a. m» E. “m G a“. 3 “Ok 7 m S 5.5 Ra m? «U 9-». .2... F m}. 1‘ 3b: an.” m a. a»... A G J. Dual .hrh .m V g:g5::,,3’32::2,2,52:32., mm A ' 0'. I»: § .0? 631a amiss a I as 9 A? 6...: LIBRARY Michigan State University A COMPARISON OF THE EFFECTS OF PROGRAMMED AND CONVENTIONAL METHODS OF INSTRUCTION ON LEARNING AND PROBLEM SOLVING IN OLDER CHILDREN '\ h' 5’ {in .l‘ '1 Marilyn D ugherty A.THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1962 ABSTRACT A COMPARISON OF THE EFFECTS OF PROGRANNED AND CONVENTIONAL METHODS OF INSTRUCTION ON LEARNING AND PROBLEM SOLVING IN OLDER CHILDREN Marilyn Daugherty Michigan State University This study was designed to compare the effects of a linear programmed method of instruction with a more conventional method when learning time, instructor, intelli- gence, information presented, and motivation factors were the same for both groups. ”Conventional” material was in traditional printed narration form without accompanying lecture-type instruction. The same material in linear-type programmed form was given to the experimental group. I.Q. scores were obtained for each subject so groups could be equated on this variable. Learning criteria was a multiple- choice test. Problem-solving performance was reflected; by correct solutions to problems presented and solving- time required. Results showed significantly better learning for con- ventional subjects than program subjects. Programming of the material to be learned resulted in no better learning for subjects of below 100 I.Q. and resulted in poorer learning for subjects of above 100 I.Q. Type of instruction did not prove to be an influencing factor on whether or not subjects solved the problemspresented or on solving- time required by those who did solve them. The above results suggest that at least in some areas of learning and particularly for students of high-average or better intelligence, programmed instruction is still a very Open question. Approved ZQZEZ£ C7§E:§LL3 ” ajor Proflzssor Date #1; :2 1 /y41\ L» ‘/ 1/ ACKNOWLEDGMENTS I wish to thank Dr. M. Ray Denny for his excellent guidance throughout this study. He also provided the linear program which was used. The cooperation of the Grand Ledge, Michigan, and De Witt, Michigan, school systems is also greatly appreciated. Without such help this study would have been impossible. I also wish to thank Drs. Ratner and Dietze of the Michigan State University Psychology Faculty for their helpful suggestions concerning the writing of the text. M. D. II. III. IV. v. CONTEXT AND STATEMENT OF IETHOD RESULTS 1. 2. TABLE OF CONTENTS Subjects . Mate ri a1 5 . Procedure . Learning . Problem-Solving DISCUSSION . . . . . S UMMARY Iii) FOOTNOTES . REFERENCES . APPENDICES . O o 0 o o 0 CONCLUSIONS PROBLEM Page 11 15 16 2O 22 24 26 II. III. LIST OF TABLES Page Frequency of Multiple-Choice Test Scores by I.Q. . . . . . . . . . . . . . 12 Frequency of Solutions to Problem I by IOQ’ . O O O O O O O O O O O O O O O O 0 11+ Frequency of Solutions to Problem II by IOQ. . O O O O O O O O O O O O O O O O O 15 APPENDICES Linear Program . . . . . . . . . . . . . Conventional Presentation . . . . . . . Joe Doodlebug Problems I and II . . . . Multiple-Choice Test . . . . . . . . . . Verbal Instructions . . . . . . . . . . Wilcoxon Test on Program and Conventional Ss' Multiple-Choice Test Scores . . . . Significance Test on PrOgram Ss' Scores an I.Q. Group . . . . . . . . . . . . . . . Significance Test on Conventional 85' Test Scores and I.Q. Group . . . . . . . . . Significance Test of Above 100 I.Q. Se and Perfect and Non-Perfect Test Scores . . Significance Test of Below 100 I.Q. Se and Perfect and Non-Perfect Test Scores . . Significance Test of Program and Conven- tional 83' Problem I Solutions . . . . . Significance Test of Program and Conven- tional Ss' Problem I and Problem II Solutions . . . . . . . . . . . . . . . Wilcoxon Test on Program and Conventional 58' Problem I Solution Times . . . . . . Significance Test on All Solvers and Non- Solvers of Problem I and I.Q. Group . . Significance Test on All Solvers and Non- Solvers of Problems I and II . . . . . . Page 26 27 28 29 32 55 d 36 57 58 59 40 #1 42 45 1+4 A Comparison of the Effects of Programmed and Conventional Methods of Instruction on Learning and Problem Solving in Older Children Marilyn Daugherty Michigan State University The basic idea in programming is that the most efficient, pleasant, and permanent learning occurs when the student proceeds through a course by a large number of small, easy-to-take steps.1 A program is a sequence of carefully constructed items designed to lead the student to mastery of the subject with minimal error. Information is given to the student in small units to which he responds in some way. Items are designed so that the student can make correct responses while progressing toward more and more complex material.2 In the 1920's Sidney L. Pressey designed several machines for automatic testing. The student selects a multiple-choice answer. If wrong, the error is tallied. and the student continues choosing until correct; if right, the device mores to the next item. This "machine" was designed against a theoretical back-ground of the principles of "frequency" and "recency"; "massed" and "spaced" 3 practice. In recent years a "scrambled" type of programming has been designed whereby incorrect responses to multiple- choice items cause the device to switch to material designed to clarify misunderstanding and then proceed from there. In "linear" programming, the student first supplies the answer; he is then given the correct answer immediately and the device continues. B. F. Skinner has said that an important feature of efficient programming should be that the student must compose his response rather than select it from a set of alternatives.# One reason for this is that it is important for the student to recall rather than to recognize, and to make a response as well as see that it is right. Another reason is that effective multiple choice material must contain plausible wrong answers which are out of place in the delicate process of ”shaping" behavior because they strengthen unwanted forms. According to Skinner, "The act of responding tends to cause learning; therefore the student should not be exposed to incorrect alternatives."5 Linear programming is the type employedlin the present study. "Program subjects” referred b in this study are subjects who received the program found in Appendix A. The term "conventional" may apply to any of several tradi- tional forms of teaching presentation including lecture, group discussion, and printed narration form. As used in the present study however, "conventional" refers to a printed narration form only. No lecture or discussion took place. There are five advantages over other types of in- struction claimed for programming: (1) If each step is small and easy to take, the student is not likely to make errors. If a student makes many errors in learning, he often decides that he does not like the subject. (2) Better learning occurs when the student is actively responding to the material. He "learns by doing." (3) Immediate feedback of correct responses results in better learning. (4) Each student can work each step as slowly or quickly as he chooses. Programs are self- paced. (5) By looking over a program, one can see exactly what came before a step on which a mistake was made. Since accurate records of the learning experience of each student are available, revisions can be made on the basis of actual student responses. If the presentation of some point is not clear, this will show up on the performance of the program. A study on automated teaching methods using linear programs by A. Roe (University of California) compared the effects of (a) teaching machines; programmed textbooks; programmed lectures; and standard lectures; (b) free response vs. multiple-choice responses; and (c) individual booth vs. classroom. He found that all program groups performed significantly better than the conventional lecture group. The free-response group took more time than the multiple-choice and programmed text groups but there were no significant differences between them when tested. There was no indication that any one of the methods was particularly better for a particular aptitude level on the basis of the test criterion or learning time.7 J. H. Hughes compared the effectiveness of programmed and conventional classroom instruction and found that program students performed significantly better than the conventional students.8 Coulson and Silverman found similar results in a study on teaching machines.9 The relation of programming to problem-solving has scarcely been investigated. Watman (Michigan State Uni- versity) found that significantly more college age subjects in the group receiving a traditional printed narration presentation of the materials solved the "Joe Doodlebug" problem correctly than program subjects.10 Vatman's study--particularly in regard to materials used and experimental controls employed-~was very similar to the present study. Douglas Porter conducted ae22 week experiment on the teaching of spelling to elementary school children using teaching machine instruction. The experimental (program) group received the same words as the control group who used standard textbooks. Results were measured by standardized achievement test scores. The experimental group was significantly superior. No relationship between intelligence scores and achievement in the experimental group appeared, but a significantly positive relationship was shown in the control group. Porter's experiment seems to indicate that this type of instruction is particu- larly good for the lower I.Q. groups.ll Much exploratory study and definitive experimentation is necessary to ascertain the merits and disadvantages of programming compared to more conventional types of instruction in a learning situation.12 It seems that of accessible published studies involving comparisons of programmed and conventional instruction, it is almost. impossible to find studies utilizing strict experimental controls. A common error seems to be that of introducing "teaching machines" or programmed materials to one group and then comparing performance of that group with another group who remains in its usual routine. Little effort seems to have been made to account for:- (a) interaction between instructor and pupils in the conventional gnmup; (b) interaction between experimental (program) subjects and experimenter (usually a new personality to this group); and (0) how similar or different the specific information used in program and conventional groups may be. For example, when "lecture" is a part of the conventional group's instruction, which of numerous possible methods of lecture was used and whether or not each point presented to the program group was also presented in the lecture group seems never to be mentioned. The relationship between programmed instruction and intelligence is another area inviting research. The following study was designed to add one small drop of understanding to the-~50 far--very small pool of knowledge surrounding programmed instruction. An attempt was made to compare the effects of conventional and programmed methods of instruction on learning and problem-solving under the most stringent controls possible. The study was under- taken to find out whether or not subjects with conventional instruction would perform as well or better than program subjects when learning time, instructor, intelligence, information presented, and motivation factors (influenced by instructions, instructor, etc.) were the same for both groups. In order to help insure equal treatment of the two groups, the conventional subjects were given the material in traditional printed narration form. "Lecture" was not implied by the term "conventional". The material thus appeared to the conventional subjects in the same way as much of their school subject-matter routinely appears in reading assignments. These subjects however, were taken out of their daily routine for the experiment as were the program subjects. Also, the material was given to them by the researcher--not by their regular classroom teacher. Thus both groups had a novel experience. This aspect of the study is rather important since it is quite possible that seeming superiority of programmed instruction may sometimes be due to a new and novel experience by the program group and not a result of the program itself. Another question of interest in this audy was whether or not programmed instruction would result in superior performance by students of low average or lower intelligence compared to the performance of subjects of similar intelli- gence receiving conventional instruction as in Porter's study previously mentioned. Method Subjects One hundred fourteen (llA) eighth grade students from 13 to 15 years of age were used (57 program and 57 conventional subjects). Approximately half of these were from the Jonas Sawdon Jr. High School of Grand Ledge, Michigan, and the other half from the De Witt Jr. High School, De Witt, Michigan. There were approximately half male and half female students in both the program and conventional groups. Both schools are located in small communities and enroll average students. Of the subjects in the program group participating in the study, the mean I.Q. was 100 and I.Q. scores ranged from 60 to 156. Of the subjects in the conventional group, the mean I.Q. was 100 and I.Q. scores ranged from 59 to 131. Materials I.Q. scores were obtained by administering the Otis Test of Mental Ability (intermediate); 13 and the Science Research Associates Primary Mental Abilities test (inter- mediate).ll1L Material to be learned was information about a hypo- thetical character named Joe Doodlebug. Both control and experimental groups received the same information but the experimental group received it in linear programmed form with the necessary additional printed instructions. The program was constructed by Dr. M. Ray Denny (Michigan State University). This program will be found in.Appendix A» Appendix B contains the information as presented to the conventional subjects. The two problems faced by Joe Doodlebug and presented to the subjects to be solved will be found in Appendix C. (Problem I was presented to all subjects at the very be- ginning of the learning session along with the information to be learned). A lZ-item multiple-choice test over the information about Joe Doodlebug was used in two forms and distributed alternately to subjects. (Appendix D). This was used as a measure of how well the material had been learned. Three hints which were given to the subjects were clearly printed on large posters and placed in front of the groups at the prOper time intervals. (Appendix E). Procedure Students were assigned to the two groups on the basis of I.Q. scores. Of the two highest scorers, one was assigned to the conventional and the other to the program group; the next two highest scorers similarly, etc., until all were assigned. This was done prior to the day of presentation of the material to be learned and the problems. Each school was treated separately in this way. Since all conditions for program and conventional subjects were the same except for method of printed presentation, it was possible to work with experimental and control subjects in the same learning session. Four 8th grade classes of approximately 30 students in each and each containing both conventional and program subjects were used (each S wore a tag with a number corresponding to his name on the researcher's record, and a symbol indicating whether he was in the experimental or control group). The exact verbal instructions will be found in Appendix E. lO Verbal instructions were given. The information to be learned and Problem I was then distributed. Seventeen minutes were allowed to learn this material. During this learning period, both groups had Problem I available. Thus the purpose for learning the material and its importance in finding the correct solution to the problem was equally clear to both groups. The information was then taken away, the test administered and collected. The next 25 minute period was allowed for the subjects to offer solutions to the Joe Doodlebug Problem I. During this period, any subject wishing to offer a solution would move to the front of the room and quietly tell the researcher or assistant. If incorrect, he would return to his seat and keep trying to find the correct solution. If correct, he would return to his seat and wait quietly during the remainder of the period. A subject could offer as many solutions as he wished until he found the correct one. Each of three hints was given at 5 minute intervals at the beginning of this period. Problem I was then collected and a 3 or 4 minute break was given. Problem II was then distributed and 5 minutes were allowed for solutions to be offered. The researcher (and an assistant) wrote down the solution time when any subject offered the correct solution. 11 Results Learning (a) The l2-item multiple-choice test scores which were used as a measure to reflect learning as a function of type of instruction showed significantly better per- formance by conventional subjects than by program subjects. The Wilcoxon matched-pairs signed ranks test was used and a Z score of -2.55 derived which is significant at the .05 level. (Appendix F) There were 5 subjects in the experimental group who did not complete the program during the learning period. Four of these students had scored in the lower 1/5 of the group on the I.Q. tests. The other student achieved an I.Q. score in the average range. [Table I on page 12 lists scores for both groups. (b) No significant differences;as to perfect or non-perfect test scores were found at the .05 level be- tween above and below 100 I.Q. subjects in the program group. (Appendix G) (0) Significant differences in test performance using perfect or non-perfect test scores were found between above and below 100 I.Q. subjects in the conventional group at the .05 significance level. (Appendix H) More high I.Q. subjects received perfect scores than low I.Q. subjects. (d) Significant differences in perfect or non-perfect 12 TABLE I Multiple-Choice Test Scores Conventional 88 Program 58 I.Q. 12 11 10 pg 8 1, 6 5 4 12 11 1o 9 a J 6 5 4 130—139 1 1 ' ‘ 1 120-129 3 1 2 1 110-119 8 1 4 1 100-109 15 2 1 6 13 1 1 20-99 9 3 1 i 1+ 1 If 80— a9 5 2 j i 1 i 1 1i 1 70-119 1 l i 1 1 1 60- 69 i 1 1 1 i 59 I 1 I FREQUENCY OF TEST SCORES BY I.Q. 13 test scores were found between program and conventional subjects of above 100 I.Q. at the .01 significance level. (Appendix I) {ore conventional subjects received perfect scores than program subjects. (e) No significant differences in test performance at the .05 significance level were found between program and conventional subjects of below 100 I.Q. (Appendix J) The above results suggest that at least in some areas of learning the value of programmed instruction is still a very open question. Programming of the material to be learned resultedlin no better learning for subjects of I.Q. below 100 and resulted in poorer learning for subjects of I.Q. above 100. Problem-Solving (a) No significant differences were found as to number of solutions between program and conventional groups on either Problem 1 (Appendix K) or Problem 11 (Appendix L) at the .05 level. Tables II and III (pp. 14 and 15) show the number of solutions obtained for the two problems in each group. (b) No significant differences as to solution time were found between program and conventional groups for sub- jects who did solve the problem. (Appendix M) (0) Significant differences at the .01 level were found between above 100 I.Q. subjects and below 100 I.Qg sub- jects of both groups as to the number of problem solutions. The data indicates that the above 100 I.Q. subjects 14 TABLE II Conventional 33 Program 83 I.Q. Solved Non-solved Solved Non-solved 130-139 1 O 1 0 120-129 2 2 2 1 110-119 7 2 6 4 100-109 1 6 ll g5 16 9o— 99 ‘ 1 12 1+ a so- 89 IL 1 6 o 1+ 70— ]9 A l - 2 I O 5 60- 69 O i 2 O l 59 0 1 TOTALS 19 58 18 39 FREQUENCY OF SOLUTIONS TO PROBLEM I BY I.Q. 15 TABLE III Conventional 83 Program 38 I.Q. Solved Non—solved Solved Non-solved 130-139 1 O l 0 120-129 3 1 2 1 110-119 6 .3 5 5 100-109 A 13 6 15 190- 99 3_ 10 h 8 80— 89 l 6 l, 3 70- 79 1 2 0 5 60— 69 O 2 O l 59 O 1 TOTALS 19 38 19 38 FREQUENCY OF SOLUTIONS TO PROBLEM II BY I.Q. 16 were much better able to solve than below 100 I.Q. subjects in both groups. (Appendix N) (d) Highly significant differences between solvers and non—solvers of Problem I were found as to whether or not they solved Problem II. Almost all who solved I also solved II and very few who did not solve I solved II. (Appendix 0) Type of instruction did not prove to be an influencing factor on whether or not subjects solved the particular problems presented or on solving-time required by those who did solve them. Also there was no evidence of a relationship between how well the material was learned (as reflected by the multiple-choice test) and successful problem-solving. Studies by Rokeach indicate that certain personality factors--attitudes basic to the individual-- account for most differences in solving this particular problem.15 Discussion One consideration in attempting to account for the superior learning of conventional subjects is the fact that both conventional and program groups were given exactly the same amount of time in which to learn the information. Generally, studies reported on prOgrammed instruction state that program groups take more time than other types of instruction but result in better learning. 17 In this study however, program subjects had to use part of their learning period turning 52 pages and filling in 40 blanks while conventional subjects had all the information in narrative form on one page and were free to read and review it many more times than program sub- jects. If each conventional subject had been allowed to terminate the learning period at his own discretion, perhaps results would have been different. When superior learning seems to occur by program groups, itmay be a function of spending more time and making more responses to the material and not a function of prOgramming itself. More time and responding may have an.adverse affect also according to David Cram who says: "Students tend to tire more quickly when working on linear programs . . . The pace is slower and more concentrated; there is more repetition; and the student is forced to make more overt responses per unit of time." The comparatively poor learning evidenced by the experimental group could possibly be due to an inferior program. Possibly a different kind of program would be more beneficial than the linear program which was used, or this one improved upon. For example, if at the end of the prOgram a page of summary were added so that the sub- jects could have all the information before them at once and allowing them a comprehensive review, better learning may result. 18 Another possible partial explanation for the results obtained on the multiple-choice test is that as each program subject progressed to the next step, he may have had a feeling of finality as to the previous step (having filled in the blanks successfully) and began to forget it. This is of course suggesting a phenomenon something like the Zeigarnik effect.l7 Gagne and Smith conducted a study with ninth and tenth grade boys all above 110 I.Q. which may shed some light on the differences in performance of conventional and program subjects confronted with the "Joe Doodlebug" problem (although their study involved a problem different from the ones used in this study). They were interested in the effects of the subjects' own verbalizing during problem-solving performance. Results appeared to indicate that requiring subjects to verbalize during practice had the effect of making them think of new reasons for their moves, and thus facilitated both the discovery of general principles and their employment in solving successive problems.18 Since conventional subjects needed less time to actually read and review the material, they were free during the remaining time to ponder the information and "verbalize" silently to themselves. It seems reasonable to suggest that (silent) verbalizing by the conventional subjects could partially explain their superior learning 19 of the material in the light of Gagne and Smith's study. Another question which seems plausible in regard to superior learning by program groups reported in some studies is whether the motivation factor increases for program subjects as a result of a new set of stimuli. If programmed instruction were continued over a long period of time, it is questionable whether or not the program group would continue to perform better than other types of instruction groups. The novelty of being in an experimental group and being presented with new materials to use was controlled in the present study by treating the control subjects in exactly the same manner as the program subjects. All subjects in both groups had their usual routine interrupted in order to participate in an experiment. All subjects worked with information quite different from usual school subject matter and with persons new to them. In this way the "novelty factor" which presumably operates in favor of superior performance by program.groups in many studies was eliminated. As 1hr the "self-pacing" claim for programmed instruction, the conventional subjects in this study were actually more self-paced than the program subjects since the latter were required to fill in all the blanks and were thus "tied" to the program. Leslie Briggs conducted an experiment in which each of two groups was given a 20 total of 13 minutes practice on a 20—item paired- associate problem. One group was self paced; pacing was automatic for the other. Both were then given a test trial. No differences between effectiveness of the two methods was demonstrated.19 It would seem from this study that students of high average or better intelligence may find programmed instruction (at least in some situations) too cumbersome to be advantageous. Certainly such investigation is a vital current need and requires more--and in general, better Controlled-—studies than are now available before the many questions surrounding programmed instruction can be answered. Summary This study was designed to compare the effects of a linear programmed method of instruction with a more conventional method when learning time, instructor, intelli- gence, information presented, and motivation factors were the same for both groups. "Conventional" material was in traditional printed narration form without accompanying lecture-type instruction. The same material in linear- type programmed form was given to the experimental group. I.Q. scores were obtained for each subject 30 groups could be equated on this variable. Learning criteria was a multiple-choice test. Problem-solving performance was £27.... 1“ 21 reflected by correct solutions to problems presented and solving-time required. Results showed significantly better learning for conventional subjects than program subjects of 8th grade level. Programming of the material to be learned resulted in no better learning for subjects of below 100 I.Q. and resulted in poorer learning for subjects of above 100 I.Q. Type of instruction did not prove to be an influencing factor on whether or not subjects solved the problems presented or on solving-time required by those who did solve them. The above results suggest that at least in some areas of learning and particularly for students of high-average or better intelligence, programmed instruction is still a very open question. 22 Footnotes lSlouth, Dewey A., "Descriptive Statistics" (Denison University), The Psychological Record, Vol. 10, pp. 224-226. 2Center for Programed Instruction, Inc. (pamphlet-- no author given), May, 1962. 3Markle, Eigen, and Komoski, A Programed Primer on Programing, New York, 1961. ?,§ uskinner, Cumulative Record, New York, 1959, p. 159. 5Cram, David, Explaining_Teaching Machines and Programming, Fearon Publishers, San Francisco, 1961, p. 39. 6Principles of Programed Learning(No author given), 1961. 7Roe, A., et. 21,, "Automated Teaching Methods Using Linear Programs"_(University of California), Programed Instruction, Vol. I, May 1961. 8Hughes, J. H., "The Effectiveness of Programmed Instruction: Experimental Findings for 7070 Training", Programed Instruction, Vol. I, New York, January 1961. 9Cou1son and Silverman, "Effects of 3 Variables in a Teaching Machine", Journal of Educational Psychology, Vol. 51, pp. 135-193. loWatman, Walt, "A Study of the Effectiveness of Programmed and Traditional Methods of Instruction in a Complex Verbal Problem-Solving Situation", Masters Thesis, NHchigan State University, 1962. 11Porter, Douglas, "Some Effects of Year-Long Teaching Machine Instruction", Teaching Machines: An Annotated BibliographyJ Department of Audio-Visual Instruction, Washington D.C., 1960, pp. 58-59. 12Center for Programed Instruction, _p. cit. 13Otis Self-Administering Tests of Mental Ability, Arthur B. Otis, Intermediate Examination, Forms A and B, World Book Company, Chicago, 1922. 23 lLASRA Primary Mental Abilities, L. L. and Thelma Thurstone (University of North Carolina), Science Research Associates, Inc., Chicago, 19#8. 15Rokeach, Milton, The Open and Closed Mind, New York, 1960, pp. 171-196. 16Cram, pp, cit., p. 52. l7Ruch, Floyd L., Psychology and Life, Scott, Foresman, and Co., Chicago, 1948, p. 579. 18Gagne and Smith, "Effects of Verbalization in Problem-Solving", Journal of Experimental Psychology, Vol. 65, p. 12. 19Briggs, Leslie J., §_t_. §_1_., "Self-Pacing Vs. Automatic Pacing of Practice on the Subject Matter Trainer", Teaching Machines: An Annotated Bibliography, op, cit., p. 21+. 24 References Center for Programed Instruction, Inc. (pamphlet--no author given), May, 1962. Coulson, John E., and Silverman, Harry F., "Effects of 5 Variables in a Teaching Machine", Journal of Educational Psychology, Vol. 51, June, 1960, pp. 135-145. Cram, David, Explaining Teaching Machines and Programming, Fearon Publishers, San Francisco, 1961. Fry, Edward B.; Bryan, Glenn L.; and Rigney, Joseph W.; TeachinggMachines: An Annotated Bibliography, Department of Audio-Visual Instruction, Washington D. C., 1960. Gagne and Smith, E. 0., "Effects of Verbalization in Problem Solving", Journal of Experimental Psyphology, Vol. 65, January, 1962, p. 12. Hughes, J. H., "The Effectiveness of Programed Instruction: Experimental Findings for 7070 Training", Programed Instruction, Vol. I, Bulletin No. 1, Issued by the Center for Programed Instruction, New York, January, 1961. Markle, Susan M; Eigen, Lewis D.; and Komoski, P. Kenneth, A Programed Primer on Programing, Published by the Center for Programed Instruction, Inc., New York, 1961. Principles of Programed Learning, (no author given), Teaching Machines Inc., New York, 1961. Roe, A., 23, 31., "Automated Teaching Methods Using Linear Programs" (Department of Engineering, U. of California), Programed Instruction, Issued by the Center for Programed Instruction, Vol. 1, Bulletin No. 1, May, 1961. Rokeach, Milton, The Open and Closed Mind, Basic Books, Inc., New York, 1960. Ruch, Floyd L., Psychology and Life, Scott, Foresman, and Co., Chicago, 1948. 25 Skinner, B. F., Cumulative Record, Appleton-Century- Crofts, Inc., New York, 1959. S1ough, Dewey A., "Descriptive Statistics" (Denison University), The Psychological Record, Vol. 10, pp. 224-226, Albuquerque, 1960. Watman, Walt, "A Study of the Effectiveness of Programmed and Traditional Methods of Instruction in a Complex Verbal Problem-Solving Situation", Masters Thesis, Michigan State University, 1962. t.’ r. .' i.-. '— t k .0 Please learn t}_2e to learn it well solve the proolcm rU --..;.OI"“Lc. {1d To help you labelled T XE avil to 11 you wha t tiis booklat. It is r it tiorou"nlv in order do this some gages a; you ziaed to problem. Ea h text pag e is follorzed by one pages (QUES'PION PAGE) turning to the” an“ —J-U...J ouestions before answers can be found (ANS 5 (a If afi.e1 seein text, it is a and do not look \ ‘J . p.-. Please W75 right to do soo K‘Tf"-L 1:. kn-.:;..-R L 10-11 -‘_-.-_v ~-»—‘- fizz..-»- Joe sees the food he is f L0... "Y I s.“ n1 n‘w.-o~- -<--.—--— -.—— u. ....— whim-o!- AA A ‘* "‘—— ~‘--—.— -.-_. - .— - 1+2. TEXT c-M After all his exercise Joe is very hungry and wants to get to the food as Quickly as he poss'bly can. Jos examines the situation and feels frustrated when he realizes that he will absolutely £333 to make four jumps and no fewer before he can get the food. h Ci} 5% & . . .. J 3‘ kn, _. 0 9r t Y 0'. .__~.-.. --—‘ J‘- UJ'—H '-_"‘ I e y. I —" _—_—.—— '— ' . , .. .._ wv .—_»~-«...=a_. '4 -; .LLx...‘ “'3t‘31fir.m--gfigmmw~rflw___" . > __ ._._ _. +5- __ _ v—— h 08.1". 3111‘ p K APPENDIX B CONVENTIONAL PRESENTATION 27 s-a' u ‘ (I. ‘ - k y '1“ 5‘! .h. 9 2‘. As 9 n v 1*; O .1? A n 1.: 3. : O I 1 l a A. gr, ' .1,“ z. is, n. thai £88 in t 93 t 14' ___‘ LA fig ,0 -on. I“), an. '$ '1 .L 1'. .1' o... 4...; ' ‘9'. A. ~'. “1‘ .u. 'A‘ -r . ~ ". ‘ Jr R ‘J 7 G" "or. “If. .u . _, v! 1 '1 'u’ ‘i‘ Jae starts Once ~ . u... r w... 0.. an.” ”at. LL. .- ..:.l .f 4 1.3 p. m; 3:. (Tu . u . . . J... L 1....“ a. -. . ., F E I... III alrrx S I...‘ F.‘ t ‘5) a I c to V A. w. L .fiw m. .3. .3 a 8 HO. a1“. mg .....l 1‘. LI" 3 s 1 s is man \ \ 2,1 .- 1) 8 Q3 1 13. . .J. . 1: ..a. . .. . a. IIVI ’. .. rm. Au 3.. s a U Ea T w G We 5 .d C.n . . 4 .3 . a a) m)... . a 6. ,9 fit r, 'i 1; Ci of ad hu 'K a i lab... I f iie a f very O ‘I ‘11 gr. 4. a E. '1 C ‘L‘ 89".) (a 3 1 1 no 3». n. a. 3 v .1 «2 «Na anuh as 1r... . ." ALA LIV Wit . . - \ e .1 ¢ D.“ it» n 1 0 .; a . a t. C. C vh ; Maria Own 9 2.1.. :2“. P... n on. .3 #0 8 L I ‘ a‘l .\ .. . .5. : 3 w... 4.. «a... my. my. 0 A 1 h 9 .3 2 m J C . «.3. H. 10 a 1L. «. ..... uh. .u a...“ 7... «wow .fll. ”H .fi. A. a a”. r n . . w... _.c a... hr... 3 .. . 1 _u.r u : D 1 (A... or r ” m.¢ n.n n A . .. . . a.) ... on .u .. .t qlxw In -6“ do“ 0%.“. *5: filo «all .mluJ VJ W ’Qb IN.“ t ‘1...“ real“ no .1 -.l\-‘. ha.“ A L. I‘ "I": ‘33 1. v V? J. .1. 8 n u. 1'? C1 Lia-3. '. "o 28$? 1‘ APPENDIX C JOE DOODLEBUG PROBLEMS I AND II 28 Since y u now know all the things that Joe can and cannot do and you also know that if Joe could get to the food in fewer than h jumps he would, the problem is to explain Egy_Joe Doodlebug takes exactly four jumps, no more and no less, in order to get the food. _- "”9." 1.! » f“... ‘ K order to turn f‘..1"<.‘-l.11; . . I. ’(,‘.-r- , ' , _ . f. ‘ g . -_ . .' . , . “.-~ . - 1 I . r. - . . .' a ' , ‘ ’ ' A L n . . 4 '. 4t. ‘ “ ' z ‘ — \ ‘.('* I, y. . - >.’ I- ‘ ' - - a< - " 'I -~ u . ~73 ' ‘ 2 . t, - .' v ‘ 3- . . J 4 , .. , , . >3 . ~ ,- - I, \.. - \- ‘ ~ 1 " . ' .. 1’. I . ‘_. s ' _ .. . - ..I , .. . -. _|._ . . . ~ - . . . ., ~ I '1‘ ’, ', I . ‘ ‘.. J '1- . _. . 1 r ~ . ‘ . . .' . ' ‘ . ... ,.. . .. g. ._. . .. , '1‘ .. ,' I 7. ._ .‘ I, - ,. . 4. .- . « ‘ A ‘ ' H - -- ‘ ~ r g. *5» V . ' . .' ...' J'AI -J ' k a .n 5- ‘- .. u' f ‘ . '4 ~ * , . —‘ -' . . u_ - , , I- -. ,‘ -. .;. _ 2.. . u » ‘ .r ‘ .l, 4 ' 1 . ~ - . . - ' ‘ r- . ’ . . . . ' . .' , . ' - ‘ I .. . . , ‘ I . _ . 7 . ' - ‘ ‘ - - . - ‘. . .- . ‘ . A . , . . I ‘ , .‘l | - - ' .. I . - ' . - . ._ . . . ‘ ,. _ ‘ , . . A . __‘ . . .. 3.! ‘-'.:.r*.‘ , ' \ ' , _ - . .. - . - . _ .v . - a . . . . -. . \ -- , -- ._ « v .. ' ‘ . '< - , . my . . ‘ . . _. . , . . 44 ,. v . .i\ ' . r _ . .- . . . ,.; ’ . I ‘ K . . n v . ‘ , ‘ . _, . I f . . .. ~ ~ ‘. . . - . . - . n w. . . — A . - .k h. u - , 'J ‘ ‘ . _ ‘ \ - I . . ‘ ' . . , r . - . ~- . _. —_~ . . ' , . -' -. . ~ I. '. . A . rt: ’03-- . - , ‘ _.. . --. . ., - ‘u ‘ -. . A‘ ,v x , . . - _- .' . ' . . _ . ' _ . ‘ . .. . - , " . I . . ‘ ~ . . - ‘ . ' - . - . . - ‘ - H ‘ ._ _ - ‘ .— . . , , . t I . . ~ . . y - . ‘ < n . ,« . . - . - . . .A . . , ‘ .\ ~ .. . ,.. l 1 . , , ‘ 4 > ' I - - '. . . _ 1 ‘ O _ ' . . . r ~ , , . * - ‘ . ‘4 _. .IJ-1,‘!‘ APPENDIX C JOE DOODLEBUG PROBLEMS I AND II 28 ’t L .. '3' "t I. *. r" ‘o ’.. I‘L... L., c '7 ‘-1 1 -h . I . . .3. ., a. '1. r. _ \# ._. .I .. A: ~.. IL . u. .. f .. .. min. A l: n . 1 xi... oow no... a. . . .. .le . .1 u. \I . .r. . . 0.. .4 .. .. . . n ,C .1’. . o - ... .. .9. . . a I L a. U h 1.... . . . . 1. I . .1 f; 3.... uv. .4 fi . l 11 '5 . . . ..J .l. x... ..- a... . .. HI- I. \: n x 7. . s. . y . . . 2.. J 1... a... .. . 9 K. fl. In t i. I. N .r V! :L - )0. “m r2. 1, .. .J .J. 1. )-! 4 O. IHI A .. .51 1.... ~V¥ tin. .. a - > . I" 4‘. ‘ 1 ‘ I qu J ’ "' [a 4 _ ., - ‘ . Q" $35.: ..v c ‘o ' 'l r ', :qw‘, m- " \‘fi. .VG ’U' ‘ it " ' ‘ '. l " 3.3.11: _) \ \— - , h . V ' u t ”‘1" ‘ 3 . i ‘ . , ‘, 1 . . ~ "I , ”r < , ‘5‘, k-(.“\ , .1" A ‘ ’ 1" ; . .1'} k.) ft 31;. .: Tt-«I‘u .In |., ’3‘ “-4"'\ 4.} I ,0 V5 -— ’u' o- k‘ I _, 1. . . v. {.fl J. '.7 -l \ - it"s- .’ ,‘, "' - r t j 9 ' ~. \ 1 . I , - ,-\ .5 .' q }‘ Iv fl. .‘,>i-l , ,- .._ ., =- V_. .4 - l'l_ ,. ,-~‘<~.‘- - - - -' 4- .1 w . I . U 1 * a. 3 . X m x —‘». t‘ I Q ‘ ;:q(‘u , ‘3‘; - ‘,, ‘1‘»? l1,{" (- l ‘4 A‘. z . ~ .. . .' -‘ , ,H._\ ’, I ‘ ’f"L"3-1.7"‘.}r»‘ ‘._‘,$ (1 I1 I. I. I *. 4 L“ \, . < , to -\._~ 5 \v' k‘ -“- ‘v \‘ :‘. ‘ IL-u'w ‘14 n...» .n . ~~1-n'u ‘ ‘ 3., Lu I.—\t :- A '-’ 4W“H.'. l o. - ,,; , ’ . . ‘ - a. .- k. . - ‘yqv_;-_ .‘ .-..».v ‘..~‘ _ ~ I q. way g. ‘4‘ V r’ F‘ E . ~ .- I ’1 c ’ -" ._, ‘$~' , .L',_ (:3 .L)" __‘ 1'7 ‘.7 ,‘- . ‘. ' -‘ 4 . ' ‘ l - - ‘ . - " , . - . .. ,~ .» 4 . .7 _. . . . ‘. '5 ,1. L; .'_ x 1 I H- :- 4 . J .. ~a ! t '1 ‘ g ‘ .' u v,‘ ,1. £5 .' ' U‘ . 7 0. a m r-s on ‘ x.‘ ha. 'J r." ‘ 1" ‘ —~~'\ 3'. L.. -.: ' - ' APPENDIX D MULTIPLE-CHOICE TEST 29 ,x‘ ‘5‘ -... .‘ _ . 4.1- J J. . p .L . I . .5 ". . ‘ fi. “inar DYK£1KEC{ Lne Siiudhjon; see rflfilchS AH «111 tu»s ta ~'~.. .- « - ”. .-"‘ p. - ~'- - ' 1 a: 1;. .,- s w. .u j,._.. jig-*1 and u} jcwar Lg: ,. :3 re, )1 0 Jan .521; ,«2 1.7;r»,;_, ; . '3 9.! i l ""T‘ I! --. ". 5.. ~ ’, ‘l.’ ’ 1.? I} f 1f .52: 5 Hiram r a \ 1“ a... A 'w ) ‘ V " b H - ~ ‘ ‘ ' s b . .n r. ' v‘ ---A1 .,. ’\.' t; * b9 84»: Jae 14 wefy a»fi&*y re mints Av gut to any fun; 1%} as ssox'y as he .au 1' ' __ - ‘ ab! He qu 43’« &8 he can (I. ‘ J .2 . i, ‘ ,. - " .,. Y ._, _, ,_ a&; in “we iCL.8 xr’er me flflPa :5 . 7‘ ’ ,‘ ‘ ‘a’ , i 1,. . 3,, d J at 1.118 {snag 0): 414*..— (13“;- : .7 ‘-... :4, 4': -..~ .. ‘, - V. x Eye 3 manter Slkzna Ldfi kcgc dzrafitly at i;m, ‘ ' uakmciu'» . .. ~4nmnu 521 Narrh A I I - ‘ ‘I _, . )h, south '1‘ ; £313.: i I. ’ {d} nggi ‘a . 7,. . '- A' ' A “EMT: Jae giant; 1.} -.. Ion-5:. 5's» )5; r“. i 1;: .A {1 .' 5:; ('4 a3. 2.? _IJ . I‘. .- ll" n ': ‘l' i? :3 ‘- r t. ,. " C 1' T"; "~13“- .0’ . '. ,_ t {if :1?) 5.“? E" 1 "9t ‘ I ‘ '7 \ ~ f 5 lb% n1;e f Ififiu 9 ‘ 'n”‘~lm‘fl'rlml'~a&“‘>‘Jr‘~ «12' II'LI 6‘. 1-...--“ ll:- (L J. :J {23.1. t ’ :I‘ x. t 1.1":- L1 l' f.‘ 3‘ a. his --.I"".’ z' Elem .>r .‘i' “ . ;35 {SJ affl fi'a“ as .~ ' d } é: ‘1‘” ‘_ “."A’ 5. , ‘ L p q f; ' ‘ “‘1 L} .. 5‘.‘ .31 ‘ ’35? 2'5"? ‘ ‘:._ x ", WA. 1 '2‘ .“:;' ’5‘ "“5 3. ; Nd: .‘\ h 3 ”5; .1: ’ . J {'1‘ ca 7' ’- , ‘ A - . .~ ‘g .. ., .. ; ‘. 'K I .‘ F7 ‘ 2‘- ‘ o 2. 1 '2 I 4‘ ‘- 1’? L! ‘1‘ N" i“ ‘l \‘ ;a.‘ .' (f' ‘3, ', 'oil'i." |( v‘ I Y‘. -,.-..... .. ’V. .1 '~ '7 .. - ’ gar..1Junu: 11rC=9 2h” .eiter af the ~J~ yum-“n. IO! ~. '41-. -'..b"-Q ¢-M.~w‘. quasi? cm- I. . . ,~ 1"'\L-'j:7 C .“ , {OFFEJ‘J [0." 96‘1”. €3.71 9 ‘5") 2' \ . u- '3 "\‘J 3C9 (“in turn arm; nd I’ ‘1 .~ ,, '0; a I n“ ' er §b§ at Lha end hf a aequenao of {our juspa :n 03% ‘h . (2) '1n the 313' {d} tflmn hp ia}uuyflw 393 can cniy (a) wa3k I’ a 1b) {53 . x , ’5. C2 .3 _'2 U mp (d5 trawl J09 can thauge dlrv;ttrn$ " I £32 Kb} (3} J « ' \ 35‘ 3 \ I C Before Joe at any Limp when he 18 movanc sauth nnly uftar JUMpxng four ttmva In cne dfirvutzon m an H: a sunny '38.? #an switvh hm ~nozhvr dxrertlou: he must jump times in the ram» dirqetlon. 3 2 3 ’W‘ ,ntgt. A (a " £3: a a" an four three tWU ()0?) Joe can jump diagona7ly 333‘” L): 4‘ . ..."l"3‘v_‘.. .A‘ “\‘WH 12m firm. Jump In {a} I‘ . vi. tr. 3 (-2 g f a at the and af R Baguenre “9'9?!“ In th9 maddfia ff 8 HEQUBD‘W 1:?“er ha hunts fink“ ‘1? Only [3'3” ‘t‘ in": ‘3; 3V4 (- 7 Our 31 5:- “$0.“- :0“. 52 APPENDIX E VERBAL INSTRUCTIONS "This is not a demonstration. This is an actual research project on problem solving. The puzzle you are to be given is difficult but it is definitely solvable. We are interested in your finding the solution. The material you will be given contains all the information necessary to obtain the solution. Therefore it should not be necessary to ask questions. You will be given sufficient time to learn the information thoroughly and you may review it as often as you like. After a period of time the information will be taken away and you will be tested over it. As already stated, the puzzle is a diffi— cult one and it cannot be over-empthized that your ability to solve it will depend entirely on how well you learn and remember all the information given to you. If you do not learn this information well, you will not be able to solve the puzzle. In a moment you will be given the information to be learned and the puzzle. You are not to turn them over until instructed to do so." (Information and Problem I distributed). "You should use this period of time to learn the information which you will need to solve the problem or puzzle. You will have a chance to tell one of us your 35 answer in the next time period. If you think you have. the answer, the best thing to do is to look back over the material and see if your answer is really right, and spend your time making sure you understand the information thoroughly. Begin." (17 minutes allowed). "The time is up." (Material collected. Tests passed). "Write your name and number. Circle the correct answer. Begin." (5 minutes allowed. Tests collected). (Subjects retain c0pies of the problem). "In this time period you should try and solve the puzzle. Any time from now on that you are sure you know the correct solution, walk quietly to the front of the room and whisper it to one of us. Make sure you speak softly enough so no one else can hear you. If you are told your answer is wrong, this doesn't mean everything you thought is necessarily wrong; just return to your seat and continue trying to find the correct solution. You can have as many tries as you like. You may also write or scribble on the problem sheet you have there if it will help you. Here is a hint:" When Joe sees the food and stOps dead in his tracks facing North, he has already taken gpg_jump East. He must take 4 more jumps to get the food. (5 minutes given). "Here is another hintz" Joe can jump sideways and backwards as well as forward. (5 minutes given). "Here is the last hintz" Joe does not have to face the food in order to eat it. 54 (15 minutes given. Problems collected. 3 or 4 minute break. Problem II distributed). "This is the 2nd and last puzzle." (Same instructions as Problem I. 5 minutes given). 35 APPENDIX F WILCOXON TEST ON PROGRAM AND CONVENTIONAL Ss' MULTIPLE: CHOICE TEST SCORES 209.8 - 1§£231 Z— - = -2.35 ‘W 2a Significant at .05 level. 36 APPENDIX G SIGNIFICANCE TEST ON PROGRAM Ss' SCORES AND I.Q. GROUP Test Above Below Scores lOO IQ _ lOO IQ“ Perfect 13 8 21 Non-Perfect 22 | 14 1 36 55 22 57 Z: -.21 Nonrsignificant APPENDIX H SIGNIFICANCE TEST ON CONVENTIONAL Ss' Test Above Below Scores _ lOO IQ, lOO IQ Perfect. 27 15 Non-Perfect I 4 ll 31 26 Z= 2.2 Significant at .05 level. 37 SCORES AND I.Q. GROUP 42 15 57 58 APPENDIX I SIGNIFICANCE TEST OF ABOVE lOO I.Q. 83 AND PERFECT AND NON-PEREECT TEST SCORES Test Program Conventional Scores 7 Ss J 83 4# Perfect l 13 I 27 I 40 Non-Perfect] 22 l 4 l 26 35 31 66 Z: 5.89 Significant at .01 level. 39 APPENDIX J SIGNIFICANCE TEST OF BELOW lOO I.Q. 88 AND PERFECT AND PEREECT AND NON-PERFECT TEST SCORES Test Program Conventional Scores 88 Sg, Perfect 8 I 15 23 Non-Perfect 14 l 11 25 22 26 48 2: 1.184 Non-significant 40 APPENDIX K SIGNIFICANCE TEST OF PROGRAM AND CONVENTIONAL Ss' PROBLEM I SOLUTIONS Problem I Program Conventional 7 88 SS Solved l 18 l 19 | 37 Non-Solved, 39 l 38 g] 77 57 57 141+ Z: O Non-Significant 41 APPENDIX L SIGNIFICANCE TEST OF PROGRAM AND CONVENTIONAL Ss' PROBLEM I AND PROBLEM II SOLUTIONS Problems Program Conventional 88 SS Solved I Only‘ 18 J 19 I 57 Solved I & II 16,| 16 I 32 54 35 69 Z: .13 Non-Significant 42 APPENDIX M WILCOXON TEST ON PROGRAM AND CONVENTIONAL Ss' PRDBLEM I SOLUTION TIMES 199.5 _ 2112§). Z: I = .252 f22)§22)§§§) 2 Non-Significant 43 APPENDIX N SIGNIFICANCE TEST ON ALL SOLVERS AND NON-SOLVERS OF PROBLEM I AND I. Q. GROUP Problem I Above Below 100 IQ, lOO IQ Solvers I 30 l 7’ 37 Non-Solvers 7 56 J 41 77 66 48 114 2.. 3.27 Significant at .01 level. APPENDIX 0 SIGNIFICANCE TEST ON ALL SOLVERS AND NON-SOLVERS OF PROBLEMS I AND II Problem Problem I Problem I II Solvers Non-Solvers Solvers I 32 6 l 58 Non-Solvers_ 5 71 7 76 37 77 114 Z: 8.13 Highly Significant. ..,,..3'\){fi USE ONLY «.553! .2 an " F M. It" ‘5 T f _. MICHIGAN STATE UNIVE Ill ll ITY Ll IN" IHIIIWIRIIES 7 O 9 31 8 3 1293 130