MSU LIBRARIES .— RETURNING MATERIALS: P1ace in book drop to remove this checkout from your record. FINES wi11 be charged if book is returned after the date stamped be10w. ANXIETY IN FIGURE DRAWINGS, THEIR STORIES AND WISHES By Kathleen Jane Hamernik A THESIS Submitted to Michigan State University in partial fulfillment of the requirements fbr the degree of MASTER OF ARTS Department of Psychology 1985 To Don and Charlotte, with love ii ACKNOWLEDGEMENTS This work benefited from, indeed often depended upon, the contributions, efforts, and encouragement of many significant people. The contributions of Dr. Joseph Reyher, my chairperson, were manifold: I am grateful for his breadth of vision, his depth of analysis, his contagious enthusiasm, and his commitment to quality research. I also thank my other committee members: Dr. John Hurley, for his many useful suggestions which enriched the project and improved the text; and Dr. Terry Allen for his unfailing good- humor and patient guidance in statistical matters. Special thanks go to Bernadette Jovanelly and Ben Blake for their diligence and professionalism during the long summer months of data collection. I remember with great appreciation the thirty undergraduate psychology majors who maintained their high spirits while steadily turning out the many measurements and calculations this study required. I am also grateful to Abby Golomb and Jeff Roach for so generously sharing their invaluable computer wisdom. Finally, I thank Sallie, Pete, Dan, and Lori for their unwavering faith in me. iii TABLE OF CONTENTS Page LIST OF TABLES ....................... vi LIST OF FIGURES ...................... viii INTRODUCTION ........................ l Hypotheses ...................... 7 Hypothesis I ................... 7 Hypothesis II ................... 8 Hypothesis III .................. 8 METHOD ........................... 9 Subjects ....................... 9 Experimenters ..................... 9 Apparatus ....................... 9 Instruments ...................... lO Handler Draw-A-Person Rating Scales ........ l0 Roach Draw-An-Automobile Rating Scales ...... l0 Experimental Design .................. ll Procedure ....................... 12 RESULTS AND DISCUSSION ................... 18 Hypothesis I ..................... 23 Hypothesis II ..................... 3l Hypothesis III .................... 34 Other Anxiety Results ................. 35 Additional Findings for Drawings ........... 39 Findings for Time ................... 4l Interpersonal Findings ................ 42 GENERAL DISCUSSION ..................... 45 Sex of Subject Findings ................ 45 Stimulus Object Differences .............. 47 Response Modality Findings .............. 5l Sex of Experimenter Findings ............. 52 iv Page APPENDIX I ......................... 55 APPENDIX 11 ........................ 63 APPENDIX III ........................ 64 APPENDIX IV ........................ 66 REFERENCES ......................... 83 LIST OF TABLES Table Page l Possible Experimental Conditions ........... l3 2 Possible Orders of Stimulus Objects Allowing Subject to Choose Sex of First Person Drawing . . . . l4 3 Specification of the Analysis of Variance Design for Drawings Including Degrees of Freedom and Error Terms ........................ l9 4 Pearson Correlations Between Raters on Seven Graphic Indices for Each Drawing Stimulus Object. . . 21 5 Specification of the Analysis of Variance Design for Stories and Wishes Including Degrees of Freedom and Error Terms ................... 22 6 Analysis of Variance for Average Conductance Scores for Drawings ..................... 66 7 Analysis of Variance for SR Frequency Scores for Drawings ....................... 67 8 Analysis of Variance for Shading Scores for Drawings ....................... 68 9 Analysis of Variance for Erasure Scores for Drawings ....................... 69 10 Analysis of Variance for Reinforcement Scores for Drawings ....................... 70 ll Analysis of Variance for Omission Scores for Drawings ....................... 7T 12 Analysis of Variance for Line Discontinuity Scores for Drawings .................... _. 72 l3 Analysis of Variance for Vertical Imbalance Scores for Drawings ................. 73 vi Table Page l4 Analysis of Variance for Transparency Scores for Drawings .................... 74 TB Analysis of Variance for Area (Square Inches) of Drawings .................... 75 T6 Analysis of Variance for Time for Drawings ..... 76 T7 Analysis of Variance for Average Conductance Scores for Stories ................. 77 l8 Analysis of Variance for SR Frequency Scores for Stories .................... 78 19 Analysis of Variance for Time Scores for Stories. . 79 20 Analysis of Variance for Average Conductance Scores for Wishes ................. 80 2l Analysis of Variance for SR Frequency Scores for Wishes ..................... 8T 22 Analysis of Variance for Time Scores and Wishes . . 82 vii Figure l Sex of subject by stimulus object interaction for average conductance scores on drawings ........ 2 Sex of subject by stimulus object interaction for SR frequency scores on drawings ......... 3 Stimulus object main effects for shading, erasure, and line discontinuity scores on drawings ...... 4 Stimulus object main effects for omissions and vertical imbalance scores on drawings ........ 5 Stimulus object main effect for average conductance scores on stories ............ 6 Response modality by stimulus object interaction for SR frequency scores on drawings ......... 7 Sex of subject by response modality interaction for average conductance scores on stories and wishes ........................ 8 Sex of subject by experimenter interaction for shading and erasure scores on drawings ........ 9 Sex of subject by stimulus object interaction for area scores on drawings ............. lO Experimenter by response modality interaction for LIST OF FIGURES subject ratings of experimenter respectfulness and consideration .................. viii Page 24 25 28 29 32 33 36 38 4O 44 ABSTRACT ANXIETY IN FIGURE DRAWINGS, THEIR STORIES AND WISHES By Kathleen Jane Hamernik The anxiety associated with completion of three projective tasks (figure-drawing, story—telling, wish-making) having four stimulus objects (male, female, animal, automobile) was studied in order to clarify its possible sources (intrapsychic conflict, interpersonal dynamics, task demands). Ninety-six subjects were assigned conditions that varied by sex, experimenter (male, female), and verbal response modality (oral, written). Anxiety was defined by mean skin conductance level (SC), spontaneous skin response frequency (SRR), and seven graphic indices. Additional measures were time per task and interpersonal ratings. Significant sex differences were produced by both electrodermal measures; males were consistently more anxious than females. Oral condition subjects demonstrated significantly more anxiety than written condition subjects. Significant stimulus object differences suggest that SC and SSR reflect dissimilar phenomena. The female experimenter received significantly higher interpersonal ratings, yet her subjects spent significantly less time on all tasks; a personal power hypothesis is offered as a possible explanation. INTRODUCTION Projective tests have long held a respected place in most clinicians'diagnostic armamentarium. Over the years, different projective tasks have been devised--figure drawing, story-telling, wish-making and so on--all with the same basic goal: to elicit important information about the client or subject's psychological functioning. While clinical wisdom plays a large role in the interpretation of projective material, it is research with projectives that affords the scientific foundation on which clinical interpretation rests. Research into the relationship of anxiety to projective tasks is an excellent example of how lab study enhances clinical understanding. Work by Reyher and his associates has indicated that the anxiety manifested by a subject completing a projective task may have several sources: intrapsychic, interpersonal, and task demand. Efforts to clarify these different sources of anxiety are valuable in that they help correct a tendency many clinicians have to overlook the effect of the interpersonal situation and the task demands on projective test results. Reyher initiated this line of research in T959 with a focus on figure drawings. He proposed that drawings of human figures be compared to a relatively neutral and common figure of equal difficulty. The anxiety manifested in the human drawing could then be separated and attributed to two sources: (a) the testing situation which would 2 affect the human and neutral figure equally, and (b) intrapsychic conflict which would affect only the human figures. Reyher selected an automobile as the neutral figure. Thus, if an automobile drawing has fewer graphic indices of anxiety than the human figure drawings, the clinician has an objective basis for formulating psychodynamic hypotheses to account for these differences. Handler and Reyher (1964) investigated the possibility that the level of difficulty of a drawing might be an influencing variable. They found, using the method of paired comparisons, that the automobile was judged to be as difficult to draw as the human figure. They also found evidence for the two sources of anxiety. Male, female, and automobile drawings were obtained from the same male subjects under nonstress and high stress conditions. Seventeen indices of anxiety for the drawings from the two conditions were compared. Individual drawing style and ability variables were eliminated by using the subject as his own control. As hypothesized, there was a differential increase in the number of indices of anxiety far the stress condition which significantly distinguished the stress and nonstress drawings. For the automobile drawings, 5 indices significantly differentiated the stress from nonstress condition; for the human drawings, l5 indices for the male and ll for the female significantly differentiated conditions. Handler and Reyher concluded that the increase in graphic indicators fOr the automobile reflected anxiety stemming from the testing situation itself, whereas the increase for the human figure drawings, beyond the increase for the automobile, reflected intrapsychic conflict which was exacerbated by the stress condition. 3 This research was expanded when Handler and Reyher (1966) demonstrated that human figure drawings were characterized by more spontaneous skin response activity (EDA) as well as more graphic indices of anxiety than the automobile. Male subjects each produced a male, female, and automobile drawing while a continuous EDA record was obtained. Analysis of the EDA frequency and mean conductance revealed significant differences in both measures between all three drawings. The female drawing yielded the highest level of anxiety, and the automobile drawing yielded the lowest level. Further, 13 out of 18 graphic indices of anxiety significantly differentiated the three drawings. Sanders and Reyher (1972) produced research results which supported some of the previous findings and conflicted with others. Male subjects were asked to draw clothed human figures, nude human figures, and an automobile. Using EDA frequency as their dependent measure, Sanders and Reyher demonstrated again that drawing a clothed human form is significantly more anxiety provoking than drawing an automobile. Moreover, drawing a human figure in the nude was shown to be significantly more anxiety provoking than drawing a human figure clothed. However, in contrast to the Handler and Reyher (1966) results, the anxiety elicited by the clothed female drawing task was not greater than the anxiety elicited by the clothed male drawing task; nor was there a difference in anxiety production between the nude male and female figures. Roach (l981) replicated Handler and Reyher's (1964) study demonstrating two sources of the anxiety manifested by a subject during a figure drawing task: (a) the testing situation itself, 4 and (b) intrapsychic conflict. He elaborated upon the nature of the anxiety experienced by the subjects by suggesting that it was of an interpersonal type when the experimenter was present during the task, and of a performance type when the experimenter was absent. Roach also expanded previous studies by using both male and female subjects and experimenters. His analysis revealed that when the experimenter was present in the room and of the opposite sex, the subject's mean conductance level was higher across all three drawings than if the experimenter was present and of the same sex; that is, an opposite sex experimenter provoked a significantly greater degree of overall anxiety. Myers (1965) added a new facet to the research by investigating the clinical usefulness of stories told about male, female, and animal drawings. Using female subjects, Myers obtained an EDA frequency measure for each of the three stories; however, no differences were fOund. These results may be due to the fact that the subjects were required to write their stories. Written communication is, in general, a less immediate or spontaneous form of communication. A writer must pay more careful attention than a speaker to the logic and order of his/her statements. A writer must also be able to achieve some distance from his/her thoughts in order to effectively edit them. Thus, the writer can be said to be making greater use of what Freud called "secondary process" (1900/1955, 1915/1933). Secondary process has two purposes: (a) to increase the logic with which a thought is presented, and (b) to attenuate the affect associated with that thought. In so far as a writer is making greater use of secondary process, then, his/her productions can be 5 expected to be more orderly and less affect-arousing. As a consequence, less EDA activity could be expected from Myers's subjects as they wrote out their stories. This present study is an attempt to integrate and extend various dimensions of these earlier studies. The primary focus here is the degree of anxiety associated with three projective tasks (figure- drawing, story-telling, and wish-making) and four stimulus objects (male, female, animal, and automobile). The basic assumption is that the activation of intrapsychic conflict causes anxiety. Previous research has already demonstrated with figure drawings that a human stimulus object activates more intrapsychic conflict and thus generates more anxiety than an inanimate (automobile) stimulus object. This study introduces, first, the effect of a nonhuman but animate (animal) stimulus object on the degree of anxiety manifested during a figure drawing task. The animal figure is expected to elicit more anxiety than the automobile, but less than the human figure. This expectation is in accordance with a recent formulation by J. Reyher (personal communication, December 12, 1984) on the mechanism underlying the clinical usefulness of figure drawings. As Reyher explains, a projective drawing is essentially a commitment to paper of an image created by the subject. Research (Moses and Reyher, 1985) has shown that when a particular figure, scene, or experience is imaged, the affect associated with that figure, scene, or experience is also evoked. Thus, for example, if the female body is associated with conflict for a subject, then the process of imaging and drawing a female figure will produce anxiety. Even more specifically, the imaging and drawing of those parts of the female body which are 6 associated with conflict will produce anxiety. The results obtained by Handler and Reyher (1966) support this fOrmulation. For the male figure, male subjects demonstrated the highest level of anxiety while drawing secondary sexual body parts; yet fOr the female figure, they demonstrated the highest level of anxiety while drawing the primary sexual body parts. Handler and Reyher pointed out that this differential anxiety response corresponds to the different foci of concern/conflict adolescent males have for the bodies of the two genders. For their own male bodies, they are most concerned about secondary sexual characteristics; for the female body, they are most concerned over the primary sexual characteristics. Applying this formulation to the present study, then, the imaging and drawing of an animal figure is expected to be associated with less affect, particularly conflict, than either of the two human figures. Thus, the animal figure is expected to elicit less anxiety than either human figure. On the other hand, the animal figure can be expected to be associated with greater affect and conflict than the automobile by nature of its being a living creature; so the animal drawing is expected to produce greater anxiety than the automobile drawing. In addition, this study tests the generalizability of previous research findings with figure drawings to other projective tasks, i.e., story-telling and wish-making. Here, the stories and wishes obtained from subjects are considered imaged productions like the drawings. In this case, however, the subject expresses his/her creative production verbally rather than graphically. This similar origin of production--the subject's personal imagination--is the basis for 7 predicting that anxiety will be manifested in similiar ways for all three projective tasks - drawings, stories, and wishes. Finally, this study tests the effect of the mode of response (oral vs. written) on the demonstration of anxiety during verbal projective tasks. Subjects who communicate their stories and wishes in writing are seen as making greater use of secondary process; as a consequence, any affect/conflict associated with the four different story stimulus objects (male, female, animal, automobile) is diminished. So, writing subjects are expected to display smaller differences in anxiety across stimulus objects than subjects who produce their stories and wishes orally. There are two operational definitions of anxiety used in this study; one is physiological, the other is graphic. The physiological index of anxiety is sympathetic nervous system activity, measured specifically by electrodermal (sweat gland) activity; the sweat glands are controlled exclusively by the sympathetic system (Hassett, 1978). Two types of EDA are considered: mean skin conductance level, and spontaneous skin response frequency. The graphic index of anxiety is defined by seven drawing variables: shading, erasure, reinforcement, omissions, line discontinuity, vertical imbalance, and transparencies. Hypotheses Hypothesis 1. During a figure drawing task, a nonhuman but animate stimulus object elicits less anxiety than a human stimulus object, and more anxiety than an inanimate stimulus object. A progression of increasing anxiety is therefore expected for the automobile, animal, and human figure drawings. 8 Hypothesis II. During a story-telling or wish-making projective task, an inanimate stimulus object, a nonhuman but animate stimulus object, and a human stimulus object will elicit a progression of increasing anxiety. It is predicted, therefore, that a subject completing a story- telling or wish-making task will manifest a pattern of anxiety similar to that already demonstrated for the figure drawing task; that is, the level of anxiety will be greater for human stimulus objects than for an inanimate (automobile) stimulus object. It is further predicted that during a story-telling or wish-making task, the nonhuman but animate stimulus object will elicit less anxiety than the human stimulus object, but more anxiety than the inanimate stimulus object. Thus, for all three projective tasks (figure drawing, story-telling, and wish-making), a pattern of increasing anxiety if expected for the automobile, animal and human stimulus objects. Hypothesis III. The demonstration of stimulus object related differences in the degree of anxiety manifested during a verbal projective task is a function of the mode of response, oral or written. It is predicted, therefore, that subjects who complete the story-telling and wish-making tasks orally will demonstrate greater stimulus object related differences in anxiety than subjects who complete the tasks in writing. METHOD Subjects Ninety-six undergraduate students (48 males and 48 females) from introductory psychology courses participated in this experiment. These subjects were volunteers who signed up on sheets posted in their classrooms for an experiment entitled, "Figure Drawings, Stories, and Wishes." Their reward for participating was two credit points which could be applied toward their final course grade. Experimenters The experimenters were two (one female and one male) undergraduate advanced psychology majors. They were selected on the basis of their interest in psychological research and their intention of obtaining advanced degrees in psychology. Both experimenters received extensive training to ensure that they could skillfully manipulate the equipment required by the experiment and that they could interact with subjects in a professional, ethical manner. Both experimenters were attractive and personable. Apparatus A Grass (model 5) six channel polygraph and Beckman electrodes (Ag/AgCl; 177 cm2) were used to record electrodermal responses. The electrodes were filled with Beckman electrolyte and attached to the medial phalanx of the second and third fingers of the nonpreferred hand. 10 Skin resistance was continuously recorded on paper tape by the polygraph. Mean skin conductance was calculated by averaging the reciprocals of resistance levels measured at sixteen second intervals. The first reading was taken eight seconds after the subject began each task. Transforming the skin resistance level to its reciprocal, skin conductance, normalized the heavily skewed resistance distribution. The frequency of skin responses was measured by counting the number of pen deflections that were greater than 1000 ohms for each task period. The sum was then divided by the amount of time taken to complete each task; the resulting number was spontaneous skin responses per second. The experimenters noted the start and finish of each task directly on the polygraph paper tape; the time for each task was calculated by measuring the distance between these start and finish notations. Instruments Handler Draw-A-Person Rating Scales. Handler (1967) delineated twenty graphic indices of anxiety for rating human figure drawings. Seven of these indices were used in this study: shading, erasures, reinforcement, omissions, line discontinuity, vertical imbalance, and transparencies. These particular indices were chosen on the basis of the ease with which they could be applied to animal as well as human figures. Roach Draw-An-Automobile Rating Scales. Roach (1981) devised scales for rating automobile drawings on twelve graphic indices of anxiety. These scales were constructed so as to approximate Handler's 11 (1967) scoring procedures for human figures. For the present study, seven graphic indices were chosen from Roach's manual, the same seven indices that were chosen from Handler's manual and that are listed above. Some modifications were made in the scoring procedures of both the Handler and Roach scales, and analagous scales were defined for the animal figures (see Appendix I). The intent was to obtain three sets of procedures (one for the human figures, one for the animal figures, and one for the automobiles) that were as similar as possible for each of the seven graphic indices of anxiety. In addition, the issue of figure size was addressed by defining a procedure for calculating the area of the page covered by a figure (see Appendix II). For each of the graphic indices, one pair out of ten potential raters was trained using the modified Handler and Roach manuals. Before rating the experimental drawings, each pair of raters had to achieve an 80% agreement level on a set of 20 practice drawings. Once the pair met this criterion, they rated the entire set of experimental drawings on that index. The dependent variables used in the analysis were computed by averaging the two ratings on each drawing for each variable; decimals were avoided by rounding down to the next whole number. Experimental Design Male and female subjects were randomly assigned first to either the male or female experimenter. They were then randomly assigned to one of the two response modes, either the written response mode or the oral response mode. These three variables, (sex of subject, sex of experimenter, and response mode) resulted in eight possible 12 experimental conditions (see Table 1). Provisions were made for an equal g_in each condition. All subjects were asked to make a drawing for each of four stimulus objects (male, female, animal, and automobile) and to formulate a story and three wishes for each drawing. Thus, each subject produced twelve pieces of data. The written response mode subjects wrote their stories and wishes, the oral response mode subjects told their stories and wishes directly to the experimenter. The request to draw a human figure was phrased so as to allow the subject to choose which sex he/she would draw first. The procedure yielded twelve possible orders of stimulus objects (see Table 2). Twelve subjects were therefore included in each of the eight conditions described above. One of the twelve stimulus object orders was randomly assigned to each subject. Procedure The subject was greeted and ushered into an eight foot square room containing a large desk with chair, the polygraph machine, a swivel chair, and a small stand with the electrode paraphanalia and a supply of paper and pencils atOp it. The subject was seated at the large desk, with his/her back to the polygraph. A cassette tape recorder rested on one corner of the desk. The experimenter was seated in the swivel chair to the side of the subject so as to allow the experimenter to make eye contact with the subject when giving instructions, and then to turn and monitor the polygraph while the subject completed the various tasks. Once the subject was seated, the experimenter explained, "for this experiment I will be attaching two electrodes to your fingers. Possible Experimental Conditions Male subject Male experimenter Oral response mode Male subject Male experimenter Written response mode Male subject Female experimenter Oral response mode Male subject Female experimenter Written response mode Female subject Female experimenter Oral response mode Female subject Female experimenter Written response mode Female subject Male experimenter Oral response mode Female subject Male experimenter Written response mode 14 Table 2 Possible Orders of Stimulus Objects Allowing Subject to Choose Sex of First Person Drawing Person Opposite sex person Animal Auto Person Auto Opposite sex person Animal Animal Person Opposite sex person Auto Auto Person Opposite sex person Animal Person Opposite sex person Auto Animal Person Animal Auto Opposite sex person Animal Person Auto Opposite sex person Auto Person Animal Opposite sex person Person Animal Opposite sex person Auto Person Auto Animal Opposite sex person Animal Auto Person Opposite sex person Auto Animal Person Opposite sex person 15 They will not hurt you at all. They will simply measure certain aspects of your physiology which interest us. I will also be asking you to draw some pictures and make-up some stories." The experimenter then answered questions and obtained the subject's signature on an informed consent form. The electrodes were then attached, the polygraph calibrated, and the tape recorder turned on. When the subject's resistance level had stabilized, the experimenter handed him/her a pencil and clean sheet of paper, and said, "On this first sheet of paper, I am going to ask you to make several drawings. I will then be giving you additional sheets of paper, and will ask you to make just one drawing on each of them. Each time I will tell you what to draw and when to begin; I will also give you a fresh pencil." The novelty of the situation was expected to produce enhanced EDA activity, so time was allowed for the subject's skin resistance to stabilize. The subject was then instructed to draw a circle, a triangle, pentagon, square, octagon, and a rectangle. After each of these geometric drawings, time was again allowed for the subject's skin resistance to stabilize. This series of preliminary drawings served the purpose of allowing habituation to occur. The subject was then given a fresh sheet of paper and a fresh pencil, and told, "Now draw a human person (an animal, an automobile), and tell me when you are finished." When the drawing was completed, the written response mode subjects were handed a clean sheet of paper. The experimenter then said, "Now I am going to ask you to make up a story about the person (animal, automobile) you have just drawn. Write your story on this sheet of paper. Be sure to 16 include what is happening at the present, how the person (animal, automobile) feels, what led up to this, and what the outcome will be." When the subject had completed the story, he/she was further instructed, "Now write down three wishes that this person (animal, automobile) might make." The oral response mode subject's instructions were somewhat different. After the drawing was completed, the experimenter said, "Now I am going to ask you to make up a story about the person (animal, automobile) you have just drawn. Tell me what is happening at the present, how the person (animal, automobile) feels, what led up to this, and what the outcome will be." Then, upon completion of the story, the subjects were asked, "What three wishes might this person (animal, automobile) make?" This procedure was repeated three more times according to the stimulus object order which had been randomly assigned to the particular subject. Any questions a subject asked about the task were answered in a nondirective manner. When all the drawings, stories and wishes were completed, the experimenter turned off the machines and removed the electrodes from the subject's fingers. The subject was then handed a four item interpersonal evaluation form. The experimenter explained, ”This scale asks how you feel about me. Please answer it honestly. I will not see it. When you have finished, put it face down in the box in the corner." The experimenter then turned away from the subject and completed a separate, five item interpersonal evaluation form on his/her feelings about the subject (see Appendix III). When both had finished the 17 forms, the experimenter answered any further questions the subject had, and then thanked the subject for his/her participation. RESULTS AND DISCUSSION Data from three of the ninety-six subjects were excluded from the analysis. One subject, an outlier, had extreme time scores (more than six standard deviations above the mean for four of the twelve tasks). Two other subjects produced polygraph tapes which were meaningless (machine malfunction or experimenter error are suspected). Fortunately, each discarded subject belonged to a different cell, minimizing the differences in sample sizes. A meaningful analysis of the data was therefore still possible, and the method of unweighted means was used. For the drawing tasks, a three factor design with repeated measures was used to evaluate the influence of the independent variables and their interactions on dependent measures (see Table 3). Sex of subject (P) and experimenter (E) were the between-subjects factors, and drawing stimulus object (O) was the within-subjects factor. These factors were completely crossed; the within-groups factor, subjects (S), was nested within the combination of sex of subject and experimenter variables (a specific subject cannot be both male and female; nor can it have both the male and female experimenter). Interrater reliability was estimated for the seven graphic indices of anxiety (see Table 4). It was obtained by computing from the entire data set the Pearson product-moment correlations between 18 19 Table 3 Specification of the Analysis of Variance Design for Drawings Including Degrees of Freedom and Error Terms Source Symbol DF Error Term Between Subjects Sex of subject P l S/PE Experimenter E l S/PE Sex of S by E PE 1 S/PE Subjects S/PE 89 None Within Subjects Drawing stimulus object 0 3 OS/PE Sex of S by stimulus object PO 3 OS/PE Experimenter by stimulus EP 3 OS/PE object Experimenter by sex of S PEO 3 OS/PE by stimulus object Stimulus object by subject OS/PE 267 None 20 Table 4 Pearson Correlations Between Raters on Seven Graphic Indices for Each Drawing Stimulus Object Drawing Stimulus Objects Index Male Female Animal Automobile Shading .98 .94 .98 .97 Erasures .75 .85 .91 .95 Reinforcement .88 .80 .90 .90 Omissions .95 .96 .95 .93 Line Discontinuity .94 .93 .92 .90 Vertical Imbalance .81 .87 .75 .82 Transparencies .93 .60 .57 .69 21 raters on each index for each drawing stimulus object. All but three of the correlations are above .75 and therefore considered acceptable for research purposes. Interestingly, the three exceptions are all related to the transparency index. For the female, animal, and automobile stimulus object, the correlations are .60, .57, and .69, respectively. Low interrater reliability on the transparency index has also been reported in other studies (Roach, 1981; Roach, 1984). One explanation may be that a large number of drawings receive a zero rating on the transparency index; these zero scores then severely skew the distribution. To provide another perspective, interrater agreement ratios (# of agreements/# of possible agreements) were calculated for the transparency index. The results were .97, .85, .85, and .91 for the male, female, animal, and auto stimulus objects, respectively. For the story-telling and wish-making tasks, a four-factor design with repeated measures was used (see Table 5). The three between-subjects factors were sex of subject, experimenter, and response modality (R); story/wish stimulus object (O) was the within—subjects factor. Again, these factors were completely crossed; subjects (5) were nested within the combination of sex of subject, experimenter, and response modality (a specific subject cannot be of both sexes, have an experimenter of both sexes, nor produce stories and wishes in both oral and written form). A summary table for the analysis of variance for each of the seventeen dependent variables can be found in Appendix IV. 22 Table 5 Specification of the Analysis of Variance Design for Stories and Wishes Including Degrees of Freedom and Error Terms Source Symbol DF Error Term Between Subjects Sex of subjects 1 S/PER Experimenter l S/PER Response modality l S/PER Sex of S by E PE 1 S/PER Sex of S by response PR 1 S/PER modality Experimenter by response ER 1 s/pER modality Sex of S by E by PER l S/PER response modality Subjects S/PER 85 None Within Subjects Drawing stimulus object 0 3 OS/PER Sex of S by stimulus object PO 3 OS/PER Experimenter by stimulus E0 3 OS/PER object Response modality by R0 3 OS/PER stimulus object Sex of S by E by stimulus PEO 3 OS/PER object Sex of S by response PRO 3 OS/PER modality by stimulus object Experimenter by response ERO 3 OS/PER modality by stimulus object Sex of S by E by response PERO 3 OS/PER modality by stimulus object Stimulus object by subject OS/PER 255 None 23 Hypothesis I The results provided minimal support for the prediction of a progression of increasing anxiety for the auto, animal, and human figure drawings. Instead, the results revealed interesting sex of subject main effects and interactions. The first physiological variable, average conductance, produced a significant sex of subject main effect, (F(1, 89) = 8.22, p_< .01; male subjects had a significantly higher overall level of average conductance than female subjects. This finding is qualified, however, by a significant sex of subject by stimulus object interaction, F(3, 267) = 3.27, p-< .03 (see Figure l). Probing the interaction with a test of simple effects revealed no appreciable differences in average conductance levels between stimulus objects for female subjects. By contrast, male subjects did produce a significant simple effect for stimulus object, F(3, 267) = 4.69, p,< .01. Applying a Tukey test to these four stimulus object means showed that the female drawing elicited a significantly higher level of average conductance than the automobile drawing. The second physiological variable, spontaneous response (SR) frequency, also produced a significant sex of subject main effect, F(1, 89) = lO.71,_p < .01; male subjects had an overall higher SR frequency than female subjects. As with average conductance, however, this finding is qualified by a significant sex of subject by stimulus object interaction, F(3, 267) = 4.84, p_< .01 (see Figure 2). A test of simple effects demonstrated no differences in SR frequency between stimulus objects for female subjects. A significant simple effect for stimulus object was discovered for male subjects, F(3, 267) = 5.24, p.< .01. Using Tukey's method to test these four means, the A /_ 24 .mucmzmec :o mmLoUm mucmuuaucou ommem>m so» cowaumgmucm uumwao mzpasmum an uummnam we xmm .P mgamwm how—do m:..=s:...m 3:< .mEE< 22:0“. 0.2: _ _ _ _ nr Idem I) C 10 I m4<2mu 01 Ci 11 cm W a a v 1 V. m l «o a I. 8 L ”.3 352 1 o o O: I I on (rum [0/ \O 1 a I. E. n - m \ v / J on o. J m. 3 «m J 25 .mmcpzmgv co mucoum xucmzcocm mm Lee :owuumemucm pumnno mapzswum z: pumnaam mo xmm .N mean?“ homfimc mains—Eh 3.3 .0873 2an". . 222 a _ _ _ Nu . 1 no H.529. Or 1 3. M v 13. N S co. 3 .352 Q / fl / .150 m / / M / Jmo m / O J mo 26 male drawing was shown to elicit a significantly higher SR frequency than the female, animal, and automobile drawings. For the drawing task, then, the two physiological measures produced stimulus object related differences in manifest anxiety only among male subjects. According to both measures, it was a human figure that elicited greater anxiety. This supports Reyher's (1959) original hypothesis that human figures elicit testing-situation anxiety as well as intrapsychic anxiety while neutral figures elicit testing- situation anxiety only. In each instance, however, only one of the human figures was differentiated from the other drawings and it was a different human figure for each physiological measure. Previous research by Handler and Reyher (1966) using both average conductance and SR frequency, and by Sanders and Reyher (1972) using SR frequency, have produced results with male subjects that differentiated each human drawing from the automobile. Careful examination of the methodology of the three studies reveals several differences which may account for the variation in results: (a) Sex_ of experimenter. Half of the male subjects in this study had a female experimenter while the experimenters in the earlier studies were apparently male. (b) Average conductance scoring procedure. Handler and Reyher obtained conductance readings at the beginning, middle, and end of each drawing. In this study, conductance readings were obtained every 16 seconds; since the average drawing time was 124 seconds, the utilized mean was typically based upon eight readings. (c) SR frequency scoring procedure. Sanders and Reyher defined a discriminable SR as an inflection of at least 500 ohms followed by a decline of at least 500 ohms. Handler and Reyher did 27 not present their definition. The definition in this study was an inflection of at least 1000 ohms, twice that of Sanders and Reyher. An examination of the results for the seven graphic indices of anxiety showed that five produced a significant main effect for stimulus object: shading, F(3, 267) = 10.53, p_< .Ol; erasure, F(3, 267) = 3.62, p_< .02; omissions, F(3, 267) = 15.69, p_< .01; line discontinuity, F(3, 267) = 4.50, p_< .01; and vertical imbalance, F(3, 267) = 7.78, p_< .Ol. Tukey tests were carried out on each of the significant main effects and revealed: (a) The animal drawing elicited significantly more shading than either the female or automobile drawing. (b) The female drawing elicited significantly more erasure than the automobile drawing. (c) Both human figures elicited significantly more omissions than either nonhuman figure. (d) The automobile drawing elicited significantly more line discontinuities than either human drawing. (e) The auto drawing was significantly more vertically balanced than the three others (see Figures 3 and 4). Of these five findings, only the omissions and vertical imbalance results clearly fit the hypothesized progression of increased anxiety from inanimate to animate to human figures. According to the omission index, subjects were more anxious drawing either human figure than either nonhuman figure. Similarly, according to the vertical imbalance index, subjects were least anxious drawing the automobile. The meaning of the shading and erasure results is less clear. Handler and Reyher (1964, 1966) found a higher score on the erasure and shading indices to be indicative of less anxiety. Applying their interpretation to these present results would suggest that subjects were less anxious 28 . .mmcwzmev cmlmwcoum zuwscwucoummv mcmp new .mezwmcw .mcwumzm com muumwwm arms uumnno mzpaawum .m assume homumo mains—Fm 3:< _uE_:< 2an“. 2o! _ _ _ _ . > J N mcamém YT J O J .232... 0.! / I \ 10 lo. I \ \ O J . W l o F a 0255..» I v .IJ.v.w mu 3 l m .l a; a J 3 IIJ N.N J ed 29 .mmcpzmgo TL co mucoum muse—mas? _muwpgo> new mcovmmveo Lo» muumemo spas uumnno mspaewum .e ocaovm howsmo mansfim 05¢ 35:2 22:0“. 035. _ a e _ . .ms..._ 0/ 1 u l l / .II III III all. all 0 I] C. . mzoamio CI 10 01 I m J 0.. V I N S J a; O J o no J m; a J J «d J 3D drawing the animal than either the female or automobile (according to the shading index); and that they were less anxious drawing the female than the automobile (according to the erasure index). Some of these interpretations, however, contradict the results of the physiological measures described above. The source of this contradiction may lie with the scoring procedure used in the present study. Although serious efforts were made to equalize the scoring procedures across all four stimulus objects, the final product was not formally tested. So the equality of application of some of the graphic indices to the four figures is questionable. For example, it can be argued that many animal figures have fur or feathers which, by definition, require more shading than the nonanimal figures. This argument robs the shading result of convincing intrapsychic meaning. Further testing is also needed to eliminate the possibility that the erasure and omission results could be explained by the unequal number of body parts defined for the animate and inanimate figures. For example, on the omission index, seventeen body parts were defined for the human figures versus fifteen for the animal figure and thirteen for the automobile. The line discontinuity index produced the most unexpected results. Subjects' automobile drawings had significantly more discontinuous lines than either of their human drawings. Since the scoring procedure was applied identically to all four figures and was not dependent on number of body parts, the previously discussed reservations about credibility seem inapplicable. A possible explanation related to time Spent on each drawing will be discussed in a later section. 31 Hypothesis II The results provided little support for the prediction of a progression of increasing anxiety for automobile, animal, and human stimulus objects in both story-telling and wish-making tasks. The average conductance variable produced a significant stimulus object main effect for the stories, F(3, 255) = 2.89, p_< .05 (see Figure 5). Additional analysis with a Tukey test showed that the female stories elicited a significantly higher level of average conductance than the male stories. This story finding is similar to the drawing finding for average conductance described earlier. In both cases, it was the female stimulus object that elicited the significantly high level of average conductance. For the drawing task, however, only the male subjects produced this difference, while both male and female subjects produced it during the story-telling task. The SR frequency variable produced a significant stimulus object by reSponse modality interaction for wishes, F(3, 255) = 4.51, p.< .01 (see Figure 6). Probing the interaction with a simple effects test revealed a significant stimulus object simple effect for the oral condition, F(3, 255) = 7.08, p_< .01. Testing these four stimulus object means by Tukey's method showed (a) the female stimulus object elicited a higher SR frequency than either the male or animal stimulus object; and (b) the automobile stimulus object elicited a higher SR frequency than the male stimulus object. For the written condition, there were no appreciable differences between stimulus objects. For the wish-making task, then, it is again the female stimulus object which elicited the greatest anxiety, although only for the oral condition (the response modality differences will be discussed in 32 MEAN AVERAGE CONDUCTANCE (deq 62" I l I I Male Female Animal Auto STIMULUS OBJECT Figure 5. Stimulus object main effect for average conductance scores on stories. 33 .mmcwzmgu co mmeoum zucmacmsm mm Low cowuumgmucw pounce mzpaswum an zuwpmuos mmcoammm Pomfimo maids—rpm oS< 38:3 29:0". 222 d _ fl _ zmtEB 0| 1 l 101 l l 10. I I I I nu lam—O .e «gamma 00. v0. no. 00. 50. mo. mo. O... 3.. AONHHOEEH HS NVSIN 34 greater detail under Hypothesis III). This SR frequency finding for oral wishes is similar to the average conductance results for male subjects' drawings and all subjects' stories: in all three situations the female stimulus object emerges as most anxiety-provoking. The SR frequency result for oral wishes conflicts, however, with the SR frequency results for drawings. In the drawing situation, male subjects manifested greater anxiety related to the male stimulus object; here, all oral condition subjects manifested greater anxiety related to the female stimulus object. Hypothesis III The results partially supported the prediction that the subjects completing the story-telling and wish-making tasks orally would demonstrate greater stimulus object related differences in anxiety than subjects who completed the tasks in writing. Only the SR frequency variable for wishes produced a significant stimulus object by response modality interaction; as described above, the oral condition produced several significant stimulus object differences while the written condition produced none. This supports the thesis that the writing of projective wishes allowed subjects to make more effective use of secondary process and so insulate themselves from the differential levels of anxiety provoked by the stimulus objects. The SR frequency variable also revealed a significant main effect for response modality during the story-telling task, F(1, 85) = 4.15, p_< .05. While telling stories, subjects in the oral condition had significantly higher frequency scores than subjects in the written condition. While this distinction between oral and written conditions is meaningful in itself, it does not address the issue of stimulus 35 object related differences in anxiety. It was hoped that this study would provide an explanation for Myers (1965) lack of significant stimulus object differences in SR frequency for written stories. From the present work, however, one can conclude only that subjects are more anxious relating stories orally than they are relating stories on paper. Subjects in the oral condition appear to be as insulated from stimulus object provoked differences in anxiety as the subjects in the written condition. Two additional important results were produced by the average conductance variable. A significant sex of subject by response modality interaction emgered for both stories, F(1, 85) = 4.65, p_< .05, and wishes, F(1, 85) = 4.03, p_< .05 (see Figure 7). Additional analyses showed that in the oral condition, male subjects had a higher level of average conductance than females while telling stories, 3(23) = 2.97, p_< .01, and while relating wishes, 3(23) = 2.32, p_< .01. These findings will be discussed further in the general discussion section. Other Anxiety Results The two physiological variables and the seven graphic indices of anxiety produced several additional significant main effects and interactions which were not directly related to the hypotheses above. There was a sex of subject main effect for the graphic index of omissions in which males made significantly more omissions than females, F(1, 89) = 5.50, p_< .03. There was also a sex of subject main effect for the SR frequency scores on the story-telling task; males had a significantly higher frequency score than females, F(1, 89) = 5.88, p_< .02. A sex of experimenter main effect was produced by the graphic index of line discontinuity, F(1, 89) = 4.50, p_< .05; subjects with 90 86 82 78 74 7O 66 62 58 MEAN AVERAGE CONDUCTANCE (x 105) 50 Figure 7. IFHI l 36 — Stories —-- Wishes MALES ’ I I ’ ’ 9 J Oral Written RESPONSE MODALITY Sex of subject by response modality interaction for average conductance scores on stories and wishes. 37 the female experimenter drew significantly more discontinuous lines than subjects with the male experimenter. Additional comments on these results will be made in the general discussion section. Two graphic indices produced significant sex of subject by sex of experimenter interactions: shading, F(1, 89) =Ip < .05; erasure, F(1, 89) = 5.66, p_< .02 (see Figure 8). Further analysis showed (a) female subjects used significantly more shading when with the male experimenter than when with the female experimenter, 3(24) = 2.58, p_< .02, and (b) female subjects also made significantly more erasures when with the male experimenter than when with the female experimenter, 5(24) = 2.25, p.< .03. As discussed earlier, Handler and Reyher (1964, 1966) found the shading and erasure indices to be associated with a reduction in the physiological manifestation of anxiety. They propose that subjects utilize shading and erasure as coping strategies. This suggests that female subjects in the presence of a male experimenter cope differently with the situation than either females in the presence of a female experimenter or male subjects. Finally, the sex of subject by stimulus object interactions for average conductance and SR frequency reported under Hypothesis 1 were reanalyzed to elucidate sex of subject differences. It was found that male subjects produced a higher level of average conductance than female subjects for the male drawing, 3(47) = 2.85, p_< .01, the female drawing, 3(47) = 3.09, p_< .01, and the animal drawing, 3(47) = 2.61, .p < .01. Further, male subjects produced a higher SR frequency than females for the male drawing, 3(47) =-4.39, p_< .01, and the auto drawing, 3(47) = 2.25, p_< .03. In the final section, these results will be discussed further. 38 . —— Female Subjects 2,4 _ - - - Male Subjects 2.2 ~— 2.0 -— 1.8 .— SHADING MEAN SCORE a l \ \ ERASURE Male Female EXPERIMENTER M. Sex of subject by experimenter interaction for shading and erasure scores on drawings. 39 Additional Findings for Drawings Subjects were permitted to determine which sex of human figure they would draw first. Six male subjects, or 12%, drew the female figure first; four of the six were with the male experimenter, two were with the female experimenter. Twenty female subjects, or 42%, drew the male figure first; eight of the twenty were with the male experimenter, twelve were with the female experimenter. The percentages of subjects drawing their own sex first in this study (males 88% and females 58%) are similar to those typically reported in the literature. Granck and Smith (1953) for example, found that 88% of their male subjeCts produced self-sex drawings first, as did 65% of their female subjects. A more recent study by Paludi (1978) reported 92% of the male subjects drew their own sex first while 63% of the female subjects drew the female figure first. When, however, the impact of the sex of the experimenter on the sex drawn first is considered, the results of the present study contradict previous findings. According to Paludi and Bauer (1979), an opposite sex experimenter elicits significantly more opposite sex first drawings for both male and female subjects. Here, however, it was the same sex experimenter who elicited more of the Opposite sex first drawings for both male and female subjects. Analysis of the total area (in square inches) occupied by the drawings revealed a significant sex of subject by stimulus object interaction, F(3, 264) = 3.45, p_< .02 (see Figure 9). Probing the interaction with a test of simple effects produced no appreciable differences between stimulus objects for male subjects. Drawings by males were all about the same size, with means between 50.1 and 4O .mmcwzmgu co mucoum omen Low cowaumemucw pumnao mapaswam an bumnnzm $0 xmm .m mgzmwu how .50 mains—Fm 35< .9524 29:0“. 035. Ll _ _ _ a i on \O / mm4