatment group, the more internal a student (i.e., the lower the student's score on the I-E scale), the higher the s t u d e n t ' s final examination score. It was further expected that the final examination performances of those students scoring low on the I-E scale would be similar across the three treatment groups. However, it was expected that among students scoring high on the I-E scale, the students in the reward and penalty groups would perform better than the students in the control group. The reason students scoring low on the I-E scale were expected to perform well regardless of treatment group was twofold. Previous research had indicated that: (a) internals are likely to c o m p l e t e w o r k promptly (Alien, Giat, & Cherney, 1974; Prociuk & Breen, 1974; Ramanaiah, Ribich, & Schmeck, 1975; Henneberry, Note 1) and (b) the prompt completion of work is positively related to achievement in PSI courses (Allen et al., 1974; Santogrossi & Roberts, 1978; Sheppard & MacDermot, 1970; Sutterer & Holloway, 1975). The final examination performance of students scoring high on the I-E scale was expected to vary across treatment groups because it was felt that the reward and penalty conditions would induce externals to pass unit tests promptly whereas the control condition would not. Since previous research had shown that final examination performance and prompt completion of work are positively related, it was expected that externals in the reward and penalty groups would score higher on the final examination than externals in the control group. METHOD
Sample INTERNAL
I-E SCALE SCORE
EXTERNAL
Subjects in this study were 75 undergraduate students enrolled in a 10-week in-
196
EDUCATIONAL COMMUNICATION & TECHNOLOGY
troductory speech communication course at Florida State University. Fifteen students withdrew from the course during the quarter, leaving a sample of 60 students (36 male and 24 female). Twenty-eight of the students were seniors, 15 were juniors, 16 were sophomores, and 1 was a freshman. Twenty-three of the students were majors in communication. Of the 15 students w h o withdrew, 8 were in the control group, 5 were in the penalty group, and 2 were in the reward group. There was no statistically significant difference in the percentage of students who withdrew from each group. The mean I-E score of students who withdrew did not differ significantly from that of those who completed the course; nor did it differ significantly across treatment groups. In order for research studies examining aptitude-treatment interactions to have adequate statistical power, Cronbach and Snow (1977) recommend that 100 subjects per treatment group be used. It should be noted that, after withdrawals, the number of subjects per treatment group was approximately one-fifth of the number recommended by Cronbach and Snow.
Instructional Methods The course in this study incorporated many of the features typically found in PSI courses. Instruction consisted of six selfinstructional units. Each unit consisted of a list of 10 behavioral objectives and a series of readings chosen to enable students to attain the objectives. Tutors were available to help students attain the objectives. Most of the objectives required students to identify examples of concepts that had been discussed in the readings. Multiple-choice tests were used to assess student attainment of the objectives. Each test was made up of 20 multiple-choice items, 2 for each of the 10 objectives in a unit. A student was considered to have attained an objective if he or she was able to answer correctly both test items related to that objective. Students received credit (points) for mastering a unit if they were able to attain at least eight objectives in that unit--12 points for attaining eight objectives and 14
FALL 1980
points for attaining more than eight objectives. Students who could not attain at least eight objectives after taking three versions of a unit test received no points for that unit but were allowed to proceed to the next unit. Of the students who completed the course, 82% were able to attain at least eight objectives for each of the six units in the course. The number of points a student earned on the unit tests was converted into a letter grade, which accounted for approximately 50% of the student's course grade. The conversion of points to grades was as follows: 82 or more = A; 72-81 = B; 60-71 = C; 48-59 = D ; less than 48 = F. At the end of the quarter, a final examination consisting of 60 multiple-choice items (one item related to each of the course objectives) was administered to all students. Performance on this examination also accounted for approximately 50% of a student's course grade.
Procedures On the first day of class, all students met in one classroom and the course instructor gave a brief overview of the course content and procedures. Each student was given a copy of the I-E scale and asked to respond. to each item on it. Then each student was randomly assigned to one of three treatment groups and was asked to go to a corresponding smaller room. In the smaller rooms, students were given a schedule of the suggested dates for mastering each of the six self-instructional units in the course. The schedule, which was the same for all three groups, gave the students from 6 to 8 days to master each unit. Whereas the schedule each group received was the same, the consequences of following the schedule varied. Students in the reward group received two additional points for each unit they mastered by the suggested deadlines. Students in the penalty group lost two points for each unit they failed to master by the suggested deadlines. Students in the control group were neither rewarded for mastering, nor penalized for failing to master, a unit by the suggested deadline. Each group was
INTEI'U~,CTIONBETWEENLOCUS OF CONTROL AND PACING PROCEDURES
told only of the consequences applicable to that group.
197
TABLE 1 Summary of Multiple Regression Analysis with GPA as Covariate (N 60) =
Dependent Variables and Statistical Analyses Two dependent variables were examined in this study. The primary dependent variable was student performance on the 60item final examination. A second dependent variable was student attitude toward the course procedures, as measured by nine items on a Likert-type questionnaire administered at the end of the course. Multiple regression analysis was used to examine final examination performance. The predictor variables used in the regression equation were student grade point average (GPA), score on the I-E scale, treatment group, and the interaction between treatment group and score on the I-E scale. The number of test deadlines met initially was considered as another predictor variable. However, this variable, even when entered before all the other variables in the prediction equation, accounted for less than 6% of the variance in final examination performance. Thus, the number of test deadlines met was not used as a predictor variable. The predictor variables were entered into the regression equation in the following order: GPA (used as a covariate), followed by treatment group and score on the I-E scale (entered as a block), followed by the interaction between treatment group and score on the I-E scale. GPA was used as a covariate because previous research has indicated a positive relationship between GPA and final examination performance in PSI courses (Austin & Gilbert, 1973; Bestow & O'Connor, 1973; Kulik, Kulik, & Hertzler, 1977). Before GPA was entered as a covariate, it was necessary to identify whether an interaction existed between GPA and the treatment groups. If there had been a significant interaction, it would have been inappropriate to use GPA as a covariate (Kleinbaum & Kupper, 1978). Results of a multiple regression analysis indicated, however, that the interaction between GPA and the treatment groups was not significant.
Source
Proportion of Variance df
GPA I-E Score (A)/ Treatment ~, Group (B) } Ax B Residual
MS
F
,13641
1
252.356 11.20"
.05276
3
32.528
1,44
.16554 ,64529
2 53
153.099 22,520
6.80*
*p < .01.
Chi square analyses were used to compare the attitudes of the students in the three groups. A separate analysis, comparing the distribution of responses (number of students strongly agreeing, agreeing, etc.) across the treatment groups, was conducted for each of the items on the endof-course questionnaire. Nine of the 11 items on the questionnaire were intended to measure student attitude toward a particular aspect of the course, such as pacing procedures, grading procedures, and instructional procedures. The other two items were intended to discover whether students felt they (a) paid close attention to the suggested test deadlines and (lo) understood the course procedures. RESULTS
Final Examination As indicated in Table 1, the interaction between the three treatment groups and students' scores on the I-E scale accounted for a significant portion (16.6%) of the variance in final examination scores, F (2, 53) = 6.80 < .01. The regression lines depicting this aptitude x treatment interaction (ATI) are depicted in Figure 2. The regression equations from which these lines were derived are listed below (Y-- score on the final examination, X = score on the I-E scale): Reward Group: Y = 41.42 + .67X Control Group: Y = 45.51 + .16X Penalty Group: Y = 55.91 - .95X
198
EDUCATIONALCOMMUNICATION&, TECHNOLOGY FALL1980
FIGURE 2
TABLE 2
Actual Relationship Between Final Examination Performance and Score on the I-E Scale
Descriptive Statistics
Treatment Group 58
~~ ~
s6 \\
Reward Penalty (N=20) (N=19)
Control (N=21)
-o'9,"
52[-
\\
SOl--
~,~," \
i
Final Examination Score a
~
~_..~\~_._~-
4S~==~f~..,,. ,-
x ~1-
Variable
~"
SO I-E._Score b
\,,,
",%
x SD
,, ~,
42 k~<"~
\ ~x~oxOO
,T 40
45.65 6.39
47.10 3.51
11.16 3.45
11.60 3.49
9.76 3.59
2.43 .36
2.74 .57
2.56 .47
Grade Point Average e
x
\\
38
47.21 6.65
X
\\
SD NN
2 4 INTERNAL
6
8 10 12 14 16 18 20 22 I - E SCALE S C O R E EXTERNAL
Each of the equations is adjusted for the covariate (i.e., the grand mean of the covariate was used in calculating each equation). 1 The descriptive statistics used in calculating the equations are presented in Table 2. The standard error of estimate for the equations was 4.75. An inspection of the scatterplots revealed no obvious differences in the standard error across treatment groups. An inspection of the scatterplots also revealed that none of the relationships appeared to be curvilinear. Since there was an ATI among the three treatment groups, additional analyses were conducted to determine predsely the characteristics of the ATI. These analyses included computing the 95% confidence interval for the difference in slopes of the regression lines between the various pairs of groups. These confidence intervals are presented in Table 3. Each confidence interval provides a point estimate of the difference in slopes between a particular pair of treatment groups and also a measure of the predsion of the point estimate. 1As expected, the covarlate (GPA) accounted for a significant portion of the variance in final examination scores, F(1, 53) = 11.20,p < .01 (see Table 1).
"Perfect score = 60. bI-E = Rotter Internal-External Locus of Control Scale. The higher the score, generally the greater the perception of reinforcement as being externally controlled. CAverage of 2.00 = C, 3.00 = B.
As Table 3 indicates, the point estimate for the difference in slope between the penalty and reward conditions is 1.62. This is the best single estimate of the difference in slope between the two groups. The actual difference, however, may be some value other than 1.62. Since the confidence interval indicates that we have 95% confidence that the actual difference is between .71 and 2.53 (i.e., 1.62 + .91), it appears that the difference in slope between the penalty and reward conditions is positive but that the magnitude of that difference is likely to be as little as .71 or as much as 2.53.
TABLE 3 95% Confidence Intervals for D i f f e r e n c e in Slopes of R e g r e s s i o n Lines
Regression Lines Compared Penalty vs. Reward Penalty vs. Control Reward vs. Control
95% Confidence Interval for Difference in Slopes 1.62 -+ .91
1.11 --- .86 .51 -- .88
INTERACTION BETWEENLOCUSOF CONTROLAND PACING PROCEDURES
Table 3 also indicates that the difference in slopes between the penalty and control conditions is likely to be positive. The best single estimate of that difference is 1.11, and we have 95% confidence that the actual difference is between .25 and 1.97 (i.e., 1.11 _+ . 8 6 ) . Although the best single estimate of the difference in slope between the reward and control conditions is .51, we cannot be as certain of the direction of the actual difference as we were in the two cases described previously. As Table 3 indicates, we have 95% confidence that the actual difference is between - .37 and + 1.39 (i.e., .51 ___ .88). Thus, the slope of the regression line for the control condition may actually be greater than the slope of the regression line for the reward condition. It should be noted that the confidence intervals presented in Table 3 can be used to test the hypothesis that there is no difference between the slopes of two regression lines. In cases in which the confidence interval does not capture zero, we can reject, at an alpha of .05, the null hypothesis of no difference (i.e., we can state that an ATI exists). Thus, it can be said that an ATI was found when the penalty and reward groups were compared and when the penalty and control groups were compared, but not when the reward and control conditions were compared. Among those pairs of groups where an ATI existed, further analyses were conducted to determine the range of I-E scores for which there were statistically significant differences in final examination performance between groups. These analyses involved computing, for the range (2 to 18) of I-E scores found in this study, the 95% confidence interval for the difference between two groups' predicted final examination scores. For those I-E scores where the confidence interval did not capture zero, a statistically significant difference in final examination scores between the two treatment groups was said to exist. Confidence intervals for the differences in final examination scores between the reward and penalty groups are presented in Figure 3. Among those students who scored from 2 to 5 on the I-E scale, students
199
FIGURE 3 95% Confidence Bands for Differences in Final Examination Scores Between the Reward and Penalty Groups
in the penalty group performed significantly (p < .05) better on the final examination than students in the reward group. Conversely, among students who scored from 12 to 18 on the I-E scale, students in the reward group performed significantly (p <.05) better than students in the penalty group. There were no significant differences in final examination performance between students in the reward and penalty groups who scored from 6 to 11 on the I-E scale. Figure 3 also provides an indication of the precision of the estimate of the differences in final examination scores. The wider the interval, the less precise the estimate. For example, among students scoring 12 on the I-E scale, we have 95% confidence that the actual difference in final examination performance favors students in the reward group and that the true value of that difference is, approximately, between 1 and 7 examination points. However, while we have 95% confidence that the actual difference in final examination performance of students who scored 2 on the I-E scale favors students in the penalty group, the true value of that difference may fall anywhere between approximately 3 and 21 examination points.
200
EDUCATIONAL COMMUNICATION & TECHNOLOGY FALL 1980
Figure 4 depicts the confidence intervals for the differences in final examination scores between the control and penalty groups. Among students who scored from 13 to 18 on the I-E scale, students in the control group performed significantly better (p < .05) than students in the penalty group. There were no significant differences in final examination performances between students in the control and penalty groups who scored from 2 to 12 on the I-E scale. An indication of the precision of the estimates is also provided by Figure 4. For example, among students scoring 13 on the I-E scale, we have 95% confidence that the actual difference in final examination performance favors students in the control group and that the true value of that difference falls between approximately 1/4 and 7 points. As I-E scores increase, the difference in performance between the control group and the penalty group increases, but the precision of the estimate decreases. Thus, among students scoring 18 on the I-E scale, while we have 95% confidence that the difference in final examination performance favors students in the control condition, the true value of that difference may range between approximately 3 and 17 points. FIGURE 4 95~ Confidence Bands for Differences in Final Examination ScoresBetween the Control and Penalty Groups
End-of-Course Questionnaire
For each of the nine questionnaire items designed to measure student attitude, there were no significant differences in student responses across the three treatment groups. Student attitudes in each treatment group generally were favorable, with the majority of students in each treatment group responding positively to each item. There also were no significant differences across the three treatment groups with regard to student responses to the other two items on the end-of-course questionnaire. Approximately 50% of the students in the reward and penalty conditions, and approximately 10% of the students in the control condition, agreed that they paid close attention to the suggested test deadlines. More than two-thirds of the students in each treatment group felt they had a clear understanding of course procedures. DISCUSSION The results of this study indicate that the effects of various pacing procedures on final examination scores in PSI courses vary according to student perception of locus of control of reinforcement. Among highly internal students (i.e., students scoring less than 6 on the I-E scale), a penalty procedure seems to result in better final examination performance than a reward procedure. However, among external students (i.e., students scoring higher than 12 on the I-E scale), a reward procedure and a neutral procedure seem to work better than a penalty procedure. The results of this study were not as expected. Although previous research had indicated that the prompt completion of work is positively related to achievement in PSI courses (Allen et al., 1974; Santogrossi & Roberts, 1978; Sheppard & MacDerrnot, 1970; Sutterer & Holloway, 1975), in this study the rate at which work was completed accounted for less than 6% of the variance in final examination scores. One of the reasons for these findings may have been the small sample size. Cronbach and Snow (1977) indicate that in
INTERACTIONBETWEENLOCUSOF CONTROL AND PACING PROCEDURES
order to obtain adequate statistical power, researchers conducting aptitude-treatment interaction studies should use 100 subjects per treatment group. In this study, the number of subjects per treatment group was approximately one-fifth the number suggested by Cronbach and Snow. Since the sample size was much smaller than r e c o m m e n d e d , the results of the study must be viewed with much caution. It would be prudent to consider the study as exploratory and conduct additional research on the same topic using larger samples. Researchers w h o wish to examine further the interaction between paring procedures and locus of control in PSI courses should consider measuring the anxiety levels of the students participating in their studies. Alpert and Haber (1960) describe two types of anxiety, facilitative and debilitative. Although these types of anxiety were not measured, it is possible that both types had an effect on the outcomes of this study. Facilitative anxiety may have been a factor that influenced the performance of students in the penalty group who scored low (below 6) on the I-E scale. Lefcourt (1976) indicates that stress-provoking conditions c a n facilitate the performance of individuals identified as having internal locus of control. If the penalty condition in this study was more stress-provoking than the reward condition, then the difference in final examination performance observed among students scoring low on the I-E scale may have been due to the facilitative anxiety provoked among internals in the penalty group. Debilitative anxiety may have been a factor that affected the performance of students in the penalty group who scored high (above 12) on the I-E scale. Lefcourt (1976) indicates that externals exposed to stress-provoking conditions are likely to be affected by debilitative anxiety. Externals in the penalty condition may have experienced debilitative anxiety, which may have been the cause of their relatively poor final examination performance. The relationship between anxiety and performance in a PSI course has been
201
examined in two previous studies (Allen et al., 1974; Keller, Goldman, & Sutterer, 1978). Neither study found a relationship. Unlike the present study, however, neither of the previous studies involved the use of a possibly anxiety-provoking penalty condition. Unfortunately, the present study did not attempt to measure anxiety. Since there is reason to believe that anxiety affected stud e n t p e r f o r m a n c e in this study, it is suggested that anxiety be measured in future studies examining student performance in PSI courses. The Achievement Anxiety Test (Alpert & Haber, 1960), which measures the presence and intensity of both facilitative and debilitative anxiety, appears to be an appropriate instrument. Regardless of the role of anxiety in this study, the results suggest that it may be fruitful to continue to examine how various pacing procedures in a PSI course interact with learner characteristics such as anxiety level and perception of locus of control.
REFERENCE NOTE 1. Henneberry, J. K, Locus of control and experience as related to the prediction of academic performance and the use of course contingendes in a Personalized System of Instruction. Unpublished manuscript, Le Moyne College (Syracuse, New York), 1975.
REFERENCES Allen, G. J., Giat, L., & Cherney, R. J. Locus of control, test anxiety, and student performance in a personalized instruction course. Journal of Educational Psychology, 1974, 66, 968-973. Alpert, R., & Haber, R. N. Anxiety in academic achievement situations. Journal of Abnormal and Social Psychology, 1960, 61, 207-215. Austin, S. N., & Gilbert, K. E. Student performance in a Keller Plan course in introductory electricity and magnetism. American Journal of Physics, 1973, 41, 12-18. Bostow, D. E., & O'Connor, R. J. A comparison of two college classroom testing procedures: Required remediation versus no remediation. Journal of Applied Behavioral Analysis, 1973, 6, 599-607.
202
EDUCATIONALCOMMUNICATION & TECHNOLOGY FALL1980
Calhoun, J. F. The combination of elements in the Personalized System of Instruction. Teaching of Psychology, 1976, 3, 73-76. Cheney, C. D., & Powers, R. B. A programmed approach to teaching in the social sciences. Improving Collegeand University Teaching, 1971, 19, 164-166. Cronbach, L. J., & Snow, R. E. Aptitudes and instructional methods. New York: Irvington, 1977. Femald, P. S., Chiseri, M. J., Lawson, D. W., Scroggs, G. F., & Riddell, J. C. Systematic manipulation of student pacing, the perfection requirement, and contact with a teaching assistant in an introductory psychology course. Teaching of Psychology, 1975, 2, 147151. Keller, F. S., "Goodbye t e a c h e r . . . " Journal of Applied Behavioral Analysis, 1968, 1, 79-89. Keller, F. S. Ten years of personalized instruction. Teaching of Psychology, 1974, 1, 4-9. Keller, J. M., Goldman, J. A., & Sutterer, J. R. Locus of control in relation to academic attitudes and performance in a Personalized System of Instruction course. Journal of Educational Psychology, 1978, 70, 414-421. Kleinbaum, D. G., & Kupper, L. L. Applied re-
gression analysis and other multivariable methods. North Scituate, Mass.: Duxbury, 1978. Kulik, J. A., Jaksa, P., & Kulik, C. C. Research on component features of Keller's Personalized System of Instruction. Journal of Personalized Instruction, 1978, 3, 2-14. Kulik, J. A., Kulik, C. C., & Cohen, P. A. A meta-analysis of outcome studies of Keller's Personalized Systems of Instruction. American Psychologist, 1979, 34, 307-318. Kulik, J. A., Kulik, C. C., & Hertzler, E. C. Modular college teaching with and without required remediation. Journal of PersonalizedInstruction, 1977, 2, 70-75. Lefcourt, H. M. Locus of control: Current trends in theory and research. Hillsdale, N.J.: Erlbaum, 1976. Lloyd, K. E. Contingency management in uni-
versity coui-ses. Educational Technology, 1971, 11(4), 18-23.
Malott, R. W., & Svinicki, J. G. Contingency management in an introductory psychology course for one thousand students. Psychological Record, 1969, 19, 545-556. Miller, L. K., Weaver, F. H., & Semb, G. A procedure for maintaining student progress in a personalized university course. Journal of Applied Behavior Analysis, 1974, 7, 87-91. Powers, R. B., Edwards, K. A., & Hoehle, W. F. Bonus points in a self-paced course facilitates exam-taking. The Psychological Record, 1973, 23, 533-538. Prociuk, T. J., & Breen, L. J. Locus of control, study habits and attitudes, and college academic performance. The Journal of Psychology, 1974, 88, 91-95. Ramanaiah, N.V., Ribich, F. D., & Schmeck, R. R. Internal external control of reinforcement as a determinant of study habits and academic attitudes. Journal of Research in Personality, 1975, 9, 375-384. Reiser, R. A., & Sullivan, H. J. Effects of selfpacing and instructor-pacing in a PSI course. The Journal of Educational Research, 1977, 71, 8-12. Rotter, J. B. Generalized expectancies for internal versus external control of reinforcement. Psychological Monographs, 1966, 80, (1, Whole No. 609). Santogrossi, D. A., & Roberts, M. C. Student variables related to rates of pacing in selfpaced instruction. Teaching of Psychology, 1978, 5, 3O-33. Sheppard, W. C., & MacDermot, H. G. Design and evaluation of a programmed course in introductory psychology. Journal of Applied Behavior Analysis, 1970, 3, 5-11. Sutterer, J. E., & Holloway, R. E. An analysis of student behavior in a self-paced introductory psychology course. In J. M. Johnson (Ed.),
Behavior research and technology in higher education. Springfield, II1.:Charles C Thomas, 1975.