CHAPTER 2

PERCEPTION TASKS

by Hugh M. Lewis

The Perception tasks included a number of different kinds tasks that were independently devised as spin-offs of the MPDT. The Minnesota Percepto-Diagnostic Technique (or MPDT) was employed in the course of the study as both an empirical 'anchor point' for the partial grounding of other symbolic frame tasks. In the course of its administration a number of other alternative techniques were devised to get at some of the patterns of response which were being observed on the task, and as a means of overcoming certain basic limitations of the MPDT.

Thus tasks were systematically devised for "testing the limits" of the MPDT, especially for categories 2 and 3 for the procedure, as well as a task intended to elicit patterns of short-term memory and working memory. Another "rotating frame" task emerged somewhat unexpectedly from these developments, and though the evolution of its design led to several "fractured" sub-samples, its simplicity and understanding are interesting enough to warrant further research. Also, another "pattern test" was devised which shows interesting results in the processes of perceptual identification and recognition.

The cross-cultural and empirical robustness of the MPDT is rooted in the simplicity and efficacy of its design, and yet these aspects of its design, as well as its scoring and interpretation, constitutes also its central weakness--namely its lack of resolution and detail in understanding the variation of typical response patterns. When errors such as circle-diamond separations or distortions occur, this is supposed to be symptomatic of some form of mental dysfunction associated with organic brain disrepair--the problem is its lack of specificity in the diagnosis of exactly what kinds of disrepair produce which kind of error of response pattern. Thus, scores are only broadly correlated with general levels of percepto-cognitive integration and diagnostically wide categories of disorder. Problems relating to general nervousness or even test anxiety or distraction may produce scores high enough to categorize an individual as neurotic or even psychotic though the individual may appear normal in almost every other aspect. While an individual that is organically impaired will almost certainly score highly on the MPDT, the MPDT cannot by itself clearly separate the high scorer with actual organic or psychological disorder from another high scorer who may only have apparent disorder as manifest mostly by such a high score.

It is apparent also that the individual scores of the MPDT can and often are conflated by the lack of attending carefully to the task, such that separations or distortions which occur may be the result of this failure of attention (and may sometimes also indicate a kind of "attention deficit disorder"). These kinds of errors are not sufficiently controlled.

In giving the technique to a number of different people, apparent patterns are: 1. its empirical efficacy in indicating the possible presence of some kind of manifest disorder; 2. its lack of specificity in diagnosing the latent cause of the underlying functional disorder; 3. its susceptibility to contextual interference and to the current anxiety state of the subject.

"Testing the limits" of the MPDT is a recommended technique in partially correcting for this weakness, except that this follow-up procedure is itself too weak and too dependent upon the analyst's initial and gross evaluation of the patterns of the figure.

Thus the idea of testing the limits was extended systematically with the intention of deliberately controlling for and diagnosing potential problems with 1. visual perception. 2. visual-perceptual integration. 3. hand-eye coordination. Furthermore, the second potential problem of visual-perceptual integration was further extended to include systematic and hypothetical evaluation of short-term and working memory.

Alongside of this, a more detailed system of analysis of the figure-pattern of the original MPDT was applied, entailing evaluation of distinct patterns of distortion or rotation error--whether along vertical or horizontal axis, whether to the left or the right of the axis. Some correlation was sought with actual problems of vision or visual perception, as well as with other problems such as alcoholism, general anxiety, age, etc.

At the same time, an alternate version of the MPDT, created on individual record sheets instead of the cards, and in reverse sequence, was tested out and compared with the results of the original MPDT.

Two other types of tasks that are considered more closely related to the MPDT and its extensions, as perceptual and figure-frame and figure-field tasks, include the Pattern Task and the Rotating Frame Task, and these were developed and refined in the course of fieldwork related to extending "testing the limits" of the MPDT.

In the evaluation and comparison of individual patterns of response to the MPDT, it is necessary to presume a basic level of test-competence by the average informant above a certain age indicative of sufficient cognitive development. It is as one illiterate informant told me as she tried to do the task--she had hardly ever held a pencil before. Because these were "performance" tasks on which subjects were scored and compared to established norms, as with all "performance" tasks there is an inherent bias in that with repeated retesting, subjects can be expected to show significant "improvements" on their scores. This is probably the case with practically every kind of performance test imaginable. It means that different people are bringing different backgrounds of experience and cognitive skills into the performance which are implicated in the "show". It also renders tests such as IQ, achievement and performance type tests evaluations of mental-psychological "showmanship" than necessarily of inherent capacities.

Furthermore, these tasks are justified not for the sake of the scores themselves, which are taken with a grain of salt, but in the detailed patterns of response which different informants showed, and in the inter-correlation between the sub-samples and the different nature of the tasks.

This brings up the question of the scores of young children and adolescents. There were no IQ scores so that the raw scores of the below 15 year olds could not be adjusted. The possibility of computing such "scores" exists in the analysis of their human figure drawings, which was done, but because these scores are held to be highly unreliable and probably relatively invalid, and also in part the consequence of a child's motivations, skills and background in drawing, they were not used. It rendered one very bright 8 year-old girl who liked to draw a lot for me an IQ of 220 and a mental age of 13 years old, and another bright 5 year old who was good at drawing little girls an IQ over 180. Even if these scores were used it would leave a gap for the 14 and 15 year olds. The "below-15" year-olds raw scores were therefore earmarked and left unadjusted for the sake of overall consistency in the evaluation of patterns of response in the overall sample.

Nevertheless, the alternate task designs that were spin-offs of the MPDT were set up in such a way that high scores were significant of poor performance, so that they could be more easily corroborated with the MPDT.

Sample sizes are small for a number of reasons. The development of new tasks in the field setting led to the fracturing of several otherwise quite substantial samples. At first the receptivity of the Jetty Chinese to these tasks was quite high, often because they thought I was measuring IQ. A number of parents brought their children to me with the idea that I could tell them how bright their children were. Chinese love to think of their kids as more intelligent than others and will do anything and everything they can to 'hot' house their children to an early road to success. They backed off and grew suspicious of my motives which they did not well understand, and thus the willingness to do these tasks rapidly decreased afterwards. The community then grew to think that I was doing some kind of brain surgery to steal secrets from their heads, perhaps so that the American government might afterwards use this information against them in some way.

The MPDT task was divided into several sub-samples: a non-Jetty Chinese sample, a Jetty Chinese sample, and a sample of my reference group drawn mostly from the Jetty. The Jetty Chinese sample was divided into a male and female sub-sample. The tasks were analyzed for their raw scores, for the number of circle-diamond distortions and separations which occur, as well as for the relative enlargement/reduction of the drawn figure from the presented figure, the direction of the axis of rotation from the vertical, and minor distortions and separations not normally scored. The total sample of the Jetty consisted of 27 scores, of which the average raw scores was 36. There was a .433 positive correlation between the raw scores and the adjust score for minor distortions, a .6072 positive correlation between the adjusted score for minor distortions and the adjusted score for diamond distortions, a .3126 positive correlation between adjusted scores for circle-diamond separations and diamond distortions. There was also a negative correlation of approximately .2 between the adjusted scores for the reduction and the number of distortions and separations in all the categories, as well as for age. There was a positive correlation between age and diamond distortions (.2672) and minor distortions (.1889).

Below is a graph of the scores of the total sample.

These scores are comparable to the raw scores of the reference group (N 19), which had an average of 32.1. Correlations between raw scores and other scores show a negative correlation between reduction and circle-diamond separations (-.263), raw scores (-.117) and minor distortions (-.11). There was a negative correlation also between age and raw score (-.253) and minor distortions (-.345), as well as between raw scores and reduction and diamond-circle distortions (-.117 and -.136, respectively). There was a high positive correlation between minor distortions and circle-diamond distortions (.509), raw scores and minor distortions (.207) and between raw scores and circle-diamond separations (.199). The graph of the raw scores of the reference group is shown below.

A small non-Jetty sample was collected as well (N= 18). The mean was about the same, 32.8, though the median of 23.5 was much lower than for the reference group, 26.5 and for the Jetty sample, 29.5. This sample was thrown off by the inclusion of an elderly woman and a young boy who had abnormally high scores. There is a perfect positive correlation between diamond distortions and age, between age and reduction scores (.51) and high positive correlations between diamond distortions and raw scores (.95), minor distortions (.58), reduction scores (.50), circle-diamond separations (.47) and circle-diamond distortions (.4). There were also high positive correlations between raw scores and all the other scores, except for age. The graph of these scores is presented below.

The more interesting difference is when the Jetty sample is split between males and females. Males then show up with a higher average score of 46 with a median of 37 and a mode of 44, compared to an average score for the women of 25.1 with a median of 21 and a mode of 8. Though the samples are too small to mean anything, it does hint at a possibly significant difference between men and women, and it is the kind of difference that is corroborated by observation. For the men there was a positive correlation between minor distortions and circle-diamond distortions (.5153), age and circle-diamond separations (3.003) and minor distortions (.2845), and between circle-diamond distortions and circle-diamond separations (.3056). There was a strong negative correlation between raw scores and circle diamond separations (-.284) and circle-diamond distortions (-.285) and between reductions and minor distortions and circle-diamond distortions (-.125 each).

For the women's sample (N= 13) there is a high positive correlation between circle diamond distortions and minor distortions (.87), and raw scores (.77) and age (.44) and circle-diamond separations (.34). There is a high correlation as well between raw scores and minor distortions (.61) and between age and raw scores (.25) and minor distortions (.61). There is a negative correlation between reduction and age and minor distortions (1.4 each) and circle diamond separations (-.2) and between reduction and circle-diamond distortions (-.3) The graph of the women's raw scores is presented below:

Another interesting difference was in the difference of frequency scores of the direction of rotation between men and women of the Jetty. This difference was not significant for the reference group, nor for the non-Jetty sample, so may therefore just be a fluke, nevertheless it is a pattern present in the data worth mentioning. The chi square test for significance is 6.611, significant above the .025 level.

The average score of the males of the Jetty is well into the region normally regarded as indicative of neurosis, compared to a lower but still marginally neurotic score for the women, or for the reference group or the non-Jetty Chinese. These scores are not to be regarded as significant samples in any sense, but they are suggestive of a larger overall pattern

Besides the MPDT, a systematic means of "testing the limits" of this tasks was devised as a series of multiple choice tasks of images of the cards with the figures in different orientations. This task was especially meant to help assess any relative level of impairment of perceptual integration of information. A small sample of 31 subjects was collected. A paired variates T-test with MPDT scores with a sub-sample of 15 sets of scores shows no significant difference in the two sub-samples at a level of .05 rejection.

There is a positive correlation of .215 between these sets of scores.

These statistics suggest that though the sample sizes are small, the task may be measuring to some degree what it was designed for, that is the relative level of perceptual ability to integrate information in an increasingly "noisy" or ambiguous field. There are four pages of three multiple choice per page. By the third item on each page, there is noticeable improvement of performance as the subject tends to "figure out" the pattern and to disambiguate the figures from the ground. But with each successive page the figures are increasingly ambiguous, first reversing the direction from left/right to right/left, and then eliminating the direction altogether, and finally changing the solid lines of the background frames into dashed lines. The frequency pattern of these 12 successive choices is given below:

Graph 12-5. Histogram of Sequenced Item Error

Besides this task for systematically testing the limits of the MPDT, especially in relation to the perceptual integration of information, an alternate "short term memory" task was devised that was parallel to the MPDT and that was designed to get at the process of the individual's retrieval of information from "short term" just perceptually processed by the presentation of cards before their removal. Unfortunately, only a small sample of this task was accomplished (N = 17), due primarily to over involvement in other task areas and neglect of its value.

The task was analyzed in the same manner as the MPDT, and placement on the page was introduced as a control in the process to help disambiguate the field for the subject. It is interesting that the average adjusted score is 21, with a median of 13 and a mode of 8, reflecting an overall tendency to reduce the size of the drawn image by about 2/3rds, compared to the usual one half average reduction score of the MPDT. It may represent the loss of perceptual information in the short-term memory. Overall, the average score was 23 with a median of 22 and a mode of 24. This reflects a better score than that of the MPDT, and this is reflected in the very slight negative correlation between high scores on the MPDT and lower scores of the short term memory task (-.0081, N = 11). These scores are in spite of the fact that the short-term memory task entails eight rather than six items, and thus is inherently more error prone towards higher scores than the MPDT. Two types of errors became apparent with this task that is not clearly a part of the MPDT (although there are some noteworthy exceptions, and this points up an important, yet unexplored facet of the MPDT). These are in the number of reversals and rotations that occur within the figure itself, though the overall figure is drawn correctly. There is a noticeable tendency to draw the figure correctly, but to reverse in the mirror image some of its elements. This same pattern occurred occasionally with the MPDT and made these difficult to score within the current system.

Correlations between the different scores on the tasks reveal that there is a negative correlation between reduction and raw scores (-0.2738), minor reversals (-.3002), separations (-.2087) and distortions (-.2216) suggesting that the reduction of the size might be important in understanding how the performance is improved when relying upon memory, as well as between frequency of full rotations and minor distortion scores (-.1815) and minor reversals (-.2794), and distortions (-0,4168). There are slight positive correlations between reductions and minor distortions (.154), raw scores and minor reversals and distortions, as well as between minor distortions and minor reversals (.4995), separations (.483) and distortions (.4556), as well as between minor reversals and separations (.83206) and (.39059).

This pattern suggests that while there is increasing potential for minor error, as evidenced by minor distortions, distortions, separations and minor reversals, the reduction of the image entails an overall improved performance which is correlated with low raw scores, and slightly with major rotations. Attending to the minor details, the overall axis of orientation may be simple "left out of the picture" or "forgotten". This approach definitely warrants further research.

Correlations of items of combined cumulative frequencies of minor distortions, reductions, separations, distortions and raw scores show remarkably high correlations. The same cumulative frequencies correlated in terms of the scoring dimensions also show relatively high correlations.

Table 12-6 Correlations of Scores

The "rotating frame task" was a direct spin off of the development of extensions of the MPDT, and concerned only the axis of rotation in relation to the frame. The axis of rotation became represented as a solid line. This line was kept in constant position as the frame about it was rotated one quarter turn. The original position of the frame was slowly revolved from a horizontal to a diamond position, and the original axis of the rod shifts from the vertical, to the horizontal, and to the diagonal, and then to the oblique. It's clear relation to embedded figures tasks, the rod and frame task, etc. is obvious. But the exact nature of this relationship remains to be well defined. Unfortunately, the elicitation of this task suffered a "fractured sample" as it underwent four successive version of development in the field. The first version was tested with a sample of eleven but it did not produce the kind of response pattern I was looking for. It seemed that the central rod was heavily emphasized with a dark line, while the surrounding frame was a very thin and relatively unnoticeable line. This combined with the fact that the rod was consistently disconnected from the frame, allowed most subjects to easily separate rod and frame and to perform in a manner that generated low scores. The average score on this task was 9 with a median of 8.5 and a mode of 4. The histogram revealed that it was heavy on the low-score end with 9 people scoring below four and only 2 scoring higher.

The central rod was made lighter in weight than the surrounding frame, which was emphasized, and the line was extended on both sides to actually touch the frame. The order in which the frame and the rod were rotated on successive frames was also altered. The informant is told simply to "draw the line in the center of the square as you see it and in the easiest way they can.

A sample of 30 was collected with the revised version. The mean was 14.9 and the median was 16 and the mode 5.5. A histogram of the frequency patterns of this sample showed that 14 of the sample scored in the below 2 range. Five persons scored in the below 9.8 bin, 7 in the below 17.6 bin, 2 below 25.5 and 2 below 32.2. It is difficult to draw a cut-off point at which score might represent something like "frame dependency".

Rather the task shows a clear gradient between relatively "frame independent" and extremely frame dependent response patterns. The scoring of the system was rather simple and straight forward, and allows for intermediate .5 scores, and is fairly clear and reliable.

Because there are forty eight items, the total score can be multiplied by a factor of 2.083 give a solid percentile score. While a gradient is clear in all the samples taken, the graph below of the raw, unadapted scores of the N=30 sample, show a clear demarcation line at the above 15 (15 x 2.083 = 31.245) which might be taken as significantly splitting the sample between "frame dependent" (N=15) and "frame independent" (N= 15) halves.

The orientation of the true frame was either horizontal or diagonal, and these were taken as relatively more basic and less ambiguous in relation to either a vertical, horizontal or clearly diagonal rod.

The task was also set up in such away that the items were arranged in four groups of 12 each, with scoring patterns in each subgroup representing different approaches to framing or drawing the rod. Correlation with the MPDT of a sub-sample of 15 shows a slight negative correlation between the tasks of -.02581, showing a lack of any obvious relationship between hypothetical "frame dependency" and the "perceptual integration" as measured by the MPDT. It may well be that there is little correlation between the two tasks because they may be eliciting precisely the opposite responses--the MPDT the ability to pay attention to the figure in relation to the ground, the rotating frame the need to attend to the figure regardless of the frame.

Inter-item correlation was done. Because 48 items in sets of two and in subgroups of 12 and sub-sub groups of 6 makes for a great deal of complexity, complete analysis is beyond the scope of this report. Suffice it to summarize that most items show a fairly strong positive correlation with one another (.2 -.6) while only a few items show no correlation at all or only a slight negative correlation (below -.2)

In terms of what this task implies, it is not clear. In scoring the tasks it was clear that some subjects were for the most part unable to ignore the information of the frame in defining the axis of the rod. It was easier for these subjects to draw the line in relation to the frame than independently of the orientation of the frame. For an intermediate group, there was a decreasing response of frame independence, and an increasing pattern of frame dependency. The graph below of the frequency pattern of the errors per item shows the pattern of increasing error toward the end of the task, successively punctuated by the low scores of presenting the "true" frame. It is apparent that with each set, more "frame dependent" subjects were repeatedly, and increasingly, unable to ignore the information of the frame in drawing the rod.

What emerges from analysis of this task is that there is a clear and strong positive correlation between the age of the subject and high scores. I would hypothesize that disturbed perception is largely acquired, as is "frame dependency", and we can refer to people who have learned how to depend upon external cues of contextual information as a primary cognitive strategy. It is not difficult to imagine how some contexts can be more difficult to ignore than others, and how some may also be more "confusing" than others, making attending to the context a difficult but necessary means of survival.

Correlation using Fisher's exact technique for analyzing significance of differences between men and women who score 15 or above on the N=30 sample and those who score below 15 show no significant difference, though the sample size was too small to really tell N=27.

The correlations relating age to high scores is ambiguous. The first and weakest task (N = 10) has a negative correlation between scores and age of -.41. The second and strongest task (N = 30) has a positive correlation between these two factors of .248. The third task (N = 19), transitional from the third to the fourth, has a stronger positive correlation of .364 and is noteworthy because this was perhaps the best task, generating a clearly bipolar histogram between high and low scores (10 below 1.75, and 9 above, out of a total score of 12). The fourth and final sample proved the most difficult, and had a low negative correlation of -.044. The sample was the smallest (N = 10) of which only 2 scored below 2.5 while 8 scored above.

If these statistics have any significance whatsoever, their implications about the effect of the environment in influencing "frame dependency" cannot be ignored. The positive correlation with increasing age seems at contrast with the dominant theory explaining field dependency which rests upon the differentiation between self in the disambiguation of the perceptual field. It suggests that the environmental factors influencing the development of these broad cognitive styles may be multi-factorial, and that different factors may be having cross-cutting effects. Increasing separation of self from environment may be at odds with the increasing control of the environment over the self, leading to a kind of double bind. Children do not feel their servitude, parents do not know their tyranny.

The final perception task given was in some ways the most interesting and reveals with greater detail the processes of recognition and object identification that may be involved in modes of perception. The pattern text consists of 9 pictures each of 8 common objects (half of an apple, a fish, a spider, a horse, a flower with stem and petals, a human skeleton oriented along the horizontal axis, a cat's face and a mouse). Each successive picture of each series presents a more definite outline of the object. The first pictures are relatively formless, while the last couple of pictures show the form in a clear and unmistakable manner. Subjects were asked to turn through the pages and tell us what they saw in each, if anything. The pages were in a booklet such that the same order was used each time.

A substantial sample was collected (N = 41, 18 males and 23 females). People were scored at the page at which they first had recognition of the object. At the same time, notes were kept of the details of individuals' response patterns, indicating what they thought they were seeing in the pictures until the "true form" of the object became apparent to them. What is interesting is how pattern recognition is preluded with much processing of details and trying out of different possible forms, and at some point, the "aha" pattern of the gestalt emerged, or else failed to clearly emerge at all.

The tasks were further analyzed by allowing individuals of the reference group to take copies of the pictures in order and to draw in the pictures of things they saw on each page.

The first set of pictures consists of half an apple. The average response score was 9, with a median and mode of 10. The round, symmetrical shape was difficult for people to analyze. The histogram of this item showed that only one person had a score of 6, 7 people had a score of 7, five scored 8, 4 scored 9 and 27 scored 10. People saw a round shape, usually by the 6th or 7th page. Otherwise people saw elephants, eggs, bears, rats, hearts, tomatoes, pomegranites, insects, butterflies and cats and dogs. The drawing analysis revealed that people were drawing part of the hemispherical outline by the second or third card, which became fully connected by the 4th or 5th page, and then a stem is attended by the 6th to 8th page, until finally the apple configuration is put together.

The second set of pictures was a fish. The average response score to the fish was 6.5 with a median and mode of 6. It appears that unlike the apple, the asymmetry of the fish provided early cues for the subjects to put together the picture. The frequency histogram of the fish shows that 1 person guessed it on the second page, 2 by the third page, none on the fourth page or fifth, and then 35 guessed it on the 6th page, 1 on the 7th and 5 by the 8th or 9th and 2 couldn't guess it at all (10). One person saw a cat, one a human, two saw a bird, one a crab and one a elephant. On a couple, seeing an alternate shape seemed to block and preclude recognition of the true form, and represents a frequent pattern in which projection of an image onto a vague field tends to obfuscate perceptual recognition of the pattern which is already there. This happened frequently enough, and especially by certain individuals, as to warrant consideration. This pattern shows the degradation of gestalt response on the fourth and fifth as the subjects are attending to details of the figure outline--for instance tail, fins, and gills of the fish--just before the final gestalt pattern emerges. This pattern of building a rough, cartoonish image, then degradation of response as details are organized, and then sudden gestalt of pattern recognition recurs repeated in the drawing analysis of the figures.

The third set of figures were of a spider which was also symmetrical. The average response score of the spider was 8.57 with a median of 9 and a mode of 10. Again, the symmetry of the figure, like that of the apple, seemed to provide no perceptual "handles" on which the subject could anchor the cognitive construction of the gestalt. The histogram of this response pattern was 1 person by page 2, 2 at 3, 1 by the fourth page, 2 by the sixth page, 10 by the seventh page, 9 people by the eighth, 12 by the ninth, and 15 could not guess it. This is significant because it was a relatively simple picture to see. One person saw a vague animal, one a bee, one a fly, one a tiger, one saw wild grass, two a girl's head, one a "vague shape", a vague "animal's head", an "arrow & blob", a cat's face, a tiger's head, a beetle, 4 saw a crab (10%) , and one saw "King Kong". One fairly bright individual saw a flower, then nothing, then a flower again, then a frog, a rose, an orchid and finally a crab, never guessing the spider.

Drawing analysis of this task revealed that people began drawing small figures, but not in the center of the picture, for one a butterfly, then a spider emerges by the third page, changes shape and position on the fourth card, and then disappears until the end. This pattern of seeing different "de-centered" things in the first sets, then nothing, and finally either seeing or missing the true gestalt is repeated by all the subjects. Fifty percent recognized the spider by 8 or 9, and fifty percent found no recognition.

The fourth set of pictures was of a horse, and again the asymmetry of the figure seemed to make it easy for people to recognize. Perhaps like the fish, there is something prototypical of the four legged horse with ears and a tail that makes it easy for people to recognize. In this regard it is interesting that the names and shapes of animals are some of the first things children acquire, even before colors and other basic shapes. Perhaps there is something primordial, of the cave paintings of our ancestor's, which is embedded in our brains. The average response score of the horse was 5.11, with a median and mode of 4. Seven people guessed it by the third page, and 16 by the fourth page, five by the fifth page, one by the sixth, four by the seventh, one on the 8th, one on the 9th, and four couldn't guess it at all.

Drawing analysis of the figure reveals that the outline gradually emerges by the second or third page at certain areas, the legs, tail, mane, head, and shoulders. These lines become joined and then becomes progressively more refined. One person of four couldn't guess the figure, though a rough outline of a dog was drawn by the second page. It is interesting that a great many people confused the horse with the dog (19 of 40, 47.5%) and of thee seeing of a dog precluded recognition of the horse for five people (12.5%). Another person saw an elephant. It is interesting also that some people would carry the recognition of the dog up until the eighth page, at which the clear outline of the horse suddenly caused them a little embarrassment.

The fifth set of pictures was of a flower that combined asymmetry of the stem and leaves with the central symmetry of the flower itself. This was the most complex of the pictures, and was one of the more difficult. The average score was 7.95 with a median and mode of 8. The histogram of the fifth set shows that only one guessed it by the fifth card, six people guessed it at the sixth card, three on the seventh, 18 by the eighth, six by the ninth and five didn't guess it at all. People saw a body or head, a bird, a blbo, a tree, a horse and a boat. One person saw a deer which was carried up until the end, and a horse was carried up until the eighth page, by which the sudden recognition of the flower occurred. On person saw a cat's face until the sixth card, then a pear, then a pineapple, and finally the flower by the ninth. On person saw a flower on the first and second page, then a coconut, then Garfield, Snoopy, then there was degradation of pattern, and finally found the flower.

Drawing analysis of this shows small fish shapes, or nothing, and then the gradual emergence of a flower, with an erose of the outline by the 6th or seventh page, and its clearer reemergence by 8 or 9. The flower changes shape, position and appearance throughout the drawing sequence.

The sixth set of pictures consisted of a human skeleton which was turned to the horizontal axis. The asymmetry of the figure led to an early recognition of the human form, usually by the fifth page, but recognition of the actual skeleton did not come until later, with an average response score of 8.3 and a mode and median of 9. The histogram reveals that two peole saw the skeleton by the fifth page, four by the sixth, two by the seventh, 10 by the eighth, 16 by the 9th and 8 didn't guess it at all. One person saw a cat, two a shark, one an "animal", two saw a crocodile, two saw a fish, one a "diver in the sea" and another "somebody swimming or floating, arms and legs". 23 (57.5%) saw a figure of a person, either a human, man, woman or "a person." One person saw a fish, then a scorpion, then a lobster, then a crocodile, then a lobster again, then a woman, and finally a skeleton by the eighth page. It appears that the asymmetry allowed early recognition of the human form, but the shape of the human threw off recognition of the actual skeletal appearance and details until the end.

Drawing analysis of the skeleton pictures shows the gradual emergence of a central figure by the third figure, then nothing at the fourth page at which the figure was intentionally reversed, and then the outline of a human form by five, becoming refined by the sixth, then erosion at the 7th and 8th page, with attention to the hands and head. Only one person in four found the skeleton by the 9th page.

The seventh figure was of a cat's face. It's symmetry and general shape led a number of people to see a butterfly before a cat (14, 35%), one a human face, one a dog's face, one a ribbon, one an elephant, one a bee, and one a map. The butterfly seemed to postpone actual recognition by the eighth or ninth card, despite the emergence of clear clues, whiskers and ears, by the fifth and sixth card. The average response score was 7.39, with a median and mode of 8. Four saw it by the fifth page, 8 by the sixth, 6 by the seventh, 12 by the eight, five by the ninth, and four didn't guess it at all.

Drawing analysis of the cat's head shows cartoonish "Mickey Mouses" emerging on page one and two on 50 percent, and a ribbon by another. There is definite erosion of the pattern by the 6th and 7th card, with the reemergence of the "true" form by 8 or 9. The "Mickey Mouse face" is an interesting projection, and is studied more thoroughly in the analysis of the children's drawings.

The final figure was of a mouse. It was relatively asymmetrical in appearance, with it's long curving tail in the foreground. The average score was 6.89 with a median and mode of 7. One person saw it by page two, because of the tail, one by page four, five by page five, four by page 6, then 18 by page 7 (45%), then 3 by page 8 and and five by 9 and two couldn't see it. Three people initially saw a fish, two a head, one a crocodiles head, one a rabbits head and one a cat. One person confused the mouse with a shrew, and one with a lizard. Three saw a rat. Drawing analysis reveals a similar type of pattern, with a Mickey Mouse face and a heart and monster appearing on the first page, then the appearance of a crude, cartoonish rat form, then its erosion and final reappearance by everyone by the 7th page.

To interpret the response patterns of this pattern test, it appears that asymmetrical forms were more difficult to recognize than asymmetrical forms, as evidenced by difference in average scores. The chi square test comparing of scores of 10 (no correct response) with scores of 6 or below, of the four symmetrical figures with the four asymmetrical figures is 34.996, significant way past the .005 level.

It also appears that certain unclear prototypical forms are frequently confused with similar types of things, and this confusion can delay perception of the true form. In attending to the whole pattern, there is a recurrent erosion of 'gestalt" typically by the sixth or seventh card, while the individual is attending more to the correctness of details of the image. Then the image suddenly reappears in more true to form manner by the 8th or 9th picture, often as an "aha" of sudden recognition, especially if there is previous confusion of forms or lack of overall gestalt recognition.

Analysis of the gross frequency patterns indicates that no one saw the correct image by the first page. Only three saw the image on any of the cards by the second, 11 by the third, 21 by the fourth, 19 by the fifth, 38 by the sixth, 56 by the seventh, 84 by the eighth, 55 by the ninth, and 75 could not see anything. The graph below shows these relative frequencies, the peak at the eighth position. The postponement and erosion of the pattern recognition is indicated by the four and fifth positions.

The total scores of subjects was correlated with the age of respondents, revealing a slight negative correlation of -.141. Older subjects in general do slightly better than younger subjects. Chi square test comparing men and women between those with an average score of 6 and below and those with an average score above 6 is .7466, showing significant difference between the two samples only above the .5 level of rejection. Thus there is no significant difference between men and women's scores. Unfortunately, the three main samples of the MPDT, rotating frame tasks and pattern tasks did not overlap enough to allow for correlations.