Art and madness: Can the interface be quantified?

The Sheppard Pratt Art Rating Scale--An Instrument for Measuring Art Integration

Numerous art therapists, estheticians, and art educators have examined rating methods for art productions. This paper presents a new rating scale designed to correlate artistic integration with personality organization. The scale was effective in discriminating artwork by 67 hospitalized from 63 nonhospitalized subjects. Using 67 portfolios of spontaneous artwork by patients with a personality disorder, raters found significant differences in the scores of the six subjects who later suicided.

Despite the interest and effort of researchers, instruments measuring the characteristics of visual art of adult populations have not demonstrated acceptable levels of validity and reliability (Barron, Gaines, Lee, & Marlowe, 1973; Goldman & Velasco, 1980; Goude, 1972; Goude & Derefeldt, 1981; Lewinsohn, 1965; Reiner, Telling, & O'Reilly, 1977; Wadlington & McWhinnie, 1973; Wright & Mcintyre, 1982). Barron et al. (1973) were unable to comment about components of artistic style and artists' personalities because of problems with their rating scheme. Wadlington and McWhinnie (1973) designed an 18-variable rating scale based on formal criteria defined by esthetician Monroe Beardsley (1958). They had only limited success with distinguishing differences between esthetic qualities in the artwork of different diagnostic groups of mental patients. Reiner et al. (1977) developed the Picture Regression Scale for Adults (PRSA) to rate regressiveness in artistic expression.

The literature contains mixed reports and some disappointment in quantifying the relationship between artwork and individual emotions. Wright and McIntyre's 1982 standardized Family Drawing Depression Scale (FDDS) discriminated hospitalized depressed subjects from normal subjects (no history of psychiatric disorder), but did not significantly distinguish between inpatient psychiatric groups. Lewinsohn and May's 1963 attempts to discriminate between the figure drawings of depressed and nondepressed patients on the basis of unpleasantness in facial expression and body posture showed no significant differences. Likewise, they found no significant differences between the drawings of patients with schizophrenia and without schizophrenia by the criterion of "rejection of external stimulation." Although the researchers suggested a correlation between emotional expression in the figure drawings of the patient groups and personality, there was no experimental validation of this relationship.

However, there have been some successes. For example, raters with varying degrees of experience with psychiatric patients can distinguish artwork made by psychiatric inpatients from that of subjects with no psychiatric history. In a study by Levy and Ulman (1967), three groups of raters were accurate 90% of the time when identifying the actual status of the 96 subjects.

Some studies have suggested that preference for different kinds of artwork may correlate with personality type. Knapp (1964) identified three classical conceptions of the esthetic that manifest different sources of conscious imagery: a) the "Apollonaire" or representational, which deals with the commonplace or objective world; b) the "Pythagorean" or geometric, which coincides with the logic of mathematics; and c) the "Dionysian" or expressionistic, which has at its base the unconscious or preconscious state of primitive nature. Knapp found a strong correlation between extroverted (practical and worldly) personalities and a preference for realistic art. Introverted (subjective, sensitive, impractical) personalities favored expressionistic art. The personality types who gravitated toward geometric paintings were those marked by intellectual control and strong defenses against affect and impulse.

Bergland's 1982 pilot study of 14 institutionalized older adults' schematically drawn lifelines, using ratings based on artistic criteria, showed more organization of shapes and less depression in artwork after the subjects' participation in an art therapy group. Better organized artwork corresponded with better functioning, less depressed participants. Wright and McIntyre's 1982 study showed that drawings of clinically depressed individuals revealed significantly less depression at the end of hospitalization than at the beginning when correlated with Zung's Self-Rating Depression Scale (1965). While using artwork as dreamwork in her analytically oriented art psychotherapy, Margaret Naumburg (1950) commented on increased artistic integration as an indicator of higher ego defensive resolution of intrapsychic conflict. Kramer (1971) and Rubin (1984) have made similar assertions.

Wadeson's observations (1980) describe pictorial manifestations of severe psychiatric disorders. For example, she listed the following characteristics of artwork of persons with schizophrenia: fragmentation, unmixed colors, unusual coloring, deterioration of composition, disturbed spatial organization, and disintegration. In the artwork of patients with depression, she described evidence of low energy, bareness, constriction, and lack of detail, color, and effort. In the artwork of patients with mania, she noted "wild colors," lack of color variety, deterioration in composition, and carelessness.

Shoemaker (1978) constructed a "frame of visually observable dimensions within which specific definitions of psychopathology may fall" (p. 156). In 1981, the Shoemaker-Schulz-Doll-Bergland Scale to Measure Visual Content in Art was developed. Correspondences between visual constructs and stages of development were outlined in Shoemaker's 1982 paper.

That chaotic artwork reflects a chaotic emotional condition is a long-held tenet of art therapy. However, none of the scales developed thus far has received wide usage or acceptance. Can raters distinguish artwork made by psychiatric inpatients from that of subjects with no history of psychiatric hospitalization? How quantifiable is the notion that integration in artwork reflects the internal organization of the artist?

In this study we undertook to determine whether the Sheppard Pratt Art Rating (SPAR) Scale could be a valid and reliable instrument for systematically quantifying art productions, especially with clinical populations.

Experiment 1
Method
Content Validity (a). The (SPAR) Scale was adapted from the Shoemaker-Schulz-Doll-Bergland Scale to determine the degree of organization of spontaneously produced artwork. The SPAR Scale was designed to assess artwork within six categories: Space, Figures, Energy, Color, Composition, and General. Each category was divided into seven progressive developmental levels of artistic creation, organization, cognitive development, and emotional maturity (1 = low, 7 = high), from expressionistic and primitive (Level 1), to geometric and controlled (Level 4), to the accurate representation of nature (Level 7). Table 1 is an early version of the scale:

Thirteen art therapists across the country were surveyed concerning the SPAR Scale's accuracy and usefulness. All participants had a master's or doctoral degree in art therapy and were professional art therapists. These therapists were asked to evaluate each category of the scale according to four criteria:

To what extent does the scale seem to describe the category?
To what extent do the steps seem to be in the right order?
To what extent do the steps seem to be equal intervals?
To what extent would you feel comfortable using such a scale?
Participants rated each category using a 7-point scale: 1 = not at all, 2 = very little, 3 = little, 4 = moderately well, 5 = well, 6 = very well, and 7 = perfectly. Participants also were invited to recommend changes to the items within each category.

Content Validity (b). To test the content validity of the revised instrument, 11 local art therapists were selected (using the same criteria as with the first group) and asked to provide their ratings (using the same scale as the first group used) of the revised categories.

Results
Content Validity (a). Seventy-eight percent of all responses were 5 or above; only seven percent were less than 5 (Table 2). Since the participants felt that the intervals were equal, Pearson Product-Moment Correlation coefficients were used in the majority of the following analyses.

Table 2 indicates the categories that elicited the rank order of interrater agreement (based on the standard error) across all scales. These were the General, Color, Space, Figures, Affect and Energy, and Form categories. The average percent "well" or better ratings for these attributes were 96, 86, 82, 75, 59, and 69, respectively. Although the data indicated that the validity raters saw the scales as feasible and relevant, further scale modifications were made based on the participants' recommendations. Particular attention was paid to rewording the Affect and Energy and Form categories, which had the least consensus of usefulness. These then became the Energy and Composition categories in the final scale used by the reliability raters. The resulting revised SPAR Scale can be seen in Table 3.

Content Validity (b): Results are presented in Table 4. The standard errors were similar to those noted on the first evaluation with the same criteria for selection.

Discussion
The results of both content validity studies indicate a very high degree of art therapists' agreement concerning the various attributes of the scale. The raters of the revised instrument noted that the steps were in the right order, that the steps had equal intervals, and that the scale was satisfactory to use.

Experiment 2
Method
Interrater Reliability Assessment(a). Two MA-level art therapists used the revised SPAR Scale to rate art samples selected from long-term psychiatric inpatients. The definitions of the categories used by the raters can be see in Table 5. Each rater received 2 1/2 hours of training before using the scale with scoring samples that represented a variety of patients with different diagnoses. The raters then rated a selection of spontaneous two-dimensional artwork of people with widely varied diagnoses.

Interrater Reliability Assessment (b). Additional interrater reliability analyses were performed with four independent raters using drawings from subjects who were not hospitalized. These subjects were students in an art class at a local community college. This assessment was designed to attempt to replicate the findings of the Interrater Reliability Assessment (a) with a larger group.

Results
Interrater Reliability Assessment (a). Table 6 presents the results of this first interrater analysis. Bartko's Intraclass Reliability Coefficient (1966) was used to analyze the results. The "All" column is the overall reliability for raters. The "Criterion" column represents the reliability between two staff art therapists, while the "Study" column represents the reliability between two hired raters. The final column compares the mean ratings of the staff raters to the mean rating of the hired raters. All reliabilities were significant at the .001 level. A high degree of agreement was shown between raters on all dimensions. Instruction was provided when a new rater was hired. Using the same technique, interrater reliability was again tested. Results are shown in Table 7. Again, there was significant agreement among the raters.

Interrater Reliability Assessment (b). Pearson Product Moment Correlations were computed on the Space, Figures, Energy, Composition, Color, and General scales for four independent raters. The correlations ranged from a low of .421 to a high of .991. The majority of the correlations were above .80. All of the correlations were statistically significant at the .05 level.

The Interrater Reliability Assessment (b) generally replicated the findings of the Interrater Reliability Assessment (a). Test scoring was reliable across four raters in three different studies.

Experiment 3
Method
Sixty-three people served as the norm-ing group for the various scales. None of these individuals were hospitalized patients. Artworks were produced spontaneously (no theme given) and a wide selection of art materials was used. Two raters scored these subjects' protocols which were computed on a single drawing. These raters' average scores were subsequently transformed into percentiles for the six SPAR Scale subscales.

Results
To read the norming table (Table 8), the user simply scores the drawing, then descends the far right-hand column under the appropriate variable (e.g., Space, Figures). This converts a raw score to a norm score in order to see where a picture might fall in relation to a normed group. The user will note that the norm ratings are sometimes incremented in units less than one. This is because norms were developed from averages of two raters who would sometimes disagree slightly. However, this norming allows the user to assign ratings that are fractions, if desired. Finding the rating in the norming table that most closely matches, the user will then read off the corresponding cumulative percentage score. This is the norm score. The user will then repeat this comparison process for every variable. As an end product, the user should have six norm scores, one for each raw score for each of the variables.

The norm score refers to the percentage of people in the norming sample whose score on that scale fell at or below the score in question. For example, assume that a person received a raw score rating of 5 on the Space variable, 4 on Figures, 5 on Energy, 4 on Color, 3 on Composition, and 4 on General. When converted to percentiles, the raw scores become: Space = 95, Figures = 92.1, Energy =90.5, Color = 50.8, Composition = 15.9, and General = 73. This means that relative to the norming population, this hypothetical person's ratings were substantially higher than the sample.

Experiment 4
Method
To determine the discriminant validity of the instrument, the drawings of 67 patients with personality disorders were used. These drawings were produced spontaneously (no theme given). Patients had a wide selection of art materials available.

Criterion-Related Validity. The drawings were rated on 99 sequential occasions. That is, the patients provided as many as 99 drawings (portfolios of patient art in chronological sequence) and their average for the six SPAR Scale variables was computed across the entire sequence. In many cases, the patients did not provide a total of 99 drawings; however,there were at least 20 from each patient. The average rating for each of the six dimensions and the score on only the first rating was also converted to a norm score according to the procedures outlined above.

Patient Suicides. Our sample contained six patients who committed suicide 1 to 3 years post data collection. To determine if their drawing scores were significantly different from those of other hospitalized patients, independent tests were performed on the two groups for the Space, Figures, Energy, Color, Composition, and General variables.

Patients Cross-classified by Diagnosis. The patients' scaled scores on the above variables were then cross-classified by DSM-III-R (Axis I) diagnosis. Specifically, patients with a diagnosis of Dysthymia (300.40) were coded with a 1. Patients with diagnoses of Major Depression (296.2x or 3x) received a code of 2. All other patients were coded 3. These codes were then cross-tabulated with the same reduced measures outlined above.

Results
Criterion-Related Validity. The results of this analysis (Table 9) were consistent for Space, Color, Composition, Energy, and General variables. There were generally no changes across measures over time. Autocorrelational analyses (lag 110) did not reveal any significant trends. The patient population, however, did score noticeably lower on the scales relative to the norming sample, with reference only to the first test measurement (first picture). Eighty-one percent of the patients scored lower than the norming sample median on the Space variable; 91%, 97%, 98%, and 98% scored lower for the Color, Energy; Composition, and General variables, respectively.

The Figures variable was unique. The results changed markedly depending on whether or not the patient was measured once or over an extended period of time. With extended drawings, 98% of the patients scored less than the median of the norming sample on this dimension. When only the first drawing was scored, however, 55% scored less than the norming sample median. This result indicated that the Figures variable was the least reliable of all the subscales. With only one measure, hospitalized patients drew slightly more figures than did nonhospitalized (55%). With extended measurement, this percentage increased markedly (98%). Seventy-five percent of patients with Axis I diagnosis of depression scored I on the Figures subscale, whereas 94% of patients with other diagnoses scored 2 or more on the same subscale (X2=39.76, [p]<.000). The depressed patients' artwork was strikingly devoid of Figures.

Patient Suicides. Two of these variables yielded significant differences between the suicide groups (N = 6) and the remaining patients (N = 61). For the General variable, the mean of the suicide group (2.66) was significantly less than the mean of the remaining patients (3.43), [t](65)=2.52, [p], <.01. Likewise, for the Space variable, the mean of those who suicided (2.33) was significantly less than the mean of the other patients (3.20), [t](65)=2.14, [p], <.05.

The cases were cross-classified by group (those who suicided versus those who didn't) and by their scores on the various scale measures. To simplify the analysis, each patient's scores were dichotomized. If the score was less than 3, the patient received a I on a scale; this indicated prominent disorganization. The patient received a 2 if the scale score was greater than or equal to 3, indicating a higher level of organization.

When the dichotomized scaled scores were cross-classified by the suicide variable, the results indicated that the Composition and Space variables were significantly related in the predicted direction. There were significantly more suicide patients with lower scores on the composition scale (Chi-Square=4.804, [p]<.02). The same effect was evident for the Space variable. Sixty-seven percent of the suicide patients scored less than 3 whereas 79% of the other patients scored 3 or higher (Chi-Square=5.935, [p]<.015). The General variable approached significance (Chi-Square=3.689, [p]=.055) and the trend indicated more of the suicidal patients (33%) scored less than 3 on the scale versus only 8% of the remaining population.

Media Effect. Although patients had the option of selecting a wide range of art, materials, there was not a discernible media effect.

Patient Cross-classified Diagnosis. The analysis did not yield any significant findings. In general, for patients who varyin their level of depression, their drawings do not reflect significantly more or less impairment relative to other patient diagnoses.

General Discussion
There were several major findings in this study. First, psychiatric inpatients' artwork can be distinguished from a normal population's artwork (cf. Levy and Ulman, 1967; Wright and Mcintyre, 1982). The nonhospitalized subjects were also more likely to use a realistic format in their two-dimensional artwork than was the hospitalized group. Relative to the normal population, hospitalized patients scored on the six variables as follows: less than 3.5 on Space, less than 4 on Energy, less than 4 on Composition, less than 4 on Color, and less than 4 on General. The inpatient group's reliance on an abstract or semi-abstract format might speak to more introverted, primitive personality types than the more intellectually defended personality types of the normal sample (Knapp, 1964). The inpatient group's artwork was also less organized and more expressive in nature. No media effect could be found in the results of Experiment 3.

Second, the artwork of patients who committed suicide could be distinguished from a general hospitalized population. The mean of the suicide population was less than the mean of the other patients for all variables. It is likely that suicidal patients' drawings are disorganized across the board. Therefore art productions of patients who scored low on the SPAR Scale's General, Space, and Composition variables may serve as a "red flag" to suggest a need for increased precautions. It is important to point out that these tests were performed on the first drawings, suggesting that the SPAR Scale is capable of discriminating this critical population with as little as one administration.

In terms of the suicide group's artistic format (realistic, abstract, or mixed) there were some interesting trends in the data. Wadeson (1975) found that when patients were more depressed, less color was used in artwork. Pictures became emptier, less complete, and less effort was invested. She also noted that these characteristics reflected hopelessness and, in some cases, psychomotor retardation. In our study, the patients who suicided made abstract art with a frequency that approached statistical significance more than twice as often as the other hospitalized patients. The other patients were almost three times more likely than those who suicided to produce realistic art. In fact, this frequency of an abstract format for those who suicided is statistically significant when measured against the normal population's frequency of a realistic format.

Third, it does appear that artistic style and an artist's personality are related as Barron et al. (1973) stated. Our research suggests that low self-esteem effects the ability to organize pictorially, as Bradow (1975) found in his work with children. The artwork of the subjects with personality disorder was more expressive, less geometric/structured, less organized, and less representational than a group of normals. The suicide group's artwork was more chaotic, diffuse, cluttered, and random relative to that of the other patients. This is contrary to the fixed and rigid drawing style Honig (1975) noted in the artwork of suicidal patients. The discrepancy might suggest a significant difference between the artwork of patients who show some suicidal ideation and the artwork of patients in the most immediate danger.

The research suggests that intensely disorganized' spatial and compositional elements in an abstract format relate to functional ability (nonhospitalized subjects versus hospitalized patients) and in extreme cases the inability to function at all (those who suicided). In other words, the SPAR Scale does rate regressiveness in artistic expression like the Picture Regression Scale for Adults (PRSA) (Reiner et al., 1977). It can distinguish levels of emotional integration between different populations, although, for patients who vary in their level of depression, their drawings do not reflect significantly more or less impairment relative to other patient diagnosis.

Fourth, the fact that patients who were depressed did not show significant differences in their artwork's organization relative to other patients' work is consistent with Wadeson (1971). Both studies conclude that artistic organization is notsignificantly diminished by depression. Major Depression and Dysthymia were the only Axis I diagnoses compared. These are similar anyway. A future project might include scoring artwork from more different diagnoses such as patients with schizophrenia and comparing the results to the group with depressive disorders already reviewed.

A curious finding of the study was the decreasing frequency of appearance of figures over time for the hospitalized group. All of the patients were assigned to long-term units in the hospital. Perhaps the presence of figures represented a more counter dependent, less abstract and less intrapsychically invested posture, which is different from the symbol-making that deep involvement in psychotherapy fosters during long-term hospitalization. In fact, the nonhospitalized group and the hospitalized group were very similar in their frequency of figures when the first pictures of both groups were compared. The first pictures for the hospitalized group were done very near the beginning of hospitalization.

Conclusion
The SPAR Scale significantly discriminated hospitalized patients from a non-hospitalized sample with one picture. It also statistically discriminated on first application patients who had suicided from the hospitalized group. Similar scores in the Space, Composition, and General variables by other patients might be used as warnings of possible suicide attempts.

Designed as a measure of pictorial integration, the seven levels of the SPAR Scale parallel levels of emotional and cognitive development as they are manifested in artwork. The SPAR Scale provides a rating instrument applicable to nonrepresentational art as well as representational art. Graduated steps indicate increasing levels of complexity, focus, and sophistication as well as the flexibility in the type and amount of controls utilized by the artist.

Previous investigators (.e.g., Wadlington & McWhinnie, 1973) tried unsuccessfully to create a workable, esthetically oriented scale. The SPAR Scale appears to be more sensitive than the PRSA because the SPAR Scale distinguishes degrees of regression in artwork.

The authors have found the SPAR Scale a useful framework for evaluating and discussing patient artwork in ongoing treatment. For clinical use, the SPAR Scale offers several advantages: a) unlike directive drawing tests, the SPAR Scale may be used in a wider range of contexts because it is applicable regardless of the content of the artwork; b) raters only needed 2 1/2 hours of training to gain competence in using it reliably; c) because it can be used with spontaneously produced art, it can quantify a patient's changes over time; and d) norming data are available, albeit for a limited number.

Future studies may provide further proof of its discriminant validity. This study's results significantly advance the notion that the quality of pictorial organization reflects the artist's internal organization.

Requests for reprints should be sent to Christy Bergland at 4204 Wickford Rd., Baltimore, MD 21210.

* A research assistance grant was awarded by the American Art Therapy Association in 1985. The Research Department and the Rehabilitation Services Department at the Sheppard and Enoch Pratt Hospital have also supported the study with grants. Acknowledgment is also given to Clarence Schulz, M.D., Gerald Whitmarsh, PhD, Norm Ringel, MA, Rick Parente, PhD and Diane Gibson, MS, OTR/L.

Table I
Rater Initials ----- Patient Name -----
Date Rated ----- Date of Artwork -----
Medium -----
Abstract ----- Realistic -----
Mixed -----

Legend for Chart:

A - SPACE
B - FIGURES
C - AFFECT AND ENERGY
D - COLOR
E - FORM
F - GENERAL

A B

C

D

E

F

- intuitive awareness of - graceful human or animal
space - expressive features, shown
- endless space or depth both by detail and gesture
can be implied without - figures well-interactive
loss of personal space with surrounding space
- elements on the surface 7
appear to have
relationships in depth - contains and integrates a
as well as in length variety of human states
and breadth - acceptance and expression of
7 all affect with subtly
- implication of movement
- empathetic, integration of
thought and feeling
7

- subtle and sensitive use of
color for expressiveness
- color as light, not pigment
7

- articulation of all the
pictorial space
- whole form chosen or implied
with derail for pleasing
effect
- detail well-integrated into
total composition
- natural forms appear
life-like or animated
7

- true integration
- picture invites
interchangeable attention
between parts and whole
- harmonious, appealing,
effective
- synergistic: the whole is
greater than the sum of its
parts
7

- clear 3-dimensional space - whole or well-detailed
- space used expressively representational figures
- interchanging relation- - figures in good reciprocal
ships in space relationship to surrounding
- directionality is apparent space
6 - figures in motion and
interactive socially
6

- reflects integration of
conflictual aspects of
emotions/self
- control in the service
of expressiveness
- not static
6

- awareness of one color in
relation to another for
enhancement of
expressiveness
6

- color choice creates an
emotional atmosphere
- harmonizing colors
6

- visually interesting
figure-ground interplay
- forms defined in almost
a natural or surreal style
- forms well-articulated
visually
6

- almost integration
- refinement of parts and
whole
- integration of conflictual
elements
- boundaries flexible
6

- relative size and distance - appearance of 3rd dimension
explicitly defined by - some shading, shadows
placement within contour of figure
- flexible groundline with - spatial relationships may
contours suggest interaction socially
- objects in front or behind - realistic or accurate
without spatial conflict proportions
- 3-dimensional forms suggested 5
by shading
5 - evidence of motor control
- acceptance of some
conflicting emotion
- relaxation of
disassociation, repressions
or suppressions of feeling
- occasional painting out or
over
5

- broad use of color, with
some shading and blending
- increased use of paired,
even conflictual color
5

- forms arranged
differentially in space
- beginning of figure-ground
interplay
5

- refinement of parts in
relation to whole
- some flexibility, fluidity
- allowance of limited
conflictual material
5

- appearance of groundline, - stylized figures
usually straight and rigid - 2-dimensional forms/ rigid
- relative size, overlapping boundaries
images, and coveraging - figures may all face front
parallels may be used to or side
imply depth of space - may group figures
4 4

- 'pretty' pictures
- rigidly controlled affect
- sharp expression of
dichotomies/oppositionals
pain, as in stereotypic
4

- oppositional complements,
often garish but contained
- color fiat but filling
forms to boundaries
- appearance of natural
color in reference to form
- high-intensity colors plus
brown and black
4

- sharply defined boundaries
- rigidly defined geometric
forms
- rigid symmetry may
appear
4

- awareness of relationship
of parts to whole
- whole rigidly defined
- exclusion or containment of
conflictual elements
- emphasis on boundaries
4

- appearance of fragments, or - stick figures, or parts
parts of an implied or of figures
suggested whole - little or no relationship to
- use of bottom d page as other figures in space
implied space organizer 3
- 2-dimensional or flat space
3 - appearance of inconsistency
in stroke or line quality
that does not explicate or
enhance the visual
understanding of the content
- picture space (ground)
unformed
3

- patches of color
- the suggestion of natural
color, maybe unnatural but
not bizarre use
- unmixed or unmodified
color (directly from the
jar)
3

- shapes are distinguishable
in part
- transparencies
- some diffuse boundary
- sketchy, unsure forms
3

- appearance of
recognizable parts or
fragments, implying a
wholeness
- awareness of scattered
elements not integrated,
except in limited focus
3

- clustering in space - early formed centers
- patterning or repetitive - not true figures
strokes - suggestions of figures
- creation of a part, or in scribbles, perhaps
fragment, floating in space 2
2
- appearance of primitive
emotion
- possible impulsivity or
violent outbursts shown in
the use of the medium
2

- 1 or 2 colors in
relationships
- no natural reference
- clustering of randomly
ordered color
2

- clustering of strokes,
patterning that may suggest
formed shapes
- lack of boundaries,
boundaries only implied
2

- minimum sustainment of
attention
- beginning of clustered
form(s)
- particle or partial
awareness
- separate unrelated content
2

- chaotic spatial sense - kinesthetic experimentation
- experimentation with art - early sensation, no body
materials in random diffused image awareness
way, all over media space - absence of figures
1 1

- chaotic, no differentiation
of feelings
- flat or absent affect
- randomness
1

- single color unrelated to
content
- color used only as material
or physical substance
- some physical stimulation
by sensation of color
1

- diffusion
- formlessness
- chaotic form
1

- chaotic
- experimental sensation
- no sustained attention
1
Sheppard-Pratt Art Rating (S.P.A.R.) scale adapted from the Shoemaker-Schulz-Doll-Bergland Scale to measure visual content seen in art work. 732-art.892

TABLE 2 First Validity Study Descriptive Statistics of Art Therapy Scale Ratings.
Attribute Scale N Mean SD SE % well
1

Affect 1 12 5.00 1.13 .33 67
2 11 5.18 1.08 .33 64
3 11 5.00 1.00 .30 55
4 12 4.42 1.56 .45 50

Space 1 12 5.58 0.90 .26 92
2 12 5.67 0.78 .23 92
3 12 4.75 1.29 .37 67
4 12 5.08 1.00 .29 75

Color 1 13 5.62 0.96 .27 85
2 11 5.64 0.81 .24 91
3 12 5.42 0.79 .25 83
4 12 5.25 0.97 .28 83

Form 1 13 5.08 1.32 .37 69
2 13 5.31 1.25 .35 85
3 13 4.92 1.19 .33 69
4 13 4.46 1.61 .45 54

Figures 1 12 5.50 1.00 .29 75
2 12 5.25 1.29 .37 75
3 12 5.25 1.06 .31 75
4 12 4.83 1.40 .40 75

General 1 13 5.69 0.48 .13 100
2 13 5.77 0.44 .12 100
3 13 5.62 0.51 .14 100
4 13 5.38 0.77 .21 85
1 Percent rating "well," "very well," or "perfectly."

TABLE 3
PATIENT NAME -----
RATER INITIALS ----- DATE OF ARTWORK -----
DATE RATED ----- MEDIUM -----
ABSTRACT --- REALISTIC --- MIXED ---

Legend for Chart:

A - SPACE
B - FIGURES
C - ENERGY
D - COLOR
E - COMPOSITION
F - GENERAL

A
B
C
D
E
F

- intuitive awareness of space
- endless space or depth can be implied without loss of personal
space
- elements on the surface appear to have relationships in depth
as well as in length and breadth. 7
- graceful human or animal figures

- expressive features, shown both by detail and gesture
- figures well-interactive with surrounding space 7
- believable movement
- directionality and intention throughout the whole
picture 7

- subtle and sensitive use of color for expressiveness
- color as light, not pigment 7

- articulation of all the pictorial space
- whole form chosen or implied with detail for pleasing
effect
- detail well-integrated into total composition
- natural forms appear life-like or animated 7

- true integration
- picture invites interchangeable attention between parts
and whole
- harmonious, appealing, effective
- synergistic: the whole is greater than the sum of its
parts 7

- clear 3-dimensional space
- space used expressively
- interchanging relationships in space
- directionality is apparent 6

- whole or well-detailed representational figures
- figures in good reciprocal relationship to surrounding
space
- figures in motion and interactive socially 6

- control in the service of expressiveness
- not static
- stroke or line quality explicates or enhances the visual
understanding of the content 6

- awareness of one color in relation to another for
enhancement of expressiveness
- color choice creates a balanced emotional atmosphere
- harmonizing colors 6

- visually interesting figure-ground interplay
- forms defined in almost a natural or surreal style
- forms well-articulated visually 6

- almost integration
- refinement of parts and whole
- integration of conflictual elements
- boundaries flexible 6

- relative size and distance explicitly defined by placement
- flexible groundline with contours
- objects in front or behind without spatial conflict
- 3-dimensional forms suggested by shading 5

- appearance of 3rd dimension
- some shading, shadows within contour of figure
- spatial relationships may suggest interaction socially
- realistic or accurate proportions 5

- evidence of motor control
- relaxation of disassociation, repressions or
suppressions of feeling
- occasional painting out or over 5

- broad use of color, with some shading and blending
- increased use of paired, or conflictual color is
modulated
- if single color, gradations are evident 5

- forms arranged differentially in space
- beginning of figure-ground interplay 5

- refinements of parts in relation to whole
- some flexibility, fluidity
- allowance of limited conflictual material 5

- appearance of groundline, usually straight and rigid
- relative size, overlapping images, and coverging parallels may
be used to imply depth of space 4

- stylized figures
- 2-dimensional forms/rigid boundaries
- figures may all face front or side
- may group figures 4

- "pretty" pictures
- rigidly controlled affect
- sharp expression of dichotomies/oppositionals
- sustained attention 4

- oppositional complements, often garish but contained
- color flat but filling forms to boundaries
- appearance of natural color in reference to form
- high-intensity colors plus brown and black 4

- sharply defined boundaries
- rigidly defined geometric forms
- rigid symmetry may appear 4

- awareness of relationship of parts to whole
- whole rigidly defined
- exclusion or containment of conflictual elements
- emphasis on boundaries 4

- appearance of fragments, or parts of an implied or suggested
whole
- use of bottom of page as implied space organizer
- 2-dimensional or flat space 3

- stick figures, or parts of figures
- little or no relationship to other figures in space 3

- appearance of inconsistency in stroke or line quality
that does not explicate or enhance the visual
understanding of the content
- floating or ungrounded 3

- patches of color
- the suggestion of natural color, maybe unnatural but not
bizarre use
- unmixed or unmodified color (directly from the jar or
tube) 3

- shapes are distinguishable in part
- transparencies
- some diffuse boundary
- sketchy, unsure forms 3

- appearance of recognizable parts or fragments, implying
a wholeness
- awareness of scattered elements not integrated, except
in limited focus

- clustering in space
- patterning or repetitive strokes
- creation of a part, or fragment, floating in space 2

- early formed centers
- not true figures
- suggestions of figures in scribbles, perhaps 2

- appearance of primitive motion
- possible impulsivity or violent outbursts shown
in the use of medium
- picture space unformed or unorganized 2

- no natural reference, clustering of randomly ordered
color
- clustering of color
- randomly ordered color 2

- clustering of strokes, patterning that may suggest
formed shapes
- lack of boundaries, boundaries only implied 2

- minimum sustainment of attention
- beginning of clustered form(s)
- particle or partial awareness
- separate unrelated content 2

- chaotic spatial sense
- experimentation with art materials in random diffused way, all
over media space 1

- kinesthetic experimentation
- early sensation, no body image awareness
- absence of figures 1

- chaotic
- randomness through-out
- lack of directionality and intention 1

- single color unrelated to content
- color used only as material or physical substance
- some physical stimulation by sensation of color 1

- diffusion
- formlessness
- chaotic form 1

- chaotic
- experimental sensation
- no sustained attention
Sheppard-Pratt Art Rating (S.P.A.R.) Scale adapted from the Shoemaker-Schulz-Doll-Bergland Scale to Measure Visual Content Seen in Art Work

TABLE 4 Second Validity Study Descriptive Statistics of Art Therapy Scale Ratings.
Legend for Chart:

A - Attribute[1]
B - Scale
C - N
D - Mean
E - SD
F - SE
G - % well

A B C D E F G

Affect 1 11 5.45 0.66 .20 91
2 11 5.82 0.57 .17 100
3 11 5.55 0.89 .27 82
4 11 5.27 0.86 .26 73

Space 1 11 5.91 0.29 .09 100
2 11 6.09 0.51 .15 100
3 11 5.45 1.16 .35 82
4 11 5.36 0.88 .27 82

Color 1 10 5.70 0.78 .25 90
2 10 5.80 0.87 .28 90
10 5.60 1.02 .32 80
4 10 5.60 0.29 .29 80

Form 1 11 5.64 0.64 .19 91
2 11 6.00 0.60 .18 100
3 11 5.45 1.30 .39 82
4 11 5.55 0.78 .24 91

Figures 1 11 6.00 0.60 .18 100
2 11 6.09 0.67 .20 100
3 11 5.91 0.51 .15 91
4 11 5.82 0.83 .25 91

General 1 11 5.82 0.57 .17 100
2 11 5.91 0.67 .20 100
3 11 5.91 0.67 .20 100
4 11 5.64 0.77 .23 91
TABLE 5 List of Definitions used for Training Raters (Rating Scale Worksheet)
SPACE*
The illusion of space has the quality of endlessness found in the natural environment. The picture frame has the quality of a window through which one can see the endless recession of forms in space. There is a sense that natural, physical laws are respected.

The visual elements on the surface of the picture plane are made to give the illusion of having relationships in depth as well as in length and breadth.

FIGURES
The full development of animal or human figures as natural, realistic entities in space.

ENERGY
Energy is described as encompassing multiple affective states in a harmonious rhythmical presentation of form. This gives evidence of a continuity or sustainment of attention.

There is a synergistic relationship between thought and feeling. Energy modulation and control of motor involvement are evident. Movement may be seen in the form as well as the content of artwork.

COLOR
Esthetically pleasing arrangements of descriptive hues and values combined so that a sense of unity is achieved.

COMPOSITION*
The organization or arrangement of all of the visual elements according to principles that develop an organic unity in the total work of art.

GENERAL
The degree of organization that draws upon salient features of space, form, figures, affect and energy, and color.

1 Percent rating "well," "very well," or "perfectly."

TABLE 6 Interrater Reliability Coefficients[1] First Group of Raters
Legend for Chart:

A - All (1)
B - Raters: Criterion (2)
C - Raters: Study (3)
D - Mean (2) vs Mean (3)

A B C D

Space .90 .91 .89 .95
Figures .96 .96 .98 .97
Energy .91 .89 .95 .94
Color .89 .90 .87 .94
Composition .90 .86 .90 .97
General .92 .91 .91 .97
1 All "p" = .001

Statistical technique: Bartko's Intraclass Reliability Coefficient.

TABLE 7 Interrater Reliability Coefficients[1] Second Group of Raters
Space .92
Figures .97
Energy .89
Color .83
Composition .96
General .95
1 All "p"= .001

Statistical technique: Bartko's Intraclass Reliability Coefficient.

TABLE 8 Norm For Variables
Legend for Chart:

A - General: Value
B - General: Percentile
C - Space: Value
D - Space: Percentile
E - Figures: Value
F - Figures: Percentile

A B C D E F

2.0 4.8 1.5 1.6 1.0 78.0
2.5 6.3 2.0 7.9 2.5 81.0
3.0 14.3 2.5 14.3 3.0 84.0
3.5 33.3 3.0 34.9 3.5 87.3
4.0 73.0 3.5 46.0 4.0 92.1
4.5 84.1 4.0 68.3 4.5 98.4
5.0 96.8 4.5 87.3 5.0 100.0
5.5 100.0 5.0 95.2 -- --
-- -- 5.5 98.4 -- --
-- -- 6.0 100.0 -- --

Legend for Chart:

A - Energy: Value
B - Energy: Percentile
C - Color: Value
D - Color: Percentile
E - Composition: Value
F - Composition: Percentile

A B C D E F

3.0 11.1 2.0 4.8 1.5 1.6
3.5 20.6 2.5 6.3 2.0 3.2
4.0 61.9 3.0 7.9 3.0 15.9
4.5 71.4 3.5 19.8 3.5 31.7
5.0 90.5 4.0 50.8 4.0 60.3
5.5 100.0 4.5 73.0 4.5 77.8
-- -- 5.0 88.9 5.0 95.2
-- -- 5.5 98.4 5.5 100.0
-- -- 6.0 100.0 -- --
TABLE 9 Percentage of Hospitalized Patients' Scores That Exceeded Norming Sample Median On All Scales, After One, And More Than One Test
Legend for Chart:

A - Variable
B - % Less Than Median
C - % Greater than Median

A B C

1 Test > 1 Test 1 Test > 1 Test

Space 81% 96% 19% 4%
Color 91% 96% 9% 4%
Energy 97% 96% 3% 4%
Composition 98% 96% 2% 4%
General 98% 98% 2% 2%
Figures 55% 98% 45% 2%
REFERENCES
Barron, F., Gaines, R., Lee, D., & Marlowe, C. (1973). Problems
and pitfalls in the use of rating schemes to describe
visual art. Perceptual and Motor Skills, 37, 523-530.

Bartko, J.J. (1966). The Intraclass Correlation Coefficient as a
measure of reliability. Psychological Reports, 19, 3-11.

Beardsley, M. (1958). Aesthetics: Problems in the philosophy of
criticism. New York: Harcourt Brace.

Bergland, C. (1982). The life review process in geriatric art
therapy: A pilot study. The Arts in Psychotherapy, 9,
121-130.

Goldman, R.K., Velasco, M.M., Jr. (1980). Toward the development
of a rational scale in the use of human-figure drawings as
a kindergarten screening measure. Perceptual and Motor
Skills, 50, 571-577.

Goude, G. (1972): A multi-dimensional scaling approach to the
study of art. Scandinavian Journal of Psychology, 13,
258-284.

Goude, G., & Derefeldt, G. (1981). A study of Wolfflin's system
for characterizing art. Studies in Art Education, 22(3),
32-41.

Honig, S. (1975). Ideation in the artwork of suicidal patients.
Art Psychotherapy, 2, 77-85.

Knapp, R. (1964). An experimental study of a triadic hypothesis
concerning the sources of aesthetic imagery. Journal of
Projective Techniques and Personality Assessment, 28,
49-54.

Kramer, E. (1971). Art as Therapy with Children. New York:
Schocken Books.

Levy, B.I., & Ulman, E. (1967). Judging psychopathology from
paintings. Journal of Abnormal Psychology, 72(2), 182-187.

Lewinsohn, P.M. (1965). Psychological correlations of overall
quality of figure drawings. Journal of Consulting
Psychology, 29(6), 504-512.

Lewinsohn, P.M., May, J.G. (1963). A technique for the judgment
of emotion from figure drawings. Journal of Projective
Techniques and Personality Assessment, 27, 79-85.

Naumburg, M. (1950). Schizophrenic art: Its meaning in
psychotherapy. New York: Grune & Stratton.

Ocvirk, O.G., Bone, R.O., Stinson, R.E., & Wigg, P.R. (1973).
Art fundamentals: Theory and practice. Dubuque, Iowa:
William C. Brown.

Reiner, E.R., Telling, J.A., & O'Reilly, J.B. (1977). A picture
regression scale for adults. Art Psychotherapy, 4, 219-223.

Rubin, J.A. (1984). Child art therapy. New York: VanNostrand
Reinhold.

Shoemaker, R. (1978). The significance of the first picture in
art therapy. In B.K. Mandel, R. H. Shoemaker, & R.E. Hays
(Eds.), The proceedings of the Eight Annual Conference of
the American Art Therapy Association (pp. 156-162).
Baltimore: American Art Therapy Association.

Shoemaker, R. (1982). Assessment and record-keeping: Building
bridges that communicate progress. In A. Dimaria, E.
Kramer, & I. Rosner (Eds.), The proceedings of the Twelfth
Annual Conference of the American Art Therapy Association
(pp. 22-26). Falls Church, Virginia: American Art Therapy
Association.

Shoemaker, R., Schulz, C., Doll, A., & Bergland, C. (1981,
June). Scale to measure visual content seen in artwork.
Unpublished lecture. Pilot study was presented at
Scientific Day Meeting, Sheppard Pratt Hospital, Towson,
Maryland.

Wadeson, H. (1971). Characteristics of art expression in
depression. Journal of Nervous and Mental Disease, 153(3),
197-204.

Wadeson, H. (1975). Suicide: Expression in images. American
Journal of Art Therapy, 4(3), 75-82.

Wadeson, H. (1980). Art psychotherapy. New York: Wiley.

Wadlington, W.L., & McWhinnie, H. (1973). The development of a
rating scale for the study of formal aesthetic qualities in
the paintings of mental patients. Art Psychotherapy, 1,
201-220.

Wright, J.H., & McIntyre, M.P. (1982). The family drawing
depression scale. Journal of Clinical Psychology, 38(4),
853-861.

Zung, W. (1965). A self-rating depression scale. Archives of
General Psychiatry, 12, 63-70.

~~~~~~~~

By CHRISTY BERGLAND and ROSANNA MOORE GONZALEZ[*]

Christy Bergland, A.T.R. is a painter and art therapist in private practice. For the last 13 years she has been Supervising Arts Therapist and Chief of Expressive Arts Therapies at Sheppard Pratt Hospital in Baltimore. Rosanna Moore Gonzalez, A.T.R. worked for 13 years as an art therapist at St. Elizabeth's, Taylor Manor, and Sheppard Pratt hospitals in the Washington, DC and Baltimore areas. She is currently involved in helping her own three children discover the arts.

Share this with your friends