An Alternative Interpretation Hypothesis Regarding Color Projection Sanae Aoki & Nobuo Kogayu

        The appearance of Color Projection（CP） on the Rorschach test has been conventionally considered to signify the denial of unpleasant affection. However, this hypothesis has not been supported empirically due to lack of empirical evidence. This study investigated this conventional hypothesis quantitatively by examining verbal expressions of affection in explanations of CP responses. The participants included 37 adult Japanese psychiatric patients. The study was conducted between April 1995 and July 2017, and 68 CP responses produced by participants in the Rorschach protocol were examined.The results showed that 22% of the color projection (CP) responses were accompanied by pleasant affection only, as previously hypothesized. Furthermore, CP responses with unpleasant and negative affection or no affection at all were revealed. CP responses were classified into three by type of affective expression. While CP responses with only pleasant affective explanations were not accompanied by any special scores and often appeared in the form of multiple color projections in the gray area, CP responses, including negative affective explanations or no affective expressions at all, could be accompanied by special scores.The results suggest that while some CP responses may indicate a denial of unpleasant affection, as previously hypothesized, there may be CP responses that are reflecting cognitive integration deficits rather than affection.

Introduction:

        Color Projection（CP） in the Rorschach test occurs when a subject projects chromatic colors onto blot areas that display only varieties of black and gray tones (Piotrowski, 1957). Examples include ―a blue bird‖ in response to Card I, ―a bright red butterfly‖ in response to Card V, and ―green leaves‖ for Card VI. Piotrowski (1957) suggested that CP indicates a deliberate, conscious attempt to overtly feel, sense, and display happiness, while suppressing spontaneous and deeply felt sadness. He also proposed that CP shows an earnest, intense attempt at self-imposed serenity to dispel depression caused by deeply felt frustration.

        Moreover, Weiner (1998) and Exner (2003) later recognized that in broader terms, CP is the rejection of not only depressive feelings but also unpleasant feelings in general. Currently, many Rorschach systems support these hypotheses. Weiner (1998) reported that ―people (subjects) with one or more CP in their records are usually individuals who keep themselves in good spirits by using denial to ward off any potentially depressing feeling.‖ Therefore, according to Weiner (1998), ―Color Projection responses almost always involve pleasant percepts very rarely do subjects report a CP with unpleasant connotations.‖ He added that ―because it is so strained and contrived, the defensiveness indicated by CP is very fragile and likely to crumble quickly when challenged.‖ Exner (2003) wrote, ―The presence of even one CP represents the use of an unusual form of denial to contend with unpleasant affective experiences.‖ Weiner (1998) and Exner (2003) interpreted CP as the rejection of not only depressive affect but also unpleasant affect overall, but they essentially supported Z.A. Piotrowski’s hypothesis.

        Thus, in almost all scoring systems that use CP, the definition and interpretation of CP is identical, regardless of differences in classification frameworks such as special scores or determinants. The most widely accepted interpretation of CP is that the perception of chromatic colors in achromatic areas is an intentional, but unusual and strained, psychological mechanism of cognitive transformation to cope with unpleasant affections activated by black images. However, few studies provide grounds for this interpretation, and most were conducted by Exner. According to Exner (2003), a review of 430 records of outpatients revealed that CP was coded at least once in 44 of the protocols. Therapist evaluations of these patients indicated that the primary symptom patterns of those giving CP responses were psychosomatic (N = 14), hysteroid-like problems in dealing with affect (N = 20), depression (N = 7) and obsession-like features (N = 3). The therapist evaluations included a section pertaining to defense tactics, of which, if discernible, as many as three were to be listed. 95% of the 44 persons who gave at least one CP response were evaluated as having marked tendencies to use denial as a defensive tactic. Only about 32% of the remaining 385 patients were evaluated as frequently using denial as a defensive tactic.

        Some reports in Japan have supported the conventional hypothesis asserted by Z.A. Piotrowski and Exner, but others do not provide support. In 10 Rorschach case studies, Ishii (2003) found that most CP responses involved positive images, but some involved negative ones, such as a ―sea of blood‖ and ―dead leaves.‖ Aoki found that some CP responses involved pleasant perceptions but that others occurred with unpleasant ones, such as a ―drowned body‖ and ―rotten leaves‖ (Aoki, 2009, 2010, 2011).

        Aoki’ research also showed through several case studies that some CP responses are better interpreted as cognitive integration failure, such as color integration failure, rather than as denial of discomfort as previously hypothesized (Aoki, 2009, 2011, 2013). Aoki (2013) suggests that some CPs in which the projected color does not match the color of the content, for example, "dead leaves, because dead leaves are a fresh green color," or "violin, because violin is definitely pink," or CPs with significantly lower form quality, or CPs with scores indicating cognitive deviance, should not be immediately applied to Z.A.Piotrowski’s conventional hypothesis, but should be interpreted carefully. Furthermore, Aoki (2013) attempted to qualitatively classify CP responses in 30 Rorschach cases of CP according to content, verbal expression, shape and color conformity, and color obsession (absolute, red, etc.). As a result, four types of responses were identified: denial of unpleasant feelings as previously hypothesized, preoccupation with content (e.g., forcibly changing color due to preoccupation with content), flashback (including vivid descriptions), and integration failure (mismatch between color and response content, projecting a color different from that commonly recognized with the object; e.g., dead leaves are fresh green). The results suggest that CP may be divided into four types of CP responses. These results support Z.A. Piotrowski's claim that CP can be interpreted as indicating a denial of discomfort. However, these investigators also suggest that not all CPs carry this connotative meaning.

CPs are unique responses that change the actual color of the ink, and even protocols that show only one CP can have important implications, which is why appropriate interpretations may be important in understanding individuals. Therefore, if there are other possible interpretations of CP that are more promising than the traditional explanations, these should be considered.

        Recently, serious criticism has been brought to bear on the variables used in the Rorschach test. In response, the validity of the Rorschach variables has been defended (Meyer, Erdberg, & Shaffer, 2007; Mihura, Meyer, Dumitrascu, & Bombel, 2013; Tibon & Zeligman, 2016; Wood, Garb, Nezworski, Lilienfeld, & Duke, 2015; Wood, Nezworski, Grab, & Lilienfeld, 2001). Mihura, Bombel, and Dumitrascu (2013) evaluated peer-reviewed literature on Rorschach validity in terms of the 65 main variables in the popular Comprehensive System (CS), and CP was found to be among those with the least support. However, Mihura et al. (2013) report that this lack of support should not imply evidence of the lack of or low or unstable levels of validity but instead absence of evidence because no studies have been conducted on CP, perhaps due to its low incidence.

        In fact, basic research on color itself has been scarce in recent years; the results of C. Piotrowski’s (2017) exploratory bibliometric "topic" analysis focusing on primary Rorschach studies published between 2000 and 2016 showed that there are limited studies on color compared to the large number of Rorschach studies on psychiatric disorders. Furthermore, the low frequency of CP contributes to the lack of studies on CP.

The incidence of CP has been reported by Exner (2001), stating that it occurs in 3–7% of clinical patients. Conversely, the incidence of CP of the nonclinical group is lower than that of the clinical group and is 0–3% (Meyer et al., 2007; Nakamura, Fuchigami, & Tsugawa, 2007). However, Ishii (2003) and Aoki (2013)

Color Projection: 21

reported that CP might not be extremely rare in the empirical context. It may be difficult to collect data on CP for effective evaluation by statistical methods in short-term studies. Data on Rorschach protocols, including CP responses, are often accumulated in spontaneous, day-to-day work in clinical settings over the long term. A range of issues, including data control, can arise when data are accumulated in the course of long-term clinical practice. It is also time-consuming to collect enough to withstand validity studies. However, to verify the conventional understanding of CP, it is necessary to collect data and perform basic research.

        Aoki and Kogayu (2021) conducted a basic study of CP and examined which cards, locations, quality of development, quality of form, determinants, content, special scores, and projected colors accompanied the CP responses. The results showed that almost half of the CP responses contained special scores indicating deviations in cognition and thinking, half of which indicated a severe level (level 2). Considering that about 30% of the CP responses were FQ- and 10% of the CPs included color projections that did not match the content of the responses, it was also inferred that some of the CPs failed to integrate their cognition and thinking. Furthermore, given that nearly 30% of the CP responses were scored with MOR, it was thought that the traditional CP interpretation hypothesis of denial of unpleasant emotions may not apply to all CP responses. However, Aoki and Kogayu (2021) did not examine the linguistic representation of CP protocols or even classify CPs. Therefore, this study focused on Weiner's (1998) assertion that ―CP responses almost always involve pleasant percepts and only very rarely do subjects report a CP with unpleasant connotations‖. In particular, the presence or absence of affective expressions and their affective dimensions (pleasant, unpleasant, or both) were examined in an attempt to classify CP responses.

        The final goal of the current study was to assess the validity of CP and to re-examine the traditional interpretation of CP. Although validity cannot be determined in this study, if there are multiple types of CP responses, each with different meanings, then it is quite possible that multiple interpretation hypotheses beyond the traditional interpretation hypotheses are possible. Furthermore, if there are multiple types of CPs, knowing their attributes and grouping them accordingly may generate new interpretive hypotheses, which may lead to improved evaluation accuracy regarding clinical support. In the current study, we used the same dataset as reported by Aoki and Kogayu (2021).

Methods: Procedure:

        This study was performed with the cooperation and agreement of the patients and their treatment organizations. The Rorschach test was administered individually for diagnosis and planning in the treatment organizations. When CP appeared in Rorschach protocols, the record was placed in the set of research data after obtaining informed consent form the patient, explaining that the data were to be used only for research purposes and that personal information would be completely protected. Data from patients who signed a participation agreement were used. The research was conducted in from April 1995 and to July 2017. In all, 68 CP responses in 37 protocols were collected from over 1500 Rorschach protocols.The survey was conducted with the approval of the Committee on Ethics of the Faculty for Research in Human Sciences, Tsukuba University.

Data analysis:

Certified Psychologists with more than 20 years of experience in Rorschach testing, including the authors, scored protocols that were collected in the form described in the procedure independently by employing the CS (Exner, 2001). Before the scoring, the scorers agreed to use the CP principle, in which CP were coded only when the presence of chromatic coloring in the achromatic blot area was identified (Exner, 2003). In addition, prior to the score, it was confirmed among the scorers that achromatic included "white" and chromatic included both color specifications such as "red" and "blue" and non-color specifications such as

22: Aoki & Kogayu

"colorful" and "multicolored." 97.0% of all scores, including CP, were in agreement. Disagreements were discussed by adding another certified psychologist to determine the final scores.

Data:

        The research data included 68 CP responses from adult mental health patients (N = 37, 4 men and 33 women, age range 15 to 47 years, means age = 28.68 years,SD = 9.80). None of the patients had any visual impairments or problems with color recognition as a result of the attending doctor's diagnosis or medical checkups.The diagnosis of the patients varied from posttraumatic stress disorder, schizophrenia, depression, dissociation disorder, panic disorder, and adjustment disorder.

Results:

Expressions of affection in explanations of CP response:

The verbal expressions of 38 CP responses (55.9%) had affective words, but 30 CP responses did not (see Table 1); 15 responses (22.1%) included only pleasant and positive expressions (e.g., ―beautiful,‖ ―bright,‖

―brilliant,‖ ―cute,‖ and ―warm‖), and 12 responses (17.6%) included only unpleasant and negative affective expressions (e.g., ―bloody,‖ ―bad,‖ ―obnoxious,‖ ―malicious,‖ ―sad,‖ and ―dirty‖). Furthermore, 11 responses (16.2%) included both pleasant and unpleasant affective expressions and ambivalent ones, such as ―beautifully sorrowful butterfly‖ or ―bad, dirty, and cool.‖ Thus, the CPs involved responses with both pleasant and unpleasant affection, and not all CP responses involved pleasant percepts.

Table 1. Affective Expression of CP Response

Note. Group= the groups of verbal expressions of affection in explanations of CP response; N=number and percentage of each explanation type of CP; Affective explanation = Examples of affective explanation (adjectives) on CP; Fre=frequency of the Affective expressions.

Features of affection groups in explanations of CP response:

Color Projection: 23

        To examine the characteristics of CP for each of the four emotional expression types, the frequency of each type of CP was compared with those of the others. The following scores were assessed: cards (CP > 5, Cards I, IV, V, VI, or VII); the lightness of the card’s color (black card: I or V; light gray card: VI or VII); locations (W, D, or Dd); projected colors (one specific color, multiple specific colors, or no specific chromatic color[s]); appropriateness of projected colors (appropriate or inappropriate); determinants (shading or no shading/Form); special scores (Level 1, Level 2, or no special score); DR (Yes or No); INCOM; FABCOM; and/or CONTAM (YES or NO); and form qualities (FQo, FQu, or FQ-). CPs of different types of affective explanations (only pleasant affection, only unpleasant affection, both pleasant and unpleasant affection/ambivalent, and no affection) are shown in Table 2.

Table 2. The Percentages of Affective Type in CP Responses

Affective type

explanation; Pleasant and unpleasant＝CP with both pleasant and unpleasant explanation; No-affection＝CP without any affective explanation.

        CP responses with only unpleasant affective expressions were more likely (83.3%) to occur with black cards than they were with light gray ones (16.7%). CP responses with only pleasant affective expressions were more likely (76.9%) to occur with the light gray cards (cards VI, VII) than they were with the black ones (23.1%). W featured 70.6% of the CP responses, and CP responses with only unpleasant affective expressions were all seen in W (100%). No affective CP was seen less commonly in W (50%) but more commonly in Dd (26.7%) than any of the other three affective CP groups. Neither CP with only pleasant affection nor with only unpleasant affection was seen in Dd. Furthermore, only two CP responses with ambivalent affection were seen there (18.2%). CP responses with pleasant affective expressions were significantly more likely to feature multiple specific chromatic colors (46.7%) than the other CP types did (0–27.3%) and were less likely to be seen in only one specific color (40.0%) than other CP types were

Color Projection: 25

        (54.5–76.7%). Furthermore, CP responses with only unpleasant affective expressions were more likely to feature inappropriate colors (33.3%) than other affective CP types were (0–9.1%). CP responses with ambivalent affective expressions were more likely to be seen with a Level 2 special score (54.5%). No affective CP was observed less with a Level 2 special score (6.7%) and more with no special score (66.7%). However, CP with unpleasant affective expressions was less likely to be seen with no special score (25%). More CP responses with ambivalent affective expressions were scored DR (72.7%), and fewer no-affective CP responses were scored DR (20.0%). More CP responses with unpleasant affective expressions were scored INCOM, FABCOM, or/and CONTAM (75.0%), and fewer no-affective CP were scored INCOM, FABCOM, or CONTAM (33.3%). No bias of more than two-thirds (66.7%) was found for form qualities.

Classification of CP response:

        Quantification Method Type III (Hayashi, 1993), a type of correspondence analysis, was used to classify the CP responses according to the characteristics of the affection groups. The analysis covered eight categories and nineteen variables: card color, location, projected color, appropriateness of projected color, special scores, and emotional expressions included in response to INCOM/FABCOM/CONTAM/DR and CP characterizations. These variables were selected based on two criteria: first, the results of a basic study on the classification of CP (Aoki & Kogayu, 2021), and second, a bias of more than two-thirds (66.6%) in the frequency of occurrence in this study (Table 2).

The first and second axes were created from the results. The eigenvalue for the first axis was 0.37, and its correlation coefficient was 0.61; the second axis had an eigenvalue of 0.28 and a correlation coefficient of 0.53. The first axis was interpreted as an axis of cognitive disarray because no special score (−1.27) had a lower category score, and Levels 1 (0.96) and 2 (1.31) had a higher one. The second axis could also be interpreted as the axis of affective explanation because no affection (−1.50) had a lower category score, and unpleasant affection (0.35), ambivalent affection (0.82), and pleasant affection (1.77) had a higher one. The category scores are shown in Table 3.

Table 3. Score of Category

 Category Axis Ⅰ    Axis Ⅱ  

Note. AxisⅠ= the axis of cognitive disarray

;AxisⅡ= the axis of Emotional explanation. Rounded to the second decimal place.

Appropriate is appropriateness of a projected color.

In-appropriate is in-appropriateness of a projected color.

A two-dimensional scatter diagram is shown in Figure 1 with 19 items for category score: category scores with similar values are placed near each other, using the Ward method of cluster analysis (Figure 1).

Figure 1. Classification of CP responses

Note. AxisⅠ= the axis of cognitive distortions;AxisⅡ= the axis of Emotional explanation.

Discussion:

Response explanations:

        More than half of the CP responses were accompanied by words expressing affect. Among the emotional expressions, the largest percentage of CPs (22.1%) included only pleasant emotions, while 17.6% of CPs included only unpleasant emotions and 16.2% were mixed. This result is inconsistent with Weiner's (1998) findings that "most CPs almost always contain beautiful flowers or colorful butterflies" and that "few people express colored projections such as bloody or sickness in their CPs.

Color Projection: 27

        The results showed that about half of the CP responses included adjectives with distinct pleasant or unpleasant expressions, suggesting that CP is often an indicator of affects. While "a beautiful blue butterfly" and "colorful and beautiful leaves" may represent a denial of discomfort, "a beautiful blue butterfly, but unhappy and sad" and "a beautiful red dress, but tattered" may be failures to deny discomfort. In addition," bloody, drowned body. bright red" and "skinned, tattered animal, brown" express unpleasant affect directly, and there seems to be no factor to deny it. On the other hand, the responses of CPs without affect expression, such as "blue sky and white clouds" and "green lamp," may be indicators of affect in terms of their use of color, but it may be necessary to consider meanings other than affect.

Classification of CPresponses:

        CP responses were classified into three types. Interpreting the first axis as cognitive confusion and the second axis as affective expression, the three types of CP responses can be interpreted from two aspects: cognitive and affective.The CP with only pleasant affective expressions, consistent with the interpretive hypotheses of Z.A. Piotrowski (1957) and Weiner (1998), constituted a group with multiple specific colors, no special scores, location D, and a light gray card area. These CP responses projected multiple different colors for each part ("pink for the flower part, blue for the ribbon"), suggesting that subjective and active color change and response making by the examinee may have taken place, and in this respect, the possibility of affective defense and denial of unpleasant affect was also suggested. On the other hand, the CP with unpleasant affect was often accompanied by severe disturbances in special scores such as DR and INCI/FAB/CON, especially Level 2. The large number of cases in which the projected colors did not match the content ("dead leaves, fresh green") may be more appropriate to be considered as an indicator of severe cognitive disorganization rather than an affective defense mechanism as in the conventional hypothesis of CP. The large number of DRs also suggested the possibility that the participants were thinking fluidly by repeating words in an attempt to cope with the discomfort, but that the discomfort was not dispelled as a result. This type of CP was also assumed to be a shock to the inkblot as a whole, since it was grouped together with location W. Furthermore, the responses of CPs that did not involve any affective expression were often located in the black ink area, location Dd, but they were not accompanied by any special scores, the colors were consistent with the content, and many of them did not involve any cognitive problems. Considering the fact that they deliberately responded to location Dd and added cognitively appropriate but unique colors to it, it is possible that they were coordinating the content of their responses, i.e., the validity of what they perceived, but it is not possible to judge this type of CP only on the basis of this paper.

        The results of this study show that not all CP responses involve pleasant perceptions, and not all CPs can express denial of unpleasant affect. CP may be classified into several types according to the characteristics of their affective explanations. In many Rorschach systems, when CP appears, it is interpreted as abnormal, rare, and vulnerable affective coping with unpleasant affective experiences (Exner, 2003; Z.A. Piotrowski, 1957; Weiner, 1998). However, the results of the present study cast doubt on this traditional one-way interpretation hypothesis．The interpretation of CP needs to be considered in a complex way, including color, response explanation, FQ, and special scores.

        In interpreting the results of this study, the following limitations should be kept in mind. First, the data collection was not preplanned and took place in a naturalistic setting; therefore, there was no control group. Second, the data collection was done in a very small number of hospitals in Japan, so there may have been a sampling bias. Future studies need to examine more data to more accurately examine the characteristics of CP and its interpretation.

28: Aoki & Kogayu

References:

Aoki, S. (2009). A case report about color projection—Focusing on cases with numerous CP production. Abstracts of the 13th conference of Japanese Society for the Rorschach and Projective Method, p. 24.

Aoki, S. (2010). Discussion about color alteration responses –focusing on the responses ignoring colors on the cards and projecting other colors. Abstracts of the 14th conference of Japanese Society for the Rorschach and Projective Method, p. 31.

Aoki, S. (2011). Reconsider on the color projection on the Rorschach test. XX international Congress of Rorschach and Projective Methods abstract book, p.164–165.

Aoki, S. (2013). Study of color projection in the Rorschach test. Journal of Japanese Clinical Psychology, 31(4), 586-596.

Aoki, S., & Kogayu,N (2021). Color projection in the Rorschach test. Rorschachiana, 42(1), 35-51

Exner, J. E. (2001). A Rorschach workbook for the Comprehensive system (5th ed). Rorschach workshops.

Exner, J. E. (2003). The Rorschach：A Comprehensive system, 1. Basic foundations and principles of interpretation (4th ed). New York: Wiley.

Hayashi, C. (1993). Quantification: Theory and method. Tokyo: Asakura Shoten.

Ishii, Y. (2003). Reconsideration of color projection response. Journal of Japanese Clinical Psychology, 21(3), 301–306.

Meyer, G. J., Erdberg, P., & Shaffer, T. W. (2007). Toward international normative reference data for the comprehensive system.

Journal of Personality Assessment, 89(Suppl. 1), S201–S216. doi: 10.1080/00223890701629342, PubMed: 18039164.

Mihura, J. L., Meyer, G. J., Dumitrascu, N., & Bombel, G. (2013). The validity of individual Rorschach variables: Systematic reviews and meta-analysis of the comprehensive system. Psychological Bulletin, 139(3), 548–605.

Nakamura, N., & Fuchigami, Y. (2007). ＆Tsugawa, R. Rorschach Comprehensive System data for a sample of, 240 adult nonpatients from Japan. Journal of Personality Assessment, 89,(supple.1), S97-S102.

Piotrowski, C. (2017). Rorschach Research through the Lens of Bibliometric Analysis: Mapping Investigatory Domain. Journal of Projective Psychology & Mental Health, 24,34-38.

Piotrowski, Z. A. (1957). Perceptanalysis. Philadelphia,no italics PA:. Libris.

Tibon Czopp, S., & Zeligman, R. (2016). The Rorschach Comprehensive System (CS) psychometric validity of individual variables.

Journal of Personality Assessment, 98(4), 335–342.

Weiner, I. B. (1998). Principles of Rorschach interpretation. Hoboken, NJ: Erlbaum.

Wood, J. M., Nezworski, M. T., Grab, H. N., & Lilienfeld, S. O. (2001). Problems with the norms of the Comprehensive System for the Rorschach: Methodological and conceptual considerations. Clinical Psychology: Science and Practice, 8, 397–402.

Wood, J., Garb, M., Nezworski, H. N., Lilienfeld, M. T., S. O., & Duke, M. C. (2015). A second look at the validity of widely used Rorschach indices: Comment on Mihura, Meyer, Dumitrascu, and Bombel. Psychological Bulletin, 141(1), 236–249