I I I l I II I I I III I II 144 884 HTHS PREDICTEVE VALIDITY OF INTELLIGENCE FOR THE DRAWING COMPLETION TEST (DCT) UNDER DIFFERENT CONDITIONS OI? INSTRUCTION Thesis {or 1413 Deans of M. A. MICHIGAN STATE UNIVERSITY GeraId G. Griffin 1963 in; I? :5: rs LIBRARY Michigan State University ABSTRACT PREDICTIVE VALIDITY'OF INTELLIGENCE FOR THE DRAWING COMPLETION TEST (DCT) UNDER DIFFERENT CONDITIONS OF INSTRUCTION by Gerald G. Griffin The problem of this study consisted mainly of two parts. The first part was to determine the effect four different Group Conditions, each one presenting different instructions designed to either induce no special motivation or learning set, induce special motivation only, induce a Special learning set only, or induce both a special motivation and a special learning set, would have upon mean Drawing Completion Test (DCT) scores. The second part was to determine the relation- ship or predictive validity of intelligence (as determined by correlation coefficients) between the DCT scores obtained under each of the four different Group Conditions and the scores of an independent measure of intelligence, the College Qualification Tests (CQT). The subjects for this study were 120 beginning psychology students who were members of four separate psychology 151 classes selected at random from Michigan State University. These four classes, consisting of approximately 30 students each, were randomly assigned to the four differ— ent Group Conditions previously described. The instructions of the control group (Group Condition I) were designed so as 'Gerald G. Griffin to induce no special motivation or learning set. The instructions of Group Condition II were designed to induce special motivation by informing the subjects in the group that the DCT was a new type of intelligence test and that they should try to do their best in their responses because these responses would be scored and used for future evaluation. Group Condition III provided a learning set by special instruction in the form of a lecture, but this instruction induced no special motivation. The instructions of Group Condition IV induced both a special motivation and provided the special type of learning set presented to Group Condition III. The subjects in each of the four groups were given approximately 30 minutes to complete their responses to the DCT. The subjects were asked to code their completed DCT blanks. These blanks were then collected and scored by the investigator using Kinget's scoring scale. The investigator had no knowledge when scoring the completed DCT blanks as to which Group Condition each one belonged. Two reliability studies were conducted during this investigation using the Kendall Coefficient of Concordance. The first reliability study involved ten scorers, trained in using the Kinget scoring scale, who had scored 19 com— pleted DCT blanks selected from various studies. The reliability coefficient of this study was .903. The second reliability study involved three of the ten scorers of the first study who had scored 40 completed DCT blanks selected at random and thoroughly intermixed, 10 from each Group Gerald G. Griffin Condition, from the present study. The reliability coeffi- cient of this study was .870. A mean DCT score was found for each Group Condition and a simple analysis of variance was applied to this data. T-tests between all Group Condi- tions were made. The overall difference between the mean DCT scores of the four different Group Conditions was found to be highly significant as was the difference between the mean DCT scores for all possible sets of Group Condition comparisons. It was shown by a simple analysis of variance, using CQT scores as a criterion, that the mean intelligence of the four different groups were statistically equal. The DCT scores obtained from the subjects under each Group Condition were correlated with the subjects CQT scores, and these correlations were statistically compared at the .05 level. The CQT—DCT correlation coefficients for Group Con— ditions I, II, III, and IV were .63, .85, .53, and .17 respectively. The first three were significant, the last was not. The comparisons between the differences of the correlation coefficients of Group Conditions II and I, II and III, II and IV, and I and IV, were statistically signifi— cant. The differences between the correlation coefficients of Group Conditions I and III, and III and IV, were not statistically significant. The specific conclusions of this investigation are: (a) The drawing responses of subjects to the DCT are sig— nificantly affected, both overall and between each possible pair of Group Condition comparisons when the DCT, in addition Gerald G. Griffin to be administered under the Group Condition with instructions inducing neither special motivation or a Special learning set, is administered under the three additional Group Condi— tions of different instruction designed to either induce Special motivation, provide a definite learning set, or bring about both conditions in combination. (b) The DCT has pre- dictive validity for intelligence when administered under the instructional conditions of induced motivation, when administered under the instructional conditions of no induced motivation or special learning set, and when administered under the instructional conditions of a special learning set. '0f these three types of instruction the one producing scoring responses with the highest predictive validity for intelligence is the one inducing special motivation brought about by informing the subjects that the DCT is a new intel- ligence test and that their responses to it will be evaluated. The instruction second highest in this respect is the one which induces neither a special motivation nor provides a special learning set (the control group). The general conclusion of this investigation is that comparison of DCT scores of intelligence among different subjects is reliable only if the DCT has been administered \ to the subjects under the same conditions. Approved by: ééé2£¢aumncz likiiféL- /U’ Tfiesis Chairman PREDICTIVE VALIDITY‘OF INTELLIGENCE FOR THE DRAWING COMPLETION TEST (DCT) UNDER DIFFERENT CONDITIONS 0F INSTRUCTION By Gerald G. Griffin 11> L--l 11‘ m U) H CO Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1963 Drawing Completion Test 11 ACKNOWLEDGMENTS I wish to express my sincere gratitude and appreciation to my Program Chairman, Dr. G. Marian Kinget, for the inter- est, encouragement, and instruction she has expended on my behalf in connection with this study. In addition, I am especially grateful to my statistical advisor, Dr. Robert E. McMichael, for the very helpful assistance and time he has rendered me in the analysis of my study and its data. Also, I wish to express my thanks to Dr° Irwin Kremen, for the consideration he has made available to this study. My Special appreciation goes to my wife, Dolores, for the many drafts of this thesis which she typed, as well as to Mr. Ward Harris and Mr. William Pieper, fellow Graduate Students, for the assistance they rendered me in the collection of data for Group Conditions I and II of this study. iii TABLE OF CONTENTS CHAPTER I. INTRODUCTION II. METHOD III. RESULTS IV. DISCUSSION V. SUMMARY REFERENCES APPENDICES Appendix A-~The Drawing Completion Test (DCT) Blank. . . Appendix B-mKinget s Scoring Scale for the DCT and a List of Stimuli-Attributes. Appendix C-~Computations for the Inter_Scorer Reliability of Ten Trained Scorers for Nineteen Completed DCT Blanks, and for the Inter-Scorer Reliability of Three Trained Scorers for Forty Completed DCT Blanks. Appendix D-MDCT Total Raw Scores for Group Conditions I, II, III, and IV. Appendix E—~Statistical Analysis Computation of the Raw DCT Scores Based Upon a Simple Analysis of Variance Appendix F—~Statistical Analysis Computations of the Raw CQT Scores Based Upon a Simple Analysis of Variance Appendix G—«Statistical Computations for Deter— mining the CQT-DCT Correlation Coef- ficients and Z1 Transformation Scores between the Six Possible Com— parisons of Group Condition Correla- tion Coefficients. iv 35 Al 43 45 A8 51 LIST OF TABLES TABLE ' PAGE 1. Summary Table of the Analysis of Variance of DCT Scores for the Four Group Conditions . . . . . . . . . . . l2 2. Table of Differences of Group Condition Means . . . . . . . . . . . . . l3 3. Summary Table of the Analysis of Variance of CQT Scores for the Four Group Conditions . . . . . . . . . . . 15 A. Comparison of CQT and DCT Group Mean Scores . l6 5. CQT and DCT Correlation Coefficients for the Four Group Conditions. . . . . . . . 16 6. Standard Scores Based on the Z1 Transforma— tion Scores of the Group Condition Correlation Coefficients. . . . . . . 17 CHAPTER I INTRODUCTION The major goal of this study was to determine the predictive validity of intelligence for the Drawing Com- pletion Test (DCT), as a result of correlating DCT scores with the scores of an independent measure of intelligence, under four different conditions of instruction involving the variables of motivation and learning. A DCT blank is exhibited following the title page of this report. This blank consists of eight square fields, 1.5 inches by 1.5 inches, each of which contains different line or line structures which are referred to as stimuli. The square fields are arranged four across and two down.‘ Subjects are asked to take these stimuli and make them into a drawing of anything they wish. The DCT was originally developed by Wartegg (1934) as a clinical tool for investigating person- ality. In the past decade, however, consideration has been given to the DCT as a measure of intelligence. The hypothesis that the DCT could be used as a measure of intelligence originated with Kinget (1952). The use of a drawing test for this purpose was not a novelty originating with the DCT since similar efforts had previously been made by Schuyter (1904), using his own drawing tests on age norms, and by 2 Goodenough (1926) and Berdie (1945), using the Draw—A—Man Test. Jones and Rich (1957) also worked with the Draw-A—Man Test in estimating the intelligence of the aged. The DCT, however, has two distinct advantages over the Goodenough and other drawing tests. First, the interpretation of free drawings has been too unilaterally focused upon content, that is upon the properly projective aspects, while their execution, or expressive aSpect, has remained largely unex- plored. The scoring system Of the DCT, though, deals with execution as well as content variables. Second, the DCT provides for highly diversified standard situations which are absent in other drawing tests. In 1962 Kinget devised an objective scale for measuring a subject's functional level of intelligence using the DCT (Appendix B, page 36). This scale is based upon stimulus integration and drawing organization. Stimulus integration refers to the manner in which the subject responds to the form qualities of the stimuli while drawing organization refers to content and the characteristics of execution. Using this scale, Hayes (1962) found that drawing organization and stimulus integration, as measured by the DCT, correlated .78 with the Wechsler Scale of the California Mental Maturity Test. Kinget's scale is the same one used by the investigator to score the completed DCT blanks obtained in the present study. The College Qualification Tests (CQT) were selected by the investigator as the criterion measure of intelligence 3 with which the DCT would be correlated for purposes of deter— mining its predictive validity for intelligence. The CQT correlates .75 with the California Mental Maturity Test (Office of Evaluation Services, Michigan State University, 1962), .82 with the School and College Ability Tests (College Qualification Tests Manual, 1957), and .78 with the American Council on Education Psychological Examination (College Qualification Test Manual, 1957). One means of accounting for the high correlations among these measures is to postu~ late that a common element of intellectual ability is being measured by all of them. The intelligence which the DCT is asserted to have measured in this Study is this inferred common element of intellectual ability, and is defined and measured in terms of CQT scores. The general questions which the investigator sought to answer as a consequence of the present study were as follows: 1. What happens to the qualities of the drawing responses, as reflected by how they are scored using Kinget s scoring scale, when the DCT is administered under different conditions of instruction? 2. What are the correlations between the DCT and CQT scores (predictive validity for intelligence) under these different conditions of instruction, and are these correlations Significant? 3. Under which condition of instruction would the most reliable drawings of intelligence, or 4 highest predictive validity (highest correlation between DCT and CQT scores), be obtained, and would this correlational value be significantly different from the DCT CQT correlation values obtained for the other Group Conditions? The investigator felt that the two variables of motiva- tion and learning would have significant effects on the drawings to the DCT. It seemed plausible that with induced motivation or knowledge about how intelligence is measured from drawings (a definite learned set), or both, a subject would score higher on the DCT than he would in the absence of such induced motivation or learned set, To test this assertion the investigator set up three different Group Con~ ditions under which the DCT would be administered, each with a different set of instructions designed to either induce special motivation (Group Condition II), induce special learning (Group Condition III), or both (Group Condition IV). These three Group Conditions were compared to a control group condition (Group Condition 1) that provided instructions which induced no special motivation or learned set aside from what would be expected to exist when groups are asked to reSpond to the DCT without any knowledge of the nature of the test or its use. The general expectations which the investigator thought would be confirmed by this study were as follows: 1. The drawing scores of subjects to the DCT would be significantly affected by the instructions of ID 5 the Group Condition under which the DCT was administered. More Specifically, the drawing scores of subjects under Group Condition II would be significantly higher than the drawing scores of subjects under Group Condition I; the drawing scores of subjects under Group Condition III would be significantly higher than the drawing scores of subjects under either Group Condition I or II; and the drawing scores of subjects under Group Condition IV would be significantly higher than the drawing scores of the subjects under either Group Condition I, II, or III. The DCT would correlate significantly with the CQT under Group Conditions I and II. Group Condition II would product more reliable drawings of intelligence than Group Condition I, and Group Conditions III and IV would produce less reliable drawings of intelligence than Group Condition I, with Group Condition IV producing the least reliable of the two. CHAPTER II METHOD Subjects 5 The subjects for this study were 120 beginning psy~ chology students who were members of four separate intro~ ductory psychology classes. These classes were selected at random from the thirty two which were taught at Michigan b State University in the Spring of 1962, and these four classes were randomly assigned to the four different Group Conditions of the experiment. There were approximately thirty students in each of the classes, of which approxim mately one~half were females. These students were available as subjects as a partial fulfillment of their course requirem ments. The experiment was conducted in the subjects‘ regular classrooms during their regular scheduled class hours for introductory psychology. Materials The test materials involved in this experiment were a pencil and a Drawing Completion Test blank. A copy of the test blank, with illustrated responses, is exhibited in Appendix A, page 34. Procedure The instructions of Group Condition I (the control group) were not designed to induce any Special motivation or to provide any special learning set. These instructions were given to the group by the experimenter and were as follows: We are asking you to take this drawing test to help us with research we are doing, This test will not affect your grade in this class or in any way affect your university standing. Your instructor will not see the test results. Since this is research we may want to contact some of you later. For this reason will you put your name and student number on the back of the test blank. Also on the back, will you please put the number 1 in the upper right hand corner. 0n the front of the test blank you will find eight squares. In each square you will find a stimulus. What we want you to do is to take the stimulus in each of the eight squares and make it into a drawing of anything you wish. You Should have eight drawings when you are through. "You can respond to the stimuli in any order you wish, only please number the squares in the order that they are responded to. You have approximately thirty minutes to complete the drawings. Please do not look at your neighbor’s paper but give only your own responses. The instructions of Group Condition II were designed to induce Special motivation but were not designed to provide a Special learning set. These instructions consisted prim~ arily of informing the subjects that the drawing test they were taking was an intelligence test. The instructions given by the experimenter were as follows; We are asking you to take this drawing test to help us with research we are doing with a new type of test. Since this is research we may want to contact some of you later. For this reason will you put your name and student number on the back of the test blank. Also on the back, will you please put the number 2 in the upper right hand corner. Now, this test is a new type of intelligence test. It is important that you do your best on this test for the scores are going to be recorded 8 as well as compared to your College Qualification Test scores. These scores will be used for future evaluations. '0n the front of the test blank you will find eight squares. In each square you will find a stimulus. What we want you to do is to take the stimulus in each of the eight squares and make it into a drawing of anything you wish. You should have eight drawings when you are through. You can respond to the stimuli in any order you wish only please number the squares in the order that you respond to them. You have approximately thirty minutes to complete the drawings. Please do not I look at your neighbor's paper but give only your own ‘ responses. Group Condition III had a special type of learning set induced by a twenty-five minute lecture on, ”The Meas- urement of Intelligence from Drawings.” This lecture ; emphasized stimulus integration and drawing organization as ry criteria used in measuring functional intelli— c”? ’J (D *d H [.4 .3 f ') gence from drawings. The lecture discussion on stimulus integration centered around the following topics, with drawing illustrations of each given on the blackboard by the experimenter: . Relation of the stimulus to the completed drawing . 'Criginality of the stimulus integration Functional significance of the stimulus in the drawing Attributes of the stimulus responded to t— U)MF4 a. rounded, wavy stimuli normally bring responses of movement and life b. straight, angular stimuli normally bring technical—conStruction and mechanical type response The lecture discussion on drawing organization centered around the following topics, with drawing illustrations of each given on the blackboard by the experimenter: l. Context of the picture, its detail, and its total affect 2. Perspective~~representation of third dimension 9 3. Static and Dynamic aspects;absence or presence of life and movement 4. '0riginality of the entire drawing After the lecture the experimenter presented the DCT blanks and gave the same instructions to this group that were given to Group Condition I, except that group III sub- jects were told to put the number 3 on the back of their test blanks rather than the number 1. This, then, consti— tuted the condition of learned set but no induced motivation. Group Condition IV was given both the instructions for providing a special learning set (as given to Group Condition III) and the instructions for inducing special motivation (as given to Group Condition II). The same twenty~five minute lecture\whicn was given to Group Condi- \ tion III to provide them with a special type of learning set was also given to Group Condition IV. Following this lecture Group Condition IV was given the same instructions as Group Condition II to induce special motivation, except that group IV subjects were told to put the number 4 on the back of their test blank rather than the number 2. After the groups had completed their drawing tests, the tests were collected, all intermixed together by Group Condition, and scored according to the objective scoring scale devised by Kinget (see Appendix B, page IRS. The scoring of each test blank was done by the author who had been trained in scoring the completed DCT blanks by use of the Kinget scoring scale. The author had no knowledge, while he was scoring the tests, of which tests belonged to 10 which group (each group’s test blanks were identified on the back by the code 1, 2, 3, or 4). Each of the drawing responses to the eight squares of the test blank had a possible 6 points or 48 points per test blank. The Kendall Coefficient of Concordance inter-scorer reliability of ten scorers who had been trained in scoring completed DCT blanks using Kinget's scoring scale, based upon l9 completed DCT blanks not taken from this study, was .903 (see Appendix 0, page 42). The Kendall Coefficient of Concordance inter— scorer reliability of two trained scorers and the investi- gatoris original scoring of 40 completed and intermixed DCT blanks taken from this study (10 from each Group Condition "'1 selected randomly) was 6870 (see Appendix C, page 42). A mean score was found for each Group Condition and a simple analysis of variance was computed to determine if there was a statistically significant difference between the means. The .05 level of significance was chosen for this test. If a significant Fmvalue was obtained it was decided to do tthSCS based on the Q-distribution to see where the significant differences between Group Conditions existed. The .05 level of significance was again used for these tests. The CQT scores were used as an independent measure of intelligence. To determine if the DCT drawing scores were actually predicting intelligence the DCT scores obtained from the subjects under each Group Condition were correlated with the subjects CQT scores. Also, to determine ll which Group Condition produced the most reliable scores of intelligence the Pearson Product Moment Correlation made between the DCT and CQT scores for each Group Condition were statistically compared. The testing of correlational sig— nificance and significant differences between correlations was done at the .05 level. Since the direction of change to take place was predicted beforehand, that is since it was predicted that the DCT-CQT correlation of Group Condi— tion II would be greater than Group Condition I, and that Group Conditions III and IV would be less than Group Condi- tion I, a one-tailed test was used in testing for signifi- cant differences between these correlations. CHAPTER III RESULTS Statistical Analysis of the DCT and CQT Scores The raw DCT scores for the students in each of the four Group Conditions are presented in Appendix D, page 44. The results of the F—max test applied to the appropriate variances of this data was 1.64. This was not significant at the .05 level and therefore the hypothesis of homogeneity of variance was acceptable. A simple analysis of variance (Appendix E, page 47) was applied to these scores to deter- mine if the difference between the means across the four Group Conditions was statistically significant. Table 1 presents a summary of this analysis. TABLE l.—-Summary Table of the Analysis of Variance of DCT Scores for the Four Group Conditions Source of Significance Variation df SS MS F at .05 Level Group Conditions (GC) 3 4148.2 1382.7 42.3*** 2.68 Within Group Conditions 116 3796.0 32.7 Total 119 7944.2 ***Significant at the .001 level. 12 13 Clearly the Group Condition under which the subjects took the DCT in this study did influence their performance on the test. Thus the general expectation that the drawing scores of subjects to the DCT would be significantly affected by the Group Conditions under which the DCT was administered is supported. Next, a statistical comparison of differences between I means for all possible comparisons of the four different Group Conditions was made. To do this the Critical Differ— ence (CD) was computed (see Appendix E, page 47). The CD value at the .05 level is 3.87 and at the .01 level it is 4.7. Referring to Table 2 it will be observed that signifi- cant differences exist between all possible comparisons of Group Condition means. Therefore, the Specific TABLE 2.-—Table of Differences of Group Condition Means Group Means 24.4 30.1 35.4 40.1 Iii Xi: XIII xiv 24.4 xi 30'1 XII 5.7** 35'“ 'XIII ll.0** 5.3** “0'1 7iv 15.7** 10 O** “-7** **Significant at the .01 level. expectations that (l) the drawing scores of subjects under Group II would be significantly higher than the drawing scores of subjects under Group Condition I; (2) the l4 drawing scores of subjects under Group Condition III would be significantly higher than the drawing scores of subjects under either Group Condition I or II; and (3) the drawing scores of subjects under Group Condition IV would be sig- nificantly higher than the drawing scores of the subjects under either Group Condition I, II, or III, are all sup- ported. The raw CQT scores for each of the students of the 3 four different classes used in this study were obtained from the Student Services Counseling Center at Michigan State University (Appendix F, page 49). Group Condition I, the control group, had a mean CQT score of 133.2. Group Condition II, the one with instructions to induce special motivation but no special type of learning set, had a mean CQT score of 131.5. Group Condition III, the one with instructions to provide a special type of learning set but no special induced motivation, had a mean CQT score of 130.0. Group IV, the one with instructions designed to induce special motivation and provide a special type of learning set, had a mean CQT score of 134.9. These scores were all near the fiftieth percentile of the distribution on which the CQT norms were based. To show that the differences disclosed in the previous DCT analysis of variance were due to the instructions of each Group Condition, rather than to initial differences among the four groups associated with differences in their CQT scores, a simple analysis of variance was applied to the raw CQT scores (see Appendix F, page 50). 15 I Table 3 presents the summary of this analysis. TABLE 3.——Summary Table of the Analysis of Variance of CQT Scores for the Four Group Conditions Source of . Significance Variation df SS MS F at .05 Level Between Group Condition . (cc) 3 379.1 126 4 .193 2.68 Within Group Condition 3 _ (w) 102 66,122.9 654.3 Total 105 67,122.0 Clearly there is no difference in the average intelli- gence of the subjects who participated under each of the four different Group Conditions. The analysis of the CQT and DCT scores for the four Group Conditions is summarized in Table 4.' This table interestingly compares the two sets of scores. The drawing responses of the subjects, as reflected by each Group Con— dition's mean DCT score, vary significantly with the instruc~ tions of the specific Group Condition under which the DCT is administered, while the intelligence of the groups (as evi- denced by each Group Condition‘s mean CQT score) is statis- tically the same. 16 TABLE 4.— Comparison of CQT and DCT Group Mean Scores Group Condition I II III IV CQT Mean Scores 133.2 131.5' 130.0 134.9 DCT Mean Scores 24.4 30.1 35.4 40.1 CQT and DCT Correlations for the Group Conditions Table 5 presents the correlation coefficients for the sets of CQT and DCT scores for each of the four Group Condi- tions (the computations for these coefficients are presented in Appendix G, page 51)° TABLE 5.m—CQT and DCT Correlation Coefficients for the Four Group Conditions Group Condition Correlation Coefficient I .63 II .85 III .53 IV .17 The correlation coefficients for Group Condition I, II, and III are statistically significant at the .05 level while the correlation coefficient for Group Condition IV is not statis- tically significant at the .05 level (see Appendix G, page 51). Thus the general expectation that the DCT correlates signifi- cantly with the CQT under Group Conditions I and II is 17 supported. Table 6 presents all the possible comparisons, using 21 transformation scores (see computations of Appendix G, page 50), between the four pairs of Group Condition cor- relation coefficients. TABLE 6.—-Standard Scores Based on the 21 Trans- formation Scores of the Group Condition Correlation Coefficients Correlation Coefficient .63 .85 .53 .17 r1 r’11: rIII rIv .63 r1 .85 rII —1.71* .53 r111 .50 2.28* .17 rIV 1.93* 3.84** 1.49 *Significant-at .05 level. **Significant at .01 level. The correlation coefficients presented in Table 5 support the validity of the DCT, under Group Conditions I, II, and III, as a predictor of intelligence when intelligence is defined by CQT scores. The higher the correlation coef- ficient between any set of CQT and DCT scores the more reliable is the Group Condition of that set for measuring intelligence using the DCT. Conversely, the lower this coefficient the less reliable is the Group Condition for measuring intelligence using the DCT. Referring to Table 5 it will be observed that the Group Condition with the 18 highest correlation coefficient is Group Condition II. Table 6 reveals that the difference between the correlation coefficient of Group Condition II and the control group (Group Condition I) is a statistically significant differ_ ence at the .05 level using a one—tailed test. Thus the general expectation that Group Condition II would produce more reliable drawings of intelligence than Group Condition I I is supported. Table 5 also discloses that Group Condi- tion IV has the lowest correlation coefficient, with Group Condition III having the next lowest coefficient. Table 6 reveals that the difference between Group Condition IV and the control condition is significant but that the difference between Group Condition III and the control condition is not significant. Thus the general expectation that Group Conditions III and IV would produce less reliable drawings of intelligence than Group Condition I is supported. How— ever, the expectation that Group Condition IV would produce less reliable drawings of intelligence than Group Condition III is not supported, although by referring to Table 6 it can be seen that the necessary significance (standard score of 1.645) to give this expectation statistical support is almost reached. CHAPTER IV DISCUSSION The interpretation to be presented in this section is based upon the investigator‘s belief that the causality for the significant differences found in mean group responses and group correlations is due primarily to the differences in instruction for each of the Group Conditions and not to artifacts or other operating variables existing in the Group Conditions which were unknown to the investigator. These latter possibilities should be and have been considered, but the investigator feels that the available data and informa— stigation favor an interpretation ) in terms of motivational and learning variables induced by tion concerning the inve special instructions. 'One of the major assumptions made in this study was that, in addition to the parameter of intelligence, all other primary parameters of variation among subjects and Group Conditions of consequence to this study, save the instructional variables of induced motivation and learned set, were being controlled by the experimental design. Actually, the variables of induced motivation and learning set were assumed to have varied among the Group Conditions while what was being varied were instructions. The investi- gator assumed that these instructions were sufficient to 19 20 vary motivation and learning set. With these assumptions in mind, and in view of the results of this statistical analysis, the most plausible deduction to be derived from the comparison made by Table 4, and one which the data supports, is that the absence or presence of the independent variables (instruc- tions designed to induce motivation and learning set) existing either independently or in combination as character- ized by each Group Condition, are the most probable signifi- cant factors influencing the subject's drawing response in each case. The Special induced motivation (that which was assumed to have existed in Group Condition II) referred to in this study is the desire or need (assumed a priori) on the part of the subject to do his serious best with the task at hand, namely, that of responding apprOpriately and dilligently to the DCT in the manner he feels is designed to make the most favorable impression upon the scorer. The instructions for the control condition (the one which existed for Group Cone dition I) failed to induce (uniformily at least) this particular desire or need. In contrast, the instructions given to the subjects of Group Condition II were designed to bring about this desire or need. The difference of this in- duced motivation was sufficient enough to produce a signifi— cant difference in the DCT intelligence scores of these two equally intelligent groups of subjects. If a learning session is given to another equally intelligent group of subjects (as was the case with Group 21 Condition 111), which illustrates for them the type of drawing responses that are scored high for intelligence, but at the same time is assumed to do nothing to induce special motivation, this group's DCT intelligence scores are significantly higher than the group which was given instructions designed to induce special motivation only. '0ne plausible explanation for this, and one which the in- omq vestigator feels appropriate, is that this latter condition provides a definite means to a goal (if this goal happens to be clearly discriminated by the subject), and this means “T (a definite knowledge of what type of drawings are regarded as depicting high intelligence) has more of an effect on a subject‘s drawing responses than does a special desire or need operating within him to make highly intelligent drawings in the absence of any definite knowledge of how to do so. In other words, although a special desire or need (induced motivation) is effective in producing responses to the DCT which are scored significantly more highly intelli- gent than responses made to the DCT under conditions where this desire or need is lacking, an awareness of the type of responses which are scored high for intelligence (even though it may not be recognized that an intelligence test is involved), results in responses to the DCT which are rated even more significantly higher in intelligence than either of the other two conditions. The means alone accomplishes more than either no motivation or just motivation alone. Group Condition IV produced the responses to the DCT which 22 were rated significantly higher in intelligence than any of the other three Group Conditions because the condition of this group not only specified definitely what the goal was, and provided the desire or need to accomplish this goal, but it also provided the means (direction) by which this goal could be reached. So far we have been discussing how different instruc- tions influence the nature of a subject‘s drawing responses to the DCT with respect to how these responses will be scored for intelligence. No mention has been made up to this point, however, as to how well the intelligence scores obtained on the DCT match the level of functioning intelli- gence of the individual as revealed by an independent measure of intelligence such as the CQT, under each of these different conditions of instruction. The correlational analysis shows that of the four Group Conditions under which the DCT were administered, the condition of induced motiva- tion alone (Group Condition II) produced responses which were scored the most reliable in reflecting the subject's functional level of intelligence as measured by his CQT score. The condition which was second best as being reliable in reflecting a subject‘s functional level of intelligence was the control condition (Group Condition I), the instruc- tions of which provided no Special induced motivation or Special type\of learning set. The condition which was third most reliable in reflecting a subject's functional level of intelligahce was the one in which the instructions provided 23 a special type of learning set concerning drawing responses (Group Condition III). The least reliable condition was the one involving instructions which were designed to both induce special motivation and provide a special type of learning set (Group Condition IV). The question arises as to what are the operating factors which cause the instructions of one condition to produce responses which are more or less reliable in pre- dicting intelligance than the instructions of the other conditions? At this state the investigator favors the following explanation: Taking the DCT under conditions of instruction designed to induce motivation alone results in the most reliable re- sponses of intelligence because these are the instructions which produce drawings by the subjects that are characterized more by homogeneity and uniformity of group effort toward a common goal under conditions which minimize (it is assumed) interfering influences which tend to obscure one’s level of functioning intelligence, while at the same time it maximizes one's desire to perform maximally in the best manner he is capable. Such an effort on the part of the subjects is brought about first by the fact that the subjects are near a maximum in their desire to perform in a serious manner, thus diminishing considerably the possibility of distractions or ' and second by the tendency in some subjects to ”goof-off,’ the fact that no known factors are operating to determine the typa of responses the subject will make except his own 24 functional level of intelligence. The condition of no induced motivation or special learning set also has this second fact operating (and apparently a very important one), but in the absence of a desire or need to do one's best, distractions and "goofing-off” interfer with a subject seriously and fully bringing his functional level of intelligence to bear on the drawings he is making. In the pilot study preceeding the present study the investigator observed this difference in responding between two such groups of subjects taking the DCT, one under the condition of no induced motivation or special learning set and the other under the condition of induced motivation. Moving on to the variable of learning it was seen from the analysis of the results that even though the condition of a learned set on how to make draw1ngs to a drawing test, which will be scored high in intelligence, produces drawings which receive a significantly higher mean score than either of the conditions of no induced motivation or learned set, and motivation alone, the scores produced by this variable are less reliable in indicating the true functional level of a subject's intelligence than the other two conditions because a definite factor exists which obscures the operation of the subject‘s functional level of intelligence in making his responses to the DCT. Instead of the drawings reflecting the subject’s natural level of functioning intelligence they are reflecting more the influence of his previous learning concerning the type of drawing responses which are scored 95 highly intelligent. This obscuring influence is present regardless of whether the subject (in Group Condition III) related this learning directly to the DCT and realized that this was one of the drawing tests from which intelligence is measured. Thus the drawing responses produced by Group Condition III reflect less of the subject's natural level of intellectual functioning and more of special learning El influences.- Instead of the subject's responses reflecting his functional level of intelligence then, as was the case 1 for the subjects in Group Conditions I and II, they are g reflecting to a large degree the influences of his special learning. In this case the scorer is scoring more of a learning influence and less of an intellectual influence. For this reason the CQT~DCT correlation for this group is lower than that of Group Conditions I and II. 0ne objection to the foregoing interpretation might ce that the more intelligent subjects Should learn more than less intelligent ones, and consequently, this additional learning Should be'detected in the responses of the more intelligent suv‘ccts, thus enabling them to maintain their prOper relatiorsxip with respect to the less intelligent subjects. If this were the case then the DCT—CQT correla— tions for Group III should be little different from that of Group Condition I. This would be a valid objection if the learning of the material presented by the instructions of Group III was highly difficult so that the more intelligent subjects could be differentiated from the less intelligent 26 ones on the basis of what they learned. In a sense this expectation might appear to receive some support Since the difference in the correlation coefficients of Group Condi~ tion I and III was not statistically significant. However, the investigator does not feel that much unequal learning took place. The material presented for learning was uncom- plicated enough so that all members of the class could learn it equally as well if they were attentive at all during the lecture. Therefore, the learning could have easily been almost equally the same for all subjects. In dealing with the condition of both induced motiva- tion and a Special type of learning set (Group Condition IV) we have the same Obscuring influence of learning over— shadowing the distribution or range of functional intelli— gence in this group. But now this influence is much more marked because the subjects are told definitely that the DCT is an intelligence test (there can be no question about it as was the case in Group Condition III), so that in addi— tion to their learning set directing them as to how to produce highly intelligent type drawings they also have induced in them the desire or need to produce these drawings more so than existed in Group Condition III. As a result the subjects concentration on the instructions and their learning concerning these drawings is much keener and goal orientated than was the case in Group Condition III. AS a consequence of all of this the subjects natural levels of intellectual functioning are evidenced even less in their 27 reSponses under this Group Condition. The near equal learning hypothesis advocated by the investigator to the possible objection discussed in the previous paragraph applies even better here since the difference in the corre- lational coefficients between Group COnditionS I and IV is significant. It appears that the addition of the induced motivation brings the learning influence more sharply to bear in the subjects' responses than was the case in Group Condition III where just the\learning influence alone existed. Again, it is the influences of Special learning, sharpened by the effects of special motivation, which is being scored in these responses, and less of the influence of the subject s natural level of functioning intelligence. As a result Group Condition IV produces drawings to the DCT which have the highest mean score of intelligence of the four Group Conditions employed in the study while at the same time the scores for these drawings tend to be the least reliable in predicting intelligence as measured by CQT SCOI‘GS . CHAPTER V SUMMARY The primary concern of this investigation may be stated as follows: (1) To determine the effect four dif— ferent Group Conditions, each one presenting different instructions designed to either induce no Special motivation or learning set, induce Special motivation only, induce a special learning set only, or induce both a Special motiva~ tion and a special learning set, would have upon Drawing Completion Test (DCT) scores. (2) To determine the rela- tionship or predictive validity of intelligence (as deter— mined by correlation coefficients) between the DCT scores obtained under each of the four different Group Conditions and the scores of an independent measure of intelligence, the College Qualification Tests (CQT). To accomplish these goals four classes of beginning psychology students, con~ sisting of approximately 30 students each, were each assigned to one of the four Group Conditions. The tests were administered, collected, and scored using Kinget‘s scoring scale. Two separate reliability studies (the latter study of which was based on 40 randomly selected DCT blanks from the present investigation) produced Kendall inter—scorer reliability coefficients of .903 and .870 for Kinget’s scoring scale when applied to completed DCT blanks. The 28 J 29 proper statistical analysis, based upon analysis of variance and computation and comparison of correlation coefficients, was carried out. Based upon the results of the statistical analysis the following Specific conclusions to this investigation are made as follows: (a) The drawing reSponses of subjects to the DCT are significantly affected, both overall and between each possible pair of Group Condition comparisons, when the DCT, in addition to being administered under the Group Condition with instructions inducing neither special motivation or a special learning set, is administered under the three additional Group Conditions of different instruction designed to either induce special motivation, provide a definite learning set, or bring about both cone ditions in combination. (b) The DCT has predictive validity for intelligence when administered under Group Condition I (instructional conditions inducing no special motivation or learning set), when administered under Group Condition II (the instructional conditions of special induced motivation), and when administered under Group Condition Ill (the instruc- tional conditions of a Special learning set). Of these three types of instruction the one producing scoring responses with the highest predictive validity for intelli— gence is the one inducing Special motivation brought about by informing the subjects that the DCT is a new intelligence test and that their responses to it will be evaluated. The instruction second highest in this respect is the one which 30 induces neither a special motivation nor provides for special learning set (the control group). The general conclusion of this investigation is that comparison of DCT scores of intelligence among different subjects is reliable only if the CT has been administered to the subjects under the same conditions. ‘E m REFERENCES Berdie, R. F. "Measurement of Adult Intelligence by Drawings," g, Clin. Psychol., 1945, 1, 288-295. College Qualification Tests Manual. New York: The Psy- chological Corporation, 1957. Cooke, E. Art Teaching and Child Nature. London: J. Education, 1885. Goodenough, F. L. Measurement of_1nte11igence by_Drawings. New York: World Book Co., 1926. Griffin, G. G. ”Pilot Study.” Mimeographed c0py available from author. Hayes, F. J. "Correlation of the Wartegg Drawing Comple- tion Test with the California Mental Maturity Scale." Unpublished thesis, 1962. Jones, A. W. and Rich, T. A. ”The Goodenough Draw-A-Man Test as a Measurement of Intelligence in Aged Adults,“ J. Consulting Psychol., 1957, 21, 235-238. Kinget, G. M. The Drawing Completion Test. New York: Grune and Stratton, Inc., 1952. Office of Evaluation Services Report, Michigan State Univer~ sity, East Lansing, Michigan, 1962. Schuyter, M. C. cited in F. L. Goodenough, Measurement of_ Intelligence by Drawings. New York: World Book Co., 192e, '“_ Wartegg, E. Gefuhl Neue Psychol. Stugies 12 99 (1934) cited in G. M. Kinget, The Drawing Completion Test. New York: Grune and Stratton, Inc., 1952. West, J. H. ”Correlation of the Draw-A—Scens—Test," J, Elin. Psychol., 1960, 16, 44-45. 31 A PPENDICES 32 APPENDIX A THE DRAWING COMPLETION TEST 33 (DCT) BLANK Figure l.--Drawing Completion Blank with Illustrated Responses APPENDIX B KINGET’S SCORING SCALE FOR THE DCT AND A LIST‘OF STIMUleATTRIBUTES 3.5 36 Kinget's Scoring Scale for the DCT Corresponding Response Category Response Score Complete disregard of stimulus 0 Near-complete disregard of stimulus I 1 Relative disregard of stimulus 2 Minimum integration of stimulus 3 Good integration of stimulus 4 Very good integration of stimulus 5 Outstanding integration of stimulus 6 Operational Definitions of the Response Categories 1. Complete disregard_of stimulus: Random strokes, scribbles or drawings without any relation to the stimulus. No dimensionality context, detail, or perspective (perSpective refers to the objects in the drawing being presented in their prOper relations to one another). For illustration see Fig. 1—A, page 34. 2. Near-complete disregard for the stimulus: Lines or line structures which reproduce or evoke some attribute (see page 40) of the stimulus without inte— grating or connecting the stimulus (e.g. rudimentary duplication). Very little dimensionality context, detail, and no perspective. For illustration see Fig. 1—B, page 34. 37 Relative disregard of_stimulus: Attempt at integration or, at least, connection of the stimulus with a themq which is representational but incompatible with any of the attributes (see page 40) of the stimulus. Poor dimensionality context, detail, and no perspective. For illustration see Fig. l-C, page 34. Minimum integration_of stimulus: Stimulus is an intrinsic part of a representational drawing which brings out at least_ong of the attributes (see page 40) of the stimulus. That is, at least one stimulus quality is responded to such that the stimulus) is integrated into a drawing in some meaningful manner.% There is two dimensionality, but no perspective. There is adequate proportion and detail. For illustration see Fig. l-D, page 34. Good integration of stimulus: Two or more attributes (see page 40) of the stimulus are brought out in the drawing. The stimulus is given good functional prominence. The function of the stimulus is clear and its integration is quite meaningful. In addition, the drawing has one or more of the following characteristics: (a) appropriate context. Context refers to everything surrounding the object constructed from the given stimuli. (b) adequate detailing (sufficient detailing to give the drawing clear and full meaning). 38 (c) adequate organization (plane or depth). This is a combination of two-dimensional and threeedimensional responses usually suggesting life, mechanical form or abstraction. The drawing presents some perspec— tive, adequate proportion, and detail. For illustration see Fig. l-E, page 34. Very good integration: Two or more stimulus attributes (see page 40) are responded to and the stimulus attributes are very wgll integrated in the drawing. That is, the stimulus integration is appropriate to the stimulus attribute and represents good form and meaning. Organic responses (responses dealing with living objects such as people or animals) are made to the curved stimuli of drawings 1, 2, 7, and 8 while technical-mechanical responses are made to drawings 3, 4, 5, and 6. In addition the drawing Shows one or more of the following characteristics: (a) apprOpriate context. (b) Pregnant, economic representation. The response is a very full presentation of life, mechanical move» ment, or abstraction, and contains good detail. (c) High form niveau. (i.e. harmony between context and execution). This refers to depth drawing and means that the response is primarily in three dimension and has proportion plus linear perspective. For illustration see Fig. 14F, page 34. Outstanding integration of the stimulus: All of the characteristic outlined under the very good integration category above plus either of the following: (a) originality (unique or rare drawing) 39 (b) marked degree of Physiognomy (everything in the drawing of a human figure that reveals something about the person represented: his age, profession, or character), of Style (manner in which objects are made that conveys something about the object), of Atmosphere (pertains to the emotional quality indirectly expressed by the drawing, especially by landscapes or situations other than human scenes), or of (recognizable) Symbolism. For illustration see Fig. l—G, page 34. 40 List of Stimuli-Attributes for the Drawing Completion Test 'OBJECTIVE , PROJECTIVE ‘/ Attributes Attributes if 1. central, round, delicate, light, Small- small drifting, unim- delicate posing, 'easy" 7. tightly structured, delicate, "precious," half—circular fancy, fuzzy, "com- CURVED \ plicated," "difficult, " ‘ORGANIC r/ prickly (female) ‘ 2. symmetrically— flexible, loose, sinuous, round— alive, soft, waving, edged flowing, crawling, lgrge- moving that 8. large, flat-curved, smooth, fluent, \ taperedmoff edges "vast,” restful, "natural,” sheltering \ (cover—like) r3. structured, equi- rigid, strict, pro- distant, regularly gressive, methodical, F increasing mounting, construcu tive Dynamic —' 5. structured,straight, ”Shooting," pointing, Opposing, diagonally conflictuous, mechan— :> oriented ical—constructive STRAIGHT - MECHANICAL 4. square, black solid, heavy, somber, (male) static, material Static 6. structured, unequal mechanical—construc- length and irregu- tive, object-like, \ 1dr distribution matter-of~fact, plain, static APPENDIX C COMPUTATIONS FOR THE INTER-SCORER RELIABILITY OF TEN TRAINED SCORERS FOR NINETEEN COMPLETED DCT BLANKS, AND FOR THE INTER—SCORER RELIABILITY‘OF THREE TRAINED SCORERS FOR FORTY COMPLETED DCT BLANKS 41 42 Inter—scorer Reliability Using the Kendall Coefficient of Concordance: ;W A. W for Nineteen Completed DCT Blanks, from Various Sources, which were Scored by Ten Trained Scorers 33 51.254 w = 1/12 k (N‘— N) = 1/12 (10) (19 — 197 = 513254 __ = ' 51.0254 = .083 (100) (6859—19) 8.3 (6840) 51:25“ = .903 59,772 B. W for Forty Completed DCT Blanks, Selected at Random from the Present Study, which were Scored by Three Trained Scorers w : SS = 41,749.91 = 1/12 k (N —N) 1/12 (3) (40 -40) 41,749.91 : .870 1/12 (9) (64,000—40) "‘“‘ 1r! APPENDIX D DCT TOTAL RAW SCORES FOR GROUP CONDITIONS I, 11, III, AND 1v 43 44 DCT Total Raw Scores Group Condition Subject I II - III IV 1 32 24 32 40 2 3o 43 3o 41 3 25 21 28 28 4 o 26 32 42 5 17 37 39 39 6 3O 33 32 33 7 23 23 36 39 8 17 35 34 44 9 25 28 44 37 10 20 26 32 37 11 24 31 43 45 12 28 34 34 38 13 23 26 30 43 14 21 24 42 33 15 22 34 44 29 16 25 30 37 42 17 25 26 34 42 18 31 43 25 47 19 19 29 34 ‘ 42 20. 29 31- 38 34 21 18 30 44 47 22 31 25 34 47 23 25 26 27 43 24 29 32 39 39 25 23 35 42 47 26 31 37 39 42 27 32 29 31 38 28 33 3O 33 46 29 2o 25 31 44 3O 23 3O 43 3 Total 731 903 1063 1203 'x 24.4 x' 30.1 x' 35.4 'x 40.1 APPENDIX.E STATISTICAL ANALYSIS COMPUTATIONS'OF THE RAW DCT SCORES BASED UPON A SIMPLE ANALYSIS OF VARIANCE 46 Mean DCT Scores for the Students Participating Under the Four Group Conditions 120 Subject Group I Group II Group III Group IV 1 32 24 32 4o 2 30 43 ' 3o 41 3 25 21 28 28 4 o 26 32 42 5 17 37 39 39 6 3o 33 32 33 l 7 23 23 36 39 - 8 17 35 34 44 9 25 28 44 37 10 2o 26 32 37 11 24 31 43 45 , 12 28 34 34 38 13 23 26 3o 43 14 21 24 42 33 15 22 34 44 29 16 25 30 37 42 17 25 26 34 42 18 31 43 25 47 19 19 29 34 42 20 29 31 38 34 21 18 30 44 47 22 31 25 34 47 23 25 26 27 43 24 29 32 39 39 25 23 35 42 - 47 26 31 37 39 42 27 32 29 31 38 28 33 30 33 46 29 20 25 31 44 3O 23 3O 43 35 nJ 30 30 30 3o N=l20 TJ 731 903 1063 1203 T=39OO 'x 24.4 30.1 35.4 40.1 2 d‘ - 1279 871 865 781 2 id M = 44.10 30.03 29.83 26.93 2 z'xe 19,091 28,051 38,531 49,021ffx =134,694 T32 534,361 814,506 1,129,969 1,447,209ZTJ2 = 3,926,045 Z'T 2 = 3:926:045 = 13o,898.2 J/n 30 T2 (3900)2 = 15,210,000; Tfin = 15,210,000 = 126,750.0 47 Statistical analysis to test the null hypothesis that the means of these four groups are equal: 2 311. T2 SSGC nj 'TN“ 130,898.2 — 126,750.0 = 4,148.2 2 T SST SZXE — 77 7,944 SS = SS SS 3,796.0 T - C Source of Significance Variation df SS MS F at .05 Level Group Conditions (90) 3 4,148 2 1382.7 42.3 . 2.68 Within Group Conditions (w) 116 3,796.0 33.0 Total 119 F-Max Test to Test for Homogeneity of Variance Among the DCT Scores 2 2 2 Ho‘ (:cxn: =‘ (:0011 = ()GCIII =(;GCIV 2 Fmax = 82 max = 44.10 = 1.637 S2 min 25-93 With k = 4 and n = 30 the Significant F-max value at the .05 level = 2.61. Therefore, 1.637 is not significant and the hypothesis that the variances of the four Group Conditions are the same is acceptable. Critical Difference (CD) CD = 0,: 1¥§£— = 3.69 -§%59 = 3.69‘11.10 = 3.69 x 1.05:3.87 APPENDIX E STATISTICAL ANALYSIS COMPUTATIONS‘OF THE RAW CQT SCORES BASED UPON A SIMPLE ANALYSIS'OF VARIANCE 48 49 Raw CQT Scores for the Students Participating Under the Four Group Conditions Group I Group II Group III Group IV Subject CQT Subject CQT Subject CQT Subject CQT 2 156 1 102 1 109 1 167 3 97 2 168 2 86 2 157 5 100 4 106 4 99 3 111 6 151 6 118 5 116 4 164 10 124 7 99 6 104 5 149 11 112 8 177 7 147 6 110 12 153 9 140 10 9o 7 95 14 148 11 129 11 166 8 143 15 148 12 142 12 112 9 134 16 144 13 110 13 140 10 138 17 122 14 93 14 151 11 133 18 163 15 169 15 184 12 111 20 135 16 150 16 113 13 132 21 140 17 109 17 137 14 110 22 143 18 173 18 133 15 159 23 135 19 130 19 109 16 122 24 152 20 132 20 112 17 149 25 93 21 143 21 154 18 128 26 142 22 129 22 128 19 120 27 160 23 121 23 91 20 140 28 175 24 132 24 149 21 179 29 99 25 172 25 141 22 136 30 72 27 106 26 166 23 164 28 140 27 148 24 108 29 105 28 92 25 115 30 124 29 184 26 136 30 148 27 135 28 93 29 167 30 141 Total 3064 3419 3509 4046 g_ 23 26 27 30 x 133.2 131.5 130.0 134.9 nJ 23 26 27 30 N=lO6 Tg 3064 3419 3509 4046 T=14,038 5.x 423,498 465,243 477,215 560,276 £{x2=1,926,232 MJ 133.2 131.5 130.0 134.9 M-l32.9 TJ 9,388 096 11,689,561 12,313,081 16,370,116 2 T2/nj 4023,1731 449,598.5 456,040.0 545,670.5 {331 = J 1.859.487.1 50 Statistical Analysis to test the null hypothesis that the means of these four groups are equal. 1,859,108.0 = 379.1 202 - 733 SS (14,038)2 1,859,487.1 4 106 = 1185952487.].- 1,926,232.o — 1,859,108.0 = 67,124.0 67,122.0 — 379.1 66.745.9 Source of Variation df SS MS ’F Significance at .05 Level Between Group Conditions (cc) 3 Within Group Conditions (w) 102 379.1 66,745.0 654.4 2.68 APPENDIX G ‘ STATISTICAL COMPUTATIONS FOR DETERMINING THE CQT-DCT CORRELATION COEFFICIENTS AND STANDARD SCORES BETWEEN THE SIX POSSIBLE COMPARISONS OF GROUP CONDITION CORRELATION COEFFICIENTS 51 52 Group Condition I DCT Standard Scores and Standard Deviation Standard Score Z Group Condition I DCT. _’ DCT Scores Deviation 2 X—X Subject (x) (x47) (x—X) ( 4.6 ) 2 30 4.3 18.49 .93 3 25 - .7 .49 - .15 5 17 —8.7 75.69 —1.89 6 30 4.3 18.49 .93 10 20 -5.7 32.49 —1.24 11 24 -1.7 2.89 — .37 F 12 28 2.3 5.29 .50 14 21 —4.7 22.09 -1.02 15 22 —3.7 13.69 - .80 16 25 - .7 .49 - .15 17 25 - .7 .49 - .15 18 31 5.3 28.09 1.15 20 29 3.3 10.89 .72 21 18 -7.7 58.52 -l.67 22 31 5.3 28.09 1.15 23 25 - .7 .49 - .15 24 29 3.3 10.89 .72 25 23 —2.7 7.29 - .59 26 31 5.3 28.09 1.15 27 32 6.3 39.69 1.37 28 33 7.3 53.29 1.59 29 20 -5.7 32.49 -1.24 30 23 ~2.7 7.29 - .59 2 Total 592 2{(x-x) =495 70 N = 23 7' 25 7 _.2 6‘ = g x—x) = 495.7 -— 2.1 60 = 4 5 N 23 z = x_x' = x4? . i 53 Group Condition I CQT Standard Scores and Standard Deviation Standard Score Group Condition I CQT Z CQT Scores Deviation 2 X4X Subject (x) (x-T) (x—x) 25.8 2 156 22.8 519.84 .88 3 97 —36.2 1310.44 —1.40 5 100 -33.2 1102.2 -1.29 6 151 17.8 316.84 .69 10 124 - 9.2 84.64 - .36 11 112 -21.2 449.44 - .82 12 153 19.8 392.04 .76 14 148 14.8 219.04 .57 15 148 14.8 219.04 .57 16 144 10.8 116.64 .42 17 122 -11.2 125.44 - .43 18 163 29.8 888.04 1.16 20 135 1.8 3.24 .07 21 140 6.8 46.24 .26 22 143 9.8 96.04 .38 23 135 1.8 3.24 .07 24 152 18.8 353.44 .73 25 93 —40.2 1616.04 -1.56 26 142 8.8 77.44 .34 27 160 26.8 718.24 1.04 28 175 41.8 1747.24 1.62 29 99 —34.2 1169.64 —1.33 30 72 —61.2 3745.44 -2.37 Total 3064 2f(xwx)2 = 15,319.92 N = 23 ‘x 133.2 ._ 2 ,. , = x-x = 15,319.92 = 666.08 = 25.8 o'- 544-. 7’:— ,2— z = x it = x if “5‘:— 25—13— 54 Group Condition 11 DCT Standard Scores and Standard Deviation Standard Group Condition II DCT Score Z DCT Scores Deviation __2 ( X—X ) Subject (x) (x-x) (x—x) 5.T" 1 24 — 6.1 37.21 -1.20 2 43 12.9 166.41 2.53 4 26 - 4.1 16.81 - .80 6 33 2.9 8.41 .57 7 23 - 7.1 49.70 -1.39 8 35 4.9 24.01 .96 9 28 — 2.1 4.41 — .41 11 31 .9 .81 .18 12 34 3.9 15.21 .76 13 26 4.1 16.81 - .80 14 24 - 6.1 37.21 —1.20 15 34 3.9 15.21 .76 16 3o — .1 .01 - .02 17 26 — 4.1 16.81 — .80 18 43 12.9 166.41 2.53 19 29 — 1.1 1.21 - .22 20 31 .9 .81 .18 21 30 - .1 .01 — .02 22 25 - 5.1 26.01 1.00 23 26 — 4.1 16 81 - .80 24 32 1.9 3.61 .37 25 35 4.9 24.01 .96 27 29 — 1.1 1.21 - .22 28 30 - .1 .01 — .02 29 25 — 5.1 26 01 -1.00 30 30 — .1 .01 - .02 Total 782 2{(x-x)2 = 675 15 N = 26 Y" 30.1 6 = { X-flg = 675.15 = 25.97 5. N 26“ z = x4? T571" 55 Group Condition II CQT Standard Scores and Standard Deviation Standard Group Condition II CQT Score Z CQT Scores Deviation '_ 2 ( X-X ) Subject (x) (x-x) (x—x) 23.9 1 102 —29 5 70.25 -1.23 2 168 36.5 1332.25 1.53 4 106 -25.5 650.25 —1.07 6 118 -13.5 182.25 — .56 7 99 —32.5 1056.25 -1.36 8 177 45.5 2070.25 1.90 9 140 8.5 72.25 .36 11 129 — 2.5 6.25 - .10 12 142 10.5 110.25 .44 13 110 -21.5 462.25 — .90 14 93 -38.5 1482.25 -1.61 15 169 37.5 1406.25 1.60 16 150 18.5 342.25 .77 17 109 -22.5 506.25 - .94 18 173 41.5 1722.25 1.74 19 130 - 1.5 2.25 — .06 20 132 .5 .25 .02 21 143 11.5 132.25 .48 22 129 — 2.5 6.25 - .10 23 121 -10.5 110.25 - .44 24 132 .5 .25 .02 25 172 40.5 1640.25 1.69 27 106 —25.5 650.25 -1.07 28 140 8.5 72.25 .36 29 105 -26 5 702.25 —1.11 30 124 — 7.5 56.25 - .31 Total 3419 got—702 = 14,844.50 N = 26 x 131.5 (J = 2(x-x‘)2 =flt,844.50 =757o.94 = 23.9 N 25 z = x—x' = x4x 6— 2309 Group Condition III DCT Standard Scores and Standard Deviation Standard Group Condition III DCT Scores Z DCT Scores Deviation __2 (_§7X' ) Subject (x) (x4x) (x—x) 2578“ 1 32 3.4 11.56 - .65 2 30 - 5.4 29.16 —1.04 4 32 3.4 11.56 - .65 5 39 3.6 12.96 .69 6 32 - 3.4 11.56 — .65 7 36 .6 .36 .42 10 32 — 3.4 11.56 - .65 11 43 7.6 57.76 1.46 12 34 1.4 1.96 - .27 13 30 5.4 29.16 —1.04 14 42 6.6 43.56 1.27 15 44 8.6 73.96 1.65 16 37 1.6 2.56 .31 17 34 - 1.4 1.96 — .27 18 25 —1o.4 108.16 —2.00 19 34 - 1.4 1.96 - .27 2o 38 2.6 6.76 .50 21 44 8.6 73.96 1.65 22 34 — 1.4 1.96 - .27 23 27 8.4 70 56 —1 62 24 39 3.6 12.96 .69 25 42 6.6 43 56 1.27 26 39 3.6 12.96 .69 27 31 - 4.4 19.36 - .85 28 33 — 2.4 5.76 — .46 29 31 — 4.4 19 36 - .85 30 43 7.6 56.76 1.46 Total 957 45(X‘7)2 = 734.72 N = 27 Y 35.4 II zi‘ NI R.) II D UL) [\D J: -\] o N] R) H [\D \1 1 H II \fl R) 57 Group Condition III CQT Standard Scores and Standard Deviation Standard Group Condition III CQT Score Z CQT Scores Deviation ._‘2 X¥X Subject (x) (xx) (x—X) (28‘0“. ) 1 109 ~21.0 441.00 -7.50 i 2 86 -44.0 1936.00 -1.57 ' 4 99 -31.0 961.00 -1.11 5 116 —14.0 196.00 - .50 6 104 —26.0 676.00 - .93 7 147 17.0 289.00 .61 10 90 -40.0 1600.00 -1.43 11 166 36.0 1296.00 1.29 12 112 -18.0 324.00 - .64 13 140 10.0 100.00 .36 14 151 21.0 441 00 .75 15 184 54.0 2916.00 1.93 16 113 —17.0 289.00 - .61 17 137 7.0 49.00 .25 18 133 3.0 _ 9.00 .11 19 109 -21.0 441.00 - .75 20 112 -18.0 324 00 - .64 21 154 24.0 576.00 .86 22 128 — 2.0 4.00 - .07 23 91 —39.0 1521.00 —1 39 24 149 19.0 361.00 .68 25 141 11.0 121.00 .39 26 166 36.0 1296.00 1.29 27 148 18.0 324 00 .64 28 92 —38.0 1444.00 —1.36 29 184 54.0 2916.00 1.93 30 148 18.0 324.00 .64 Total 3509 Z(X-X)2 = 21,175.00 Cw :9 z n /V\ 2 >4 I XI n m }_1 my: «14 Oi o O H Q (n 1: N O\ l m a) C) 58 Group Condition IV DCT Standard Scores and Standard Deviation Standard Group Condition IV DCT Scorg Z DCT Score Deviation __2 ( X-X_) Subject (x) (x-x) (x-x) ‘57l' 1 40 - .1 .01 - .02 2 41 .9 .81 .18 3 28 —12.1 146.41 —2.37 4 42 1.9 3.61 .27 5 39 1.1 1.21 - .22 6 33 - 7.1 50.41 -1.39 7 39 — 1.1 1.21 - .22 8 44 3.9 15.21 .76 9 37 - 3.1 9.61 — .61 10 37 — 3.1 9.61 - .61 11 45 4.9 24 01 .96 12 38 — 2.1 4.41 - .41 13 43 2.9 8.41 .57 14 33 - 7.1 50.41 —1.39 15 29 —11.1 123.21 —2.18 16 42 1.9 3.61 .37 17 42 1.9 3.61 .37 18 47 6.9 47.61 1.35 19 42 1.9 3.61 .37 20 34 — 6.1 37.21 —1 20 21 47 6.9 47.61 1.35 22 47 6.9 47 61 1.35 23 43 2.9 8.41 .57 24 39 — 1.1 1.21 - .22 25 47 6.9 47.61 1.35 26 42 1.9 3.61 .37 27 38 — 2.1 4.41 — . 1 28 46 5.9 34.81 1.16 29 44 3.9 15.21 .76 3o 35 — 5.1 26.01 -1 00 Total 1203 {(x—x)2 = 780.70 N = 30 x 40.1 G— : x-i'i 2 7/ 780.70 = 726.02 = 5 1 30 z = x—T' = 4? 'i 59 Group Condition IV CQT Standard Scores and Standard Deviation Standard Group Condition IV CQT Scorg Z CQT Scores Deviation __ 2 X—X Subject (x) (x—x) (x—x) (2211”? 1 167 32.1 1030.41 1.45 2 157 22.1 488.41 1.00 3 111 —23.9 571.21 -1.08 4 164 29.1 846.81 1.32 5 149 14.1 198.81 .64 6 110 —24.9 620.01 —1.13 7 95 —39.9 1592.01 —1.81 8 143 8.1 65.61 .37 9 134 - .9 .81 — .04 10 138 3.1 9.61 .14 11 133 - 1.9 3.61 - .09 12 111 —23.9 571.21 —1.08 13 132 — 2.9 8.41 — .13 14 110 -24.9 620.01 -1.13 15 159 24.1 580.81 1.09 16 122 —12 9 166 41 '- .58 17 149 14.1 198.81 .64 18 128 — 6.9 47.61 - .31 19 120 -14.9 222.01 - .67 20 140 5.1 26 01 .23 21 179 44.1 1944.81 2.00 22 136 1.1 1.21 .05 23 164 29.1 846 81 1.32 24 108 -26.9 723.61 —1.22 25 115 -19.9 396.01 - .90 26 136 1.1 1.21 .05 27 135 .1 .01 .01 28 93 -41.9 1755 61 —1.90 29 167 32.1 1030 41 1.45 30 141 6.1 37 21 .28 2 Total 4046 £(x—x) = 14,605 50 N = 30 x 134.9 9 o" : g(x—'x)" = 7/14,605.5o =1486.85 = 22.1 N 30 z = x4x 60 Group Condition I CQT—DCT Correlation Coefficient Computation and Significance Subject x (DCT deviation) y (CQT deviation) x.y 2 4.3 22.8 98.04 3 - .7 -36.2 25.34 5 -8.7 -33.2 288.84 6 4.3 17.8 76.54 10 -5.7 — 9.2 52.44 11 —1.7 —21.2 36.04 12 2.3 19.8 45.54 14 —4.7 14.8 —-69 56 15 -3.7 14.8 — 54.76 16 — .7 10.8 - 6.48 17 - .7 -11.2 7.84 18 5.3 29.8 157.94 20 3.3 1.8 5.94 21 -7.7 6.8 - 52.36 22 5.3 9.8 51.94 23 - .7 1.8 - 1.26 24 3.3 1 .8 62.04 25 —2.7 ~40.2 108.54 26 5.3 8.8 45.76 27 6.3 26.8 168.84 2 7.3 41.8 305.14 29 —5.7 -34.2 194.94 30 -2.7 _61.2 165.24 ix-y = 1712.52 r1 = ix-y = 1712 52 = 1712.52 2.;63 11 635 53- 23 (4.6) (25 8) 2729 64" Testing HO: rI = 0 df = n-2 = 23—2 2 21 Significance at °O5 level = .41 Therefore, reject HO and conclude rI of .63 is significant. 61 Group Condition II CQT-DCT Correlation Coefficient Computation and Significance Subject x (DCT deviation) y (CQT deviation) Xoy 1 — 6.1 -29.5 179.95 2 12.9 36.5 470.85 4 - 4.1 —25.5 104.55 6 2.9 -13.5 - 39.15 7 - 7.1 -32.5 230.75 8 4.9 45.5 222.95 9 — 2.1 8.5 - 17.85 11 .9 - 2.5 - .2.25 12 3.9 10.5 40.95 13 - 4.1 -21.5 88.15 14 6.1 -38.5 234.85 15 3.9 37.5 146.25 16 - .1 18.5 - 1.85 17 — 4.1 -22.5 92.25 18 12.9 41.5 535.35 19 — 1.1 - 1.5 1.65 20 .9 .5 .45 21 - .1 11.5 - 1.15 22 - 5.1 - 2.5 12.75 23 - 4.1 -10.5 43.05 24 1.9 .5 .95 25 4.9 40.5 198.45 27 — 1.1 -25.5 28.05 28 - .1 8.5 — .85 29 - 5.1 —26.5 135.15 30 - .1 — 7.5 .75 fix y = 2705 00 r11 = 23037 = 2705.00 2 ___2_705.00 = _._8___ N- 26 .1 2 . 16 .14 Testing HO: rII = df = n-2 = 26—2 = 24 Significance at .05 level = .39 Therefore, reject HO and conclude rII of .85 is Significant. 62 Group Condition III CQT-DCT Correlation Coefficient Computation and Significance Subject x (DCT deviation) y (CQT deviation) X°y 1 -3.4 —21.0 71.4 2 -5.4 —44 0 237.6 4 -3.4 -31.0 105.4 5 3.6 -14.0 - 50.4 6 -3.4 -26.0 88.4 7 .6 17.0 10.2 10 -3.4 -40.0 136.0 11 7.6 36.0 273.6 12 -1.4 -18.0 25.2 13 —5.4 10.0 — 54.0 14 6.6 21.0 138.6 15 8.6 54.0 464.4 16 1.6 -17 o — 27.2 17 -1.4 7.0 - 9.8 18 —1o.4 3.0 - 31.2 19 -1.4 -21.0 29.4 20 2.6 —18.0 - 46.8 21 8.6 24.0 206.4 22 —1.4 — 2.0 2.8 23 -8.4 —39.0 327.6 24 3.6 19.0 68.4 25 6.6 11.0 72.6 26 3.6 36.0 129.6 27 -4.4 18.0 - 79.2 28 -2.4 -38.0 91.2 29 -4.4 54.0 -237.6 30 7.6 18.0 136.8 fix = 2079.4 rIII == £><- y’ = 2 2307394 25 — 2,079.4 =___3 . . _ . 3, 31.2 N 6'32— C—y- 7 (.5 ) ( O) 9 Testing HO: rIII = 0 df = n-2 = 27-2 = 25 Significance at .05 level .38 Therefore, reject HO and conclude rIII of .53 is significant. 63 Group Condition IV CQT—DCT Correlation Coefficient Computation and Significance Subject x (DCT deviation) y (CQT deviation) x-y 1 - .1 32.1 - 3.21 2 .9 22.1 19.89 3 12.1 -23.9 289.19 4 1.9 29.1 55.29 5 - 1.1 14.1 - 15.51 6 7.1 —24 9 176.79 7 - 1.1 -39 9 43.89 8 3.9 8.1 31.59 9 - 3.1 — .9 2-79 10 - 3.1 3.1 9.61 11 4.9 - 1.9 9.31 12 — 2.1 —23.9 50.19 13 2.9 — 2.9 - 8.41 14 — 7.1 -24.9 176 79 15 —11.1 24.1 266.40 16 1.9 —12.9 — 24.51 17 1.9 14.1 26.79 18 6.9 - 6.9 — 47.61 19 1.9 —14.9 — 28.31 20 — 6.1 5.1 - 31.11 21 6.9 44.1 299.88 22 6.9 1.1 7.59 23 2.9 29-1 84.39 24 - 1.1 -26 9 29.59 25 6.9 —19 9 137.31 26 1.9 1.1 2.09 27 — 2.1 01 — .21 28 5.9 —41.9 247.21 29 3.9 32.1 125.19 30 - 5.1 6.1 31.11 jixny = 562.10 T‘lv = z . y = 562 . 10 = 562 . 10 = _._1__7 No C§'-C7§ 30(5.1) (22 1) 3381530‘ Testing Ho: rIV = 0 df = n-2 = 30—2 = 28 Significance at .05 level = .36 Therefore, Ho is not rejected and it is concluded that rIV of .17 is not significant. 64 Testing the Statistical Significance between the Six Possible Pairs of Group Correlation Coefficients using Standard Scores (Z) Based Upon zl Transformation Scores. ll |._J 1. Ho : rI — rII = O z1 — leI = 1 + 1 = 1 I 73:3— 2673' ‘20— “‘73 = 7(650 + .043 = .093 = .3 1 _ 1 Z I of 63 — 741, Z II of 85 1.256 1 1 z = Z I - Z II = .741 — 1 256 = - .515 C7721 21 °3 '3 I 11 = —1.71 This is Significant at the .05 level using a one-tailed test so we reject the Ho and conclude that rI is Signifi- cantly greater than rII Ho‘ rI ‘ r111 = O 1 1 . z — z = 1 + 1 = 1 + 1 (77771 III 23-3 27—3 l/Vf2o '"24‘ = ‘7.050 + .042 = .092 = .3 1 1 .63 = .741; z III of .53 = 590 z = 741 — .590 = .151 = ._59 3 .3 This is not significant at the 005 level and, therefore, we tentatively accept Ho that rI is not significantly greater than rIII . r" A 55 o‘ r11 ‘ r11:1 = O ‘ 1 1 ‘77 flu - Z111 =m+ .042 = 7.0850 = .292 l 1 = .2 ; = . = z 11 1 56 2 III 590 z = .666 = 2.28 .292 "“‘ This is Significant at the .05 level using a one-tailed test so we reject the HO and conclude that rII is Sig- nificantly greater than r III H : - = o rI rIV O I - IV — + l — .050 + .037 — .0-7 20 27 = °295 z1 of 63 — 741 z1 of 17 — 172 I I ’ IV ° _ ° 2 = .741 — .172 = .569 = 1.93 .295 "73T5‘ This is significant at the .05 level using a one-tailed test so we reject the HO and conclude that cantly greater than rIV Ho r11 ‘ r1v = O 1 1 _ _ (:72 11 — 2 1V .17{o43 + .037 _ .080 z = 1 256 — .172 = 1.084 = .282 “"7282“ This is significant at the .05 level using test so we reject the HO and conclude that cantly greater than rIV rI is signifi- .282 3.84 a one—tailed rII is Signifi- 66 HO‘ I'111 ‘ r1v = 0 g__ 1 1 _ _ _ (7/2 111 — 2 IV _ )77042 + .037 _ .079 _ .281 z = .590 - .172 = .418 = 1.49 .281 “7281" This is not Significant at the .05 level so we tenta— tively accept Hg and conclude that r is not signifi- .h cantly greater an rIV’ although itITS close to being significant. MICHIGAN STATE UNIVERSITY LIB O 3 1293 ARIES I IIIIIII 3 3061816