Psycho-Geometrics® Research Report
The Concept/Assumptions: In 1978, Dr. Susan Dellinger combined her knowledge of the current personal style analysis tests on the market (MBTI, DISC, Wilson Learning, DrakeBeamMorin) with her doctoral research in the classical Jungian psyche types, left/right brain dominance and the current interpersonal communication models. The result was the creation of five “communication styles” subsumed within five (5) geometric shapes (square, triangle, rectangle–left brained; the circle and squiggle–right brained). Dr. Dellinger defined a series of traits and characteristics associated with each of her 5 communication styles.
Research Period: To validate Dr. Dellinger’s assumptions about the shapes styles and their
definitions, the following research was completed in 2 phases (Part 1 and Part 2) over a
three-year period from 1984 thru 1986.
Methodology: Self-Report. Unless using a multi-rater system, individual style analysis instruments/tests are based on what psychologists call “self-report.” Individual results can only reflect how the individual perceives himself/herself during the period of experiencing the test. Results are neither wrong nor right, merely providing some suggestive insight for deeper reflection and growth on one’s life journey.
Test Subjects: Dr. Dellinger presented Psycho-Geometricstm System in public seminars sponsored by CareerTrack Seminars International throughout the United States from 1984 through 1998. Participation in these seminars ranged between 50 and 200 individuals per
session consisting of adults from a wide variety of professions including business, education, healthcare and government. Participants represented all social/economic/racial groups. Approximately 60% were female and 40% were male.
Part 1: 1984-1985 (2,520 participants)
Part 2: 1986 (1,110 participants)
1. Participants receive an evaluation form before Dr. Dellinger’s presentation.
2. Participants are asked to observe an simple image of the five Shapes and are asked to (quote) “choose the Shape that best describes you as a person.” They are then instructed to draw their Shape choice on their evaluation form.
- Dr. Dellinger then proceeds to describe the characteristics of each of the Shapes.
After all five Shapes are described, the participants are asked to circle the Shape on their evaluation if it represented them or to put an X through it if it did not. Forms are collected and tabulated.
Part 1 Results: 79% of the participants reported that their original Shape choice did correctly describe them as a person, i.e. validation via Self-Report.
Part 2 Results: In 1986, participants were asked to identify their 1st and 2nd Shape choices by drawing both (same as #2 above). After Susan’s presentation, they were asked if either of their choices accurately described them. 84% reported in the affirmative.
Statistics: Shape choices by percentage of respondents in study: (1) Circle = 32%,
(2) Triangle = 28%, (3) Squiggle = 18%, (4) Box = 15%, and (5) Rectangle = 7%.
Psycho-Geometrics® Online Test/Instrument Research
The Rationale: In response to the request from many Psycho-Geometrics® users, Dr.
Dellinger and her colleague, Dr. Jack Wolf, designed a written instrument consisting
of 168 questions/items. In 2004, they enlisted the help of a team of graduate students
specializing in psychological test and measurement methodologies at the University of
California, Santa Cruz, in order to test the validity and reliability of this instrument.
The Concept/Assumptions: In order to establish the face and content validity of the
instrument, the 168 questions were grouped into five (5) behavioral factors (introversion,
non-conventionality/freedom, relationship orientedness, thinking-orientation/control,
and self-centeredness/interpersonal authenticity) and measured against five (5) “shape”
factors (box, triangle, rectangle, circle and squiggle).
Test Subjects: A sampling of 264 test subjects agreed to participate in the initial
testing of the written instrument. Subjects were chosen from the combined address
book lists of Dr. Dellinger and Dr. Wolf. The lists were comprised of both personal
and professional acquaintances of each author. Of the 743 individuals who were invited
to participate, 35.5% agreed to participate: 68% of participants female, 32% male.
Process: Participants were asked to take the Psycho-Geometrics® test via the Internet.
The test administrator, Mr. Steven Bearman, set up a private test site on the Internet where
subjects could come to take the test. The results were immediately transferred and
reported to a U. of CA collection site. Interpretation of the results was the responsibility
of Mr. Bearman and his graduate level team. Participants were given two months to
accomplish the task which required approximately 20 minutes to complete.
Methodology: Factor analysis to answer the question - when the actual factor model derived from the data is compared with the ideal actor model of the 5 shape factors, is there a relationship between the two models. Do the behavioral factors correspond well with the shape actors in our ideal model?
Summary Findings: If a question corresponds to only one shape, the ideal correlation is 1.0. For two or three shapes, the correlations are split equally resulting in correlations of 0.5 or 0.33 (rounded down) respectively, for each shape-question correlation. If Factor 1 introversion) was, for instance, closely related to the Circle Factor, we might expect many or most Circle questions to fall under Factor 1 and few questions in any of the other shape categories to fall under Factor 1. As it is, all shapes have between 9 and 16 questions represented in Factor 1. Introversion, the true factor underlying the data, does not correspond exactly with any one shape. While both the Box and Circle Factors correlate positively with some introversion questions, Triangle and Squiggle correlate negatively with some others (or positively with extroversion).
Overall, Factor 1 (introversion) does not correspond well with any one shape. Factor 2 (non-conventionality and freedom) , however, has far more Squiggle Factor questions than any others, with Rectangle Factor matching somewhat well. Factor 3 (relationship orientedness) has a decent match with the Circle Factor. Factor 4 (thinking-orientation and control) has a decent match with Triangle Factor and an okay match with Box Factor. Factor 5 (self-centeredness and interpersonal authenticity) loads across all the Shape factors and does not correspond well with any one shape. The best matches are Factor 2-Squiggle, Factor 3-Circle, and Factor 4-Triangle.
There is a clear lack of relationship between Circle and Factors 4 and 2 (thinking orientation / control and non-conventionality / freedom), a weak relationship between Squiggle and Factor 4 (thinking orientation and control), and a weak relationship between Rectangle and Factors 3, 4, and 5 (relationship orientedness, thinking orientation / control, and self-centeredness / interpersonal authenticity).
Recommendations: These results are optimistic for being able to revise the shape model, if such a project is to be embarked upon. It means that at least these three shapes have a medium correspondence with real factors underlying our survey data. It also shows that the questions associated with the shapes need to be reevaluated somewhat if they are to match real factors. It is also important again to remember here that the 5 factors in our analysis explain just 29% of the data.
In order to improve the model (instrument), we recommend the following: (1) the first step is to drop all variables (questions) which do not correlate well with any factor. These are the questions which fall below the double horizontal line in the first five tables. After dropping these from the results, a second factor analysis is run to see if a small number of factors now provides increased explanatory power (well above 29%); (2) Alternately, questions can be eliminated which do not support a match between numbered factors and shape factors.
Action: Doctors Dellinger and Wolf accepted the findings and recommendations of the Bearman
team. A valid and reliable instrument was created via the systematic reduction of noncorrelating variables from 168 to 89 (52.9%).
1) Shape and Color Choice
The Bearman team found only a “slight correlation of statistical significance with a p value of .631 between shape choice and color choice.” Of all colors offered, each shape preferred blue to any other color with gray being the least chosen color. Therefore, the assignment of colors to shapes (blue box, red triangle, yellow squiggle, green circle, gray rectangle) is determined to be largely arbitrary.
2) Visual versus Instrument Choice Correlation
The Bearman team found that visual shape choice was, indeed, a good predictor of the objective
outcome of the written instrument. The team reported a significant correlation of .345 (p < .0005) between respondents’ shape choices and their highest scoring shape on the instrument. Another way of describing the relationship between the two variables is that 12% of the variation in top scoring shape is explained by shape choice. (“Variation” is the statistical difference between random distributions of answers and actual answers.) Shape choice matches highest scoring shape 36.4% percent of the time. If all 5 shape scores are used to predict respondent shape choice, the results are similar (r = .388, p < .0005).
One exception with minor variance is the tendency of subjects to choose the Circle most
frequently in the visual test whereas the Triangle scores highest on the written instrument.