variations. Whether it is the clothes we wear, the car we drive, or the food we eat, we are all attached to certain colors and take our choices seriously. In our consumer culture of aggressive advertising and savvy salesmanship, product color schemes are presented in an endless array that companies hope will be forever associated with their merchandise. Color terms are used in popular culture to describe a variety of emotional states and moods. If you are sad, you are feeling blue, if you are yellow, you are a coward, and if you covet your neighbor’s new lawnmower, you are green with envy. Despite the unavoidable impact of color on our lives and our everyday visual environment, the question remains exactly how and to what extent color perception influences our mood. The focus of this paper is to discuss notable studies from the body of work on color perception and emotional state, its history, empirical results, and the difficulties and limitations of research in this area.
Psychologists have a long-standing interest in color. With a wide variety of approaches and diverse angles of examination, researchers have established a sizeable body of work on the subject. Exploration of color-mood associations, color meaning, color preference, and color-personality relations are but a few of the topics covered in the literature. Researchers have also used an assortment of methods and measurements in color experiments. Wexner (1954) conducted one of the earliest color-mood studies published in a psychological journal. In this study, ninety-four participants were instructed to choose a mood descriptive adjective from a presented list to best describe a series of eight colors randomly displayed on a gray cardboard background at the front of a classroom. Results of the Wexner study showed a significant association between certain colors and moods. Participants indicated a strong relationship between red and exciting/stimulating, blue and secure/comfortable, orange and distressed/disturbing/upset, and black and despondent/dejected and powerful (1954).
Murray and Deabler (1957) replicated the study done by Wexner (1954) except for two differences; eight sheets of art paper were placed on one piece of cardboard due to room constraints, and the colors red and blue were imperceptibly lighter in comparison to the shades in Wexner’s experiment (Murray & Deabler, 1957). They hypothesized that subculture, socioeconomic status, and mental health influences an individual’s choice of words in association with color. Results suggested that socioeconomic status was most often associated with between group variance, however, similar to the Wexner study, specific colors (red and black) had the same word associations across all groups (Murray & Deabler, 1957). These early studies gave some empirical credence to the intuitive belief that color affects mood, but the research methods used were not without flaws. Neither study, nor indeed much the subsequent research, took into account several important and potentially confounding variables.
A lack of solid methodology has hindered color-mood research due in part to the complexity of visual perception, but also because much of the research on how color affects mood has failed to control for a variety of confounding variables. Researchers failed to account for lighting, their source of color stimuli, and a systematic means of mood descriptors. Gelineau (1981) controlled for lighting and color stimuli by using a clearly described light source to address the former problem (that could easily be replicated by others), and by using the Munsell color system to account for the latter. Results from the Gelineau study indicated that hue variations and lighting conditions affect color preference when saturation was controlled for, but these findings were limited to participant self-report measures of like and dislike and not applicable to descriptions of emotional state (1981).
A series of experiments by Valdez and Mehrabian (1994) hypothesized that the brightness and saturation of a color would be positively correlated with pleasure and emotional arousal. They listed several problems with prior research that they sought to control for including imprecise color depiction, unreliable mood descriptor terminology, and uncontrolled lighting conditions. Like Gelineau’s study, Valdez and Mehrabian meticulously controlled lighting conditions (experiments were carried out in a windowless room with fluorescent lights designed to imitate natural daylight) and used the Munsell Color System to display ten hues with varied levels of saturation totaling seventy-six colors (1994). To address the problem of valid mood descriptors, they employed the Pleasure-Arousal-Dominance (PAD) Emotional Model (see Mehrabian, 1978 for description of PAD). Their results showed a correlation between color and emotion, but with more emphasis on color brightness and saturation than variety of color (Valdez & Mehrabian, 1994). While these findings yielded significant results, they also brought up difficult questions regarding environmental influence (i.e., how simply altering lighting conditions can change a person’s emotional response to a given color) on perception of color.
Despite the positive findings of the above-mentioned studies, some researchers sought a different approach to investigating color-mood associations. In their cross-cultural study on color-word associations, Hupka et al. (1997) controlled for confounding light and color perception variables by having the colors listed in word form. They reasoned that verbal associations are equivalent to visual associations in how they elicit emotional response to the primary emotions, and they hypothesized that fear and anger would show similar color associations across cultures and that the hue associations of envy and jealousy would vary due to cultural differences (Hupka et al., 1997). Their study consisted of participants from Germany, Poland, Russia, Mexico, Turkey, and the United States rating on a six point scale the degree to which they associated twelve select colors with fear, anger, envy, and jealously.
The benefit of using written color terms was consistency of stimuli, but this also called into question to what extent color stimuli or color-word associations were being measured. Differentiating between learned (envy and jealousy) and inherent (fear and anger) response to color was a logical step based on the literature. Hupka et al. found that there was indeed cross-cultural agreement in which colors are linked with fear and anger (black with fear, red with anger and fear), but cultural differences in color associations with jealously and envy (1997). The authors discussed color and emotion associations in the folklore and mythology of the studied cultures, and listed the following possible limitations of their study: only university students participated, more female (441) than male participants (220), only industrialized nations participated, and possible differences in cultures’ definitions of the terms color and emotion (Hupka et al., 1997). Although these limitations prevent generalizing results to all peoples in all parts of the world, findings from this study show some intriguing connections between color and emotion in a variety of cultural settings.
Nolan, Dai, and Stanley (1995) also used a novel approach to look at the relationship between color and emotional states. In their study, two hundred and sixty-one participants selected from seven printed color squares to answer a series of questions related to mood and color preference. Participants then completed the Beck Depression Inventory. The outcome showed a strong relation between black and brown color selection for three specific questions (color choice for current mood, color that best represents the participant, and favorite color) and participant’s score on the depression scale (Nolan et al., 1995). The researchers concluded, “activation of internal depressive schema may have motivated the choice of colors most closely reflecting popular conceptions of the representation of mood with color” and suggested that adding the most common favorite colors of blue and red to paint schemes in a variety of settings may stimulate positive moods in some people (Nolan et al., 1995).
Some researchers have sought to isolate the focus of emotional states and color perception relationships to particular demographic groups. Boyatzis and Varghese (1994) investigate color preferences in children. They conducted a study that examined how children felt about the colors pink, red, yellow, black, gray, green, blue, purple, and brown. Results indicated that children preferred bright colors to dark colors, but there were some differences when it came to gender (Boyatzis & Varghese, 1994). This was a comprehensive study that researched the preferences of young children, but whether adults would respond similarly had yet to be seen.
Hemphill addressed the adult population in a study on the emotional response of college students to color (1996). This study found that college students were not different than children when it came to the way they felt about certain colors, but like the Boyatzis and Varghese study, there were some differences in gender response. Results suggested females associated positive feelings with bright colors and negative emotions with dark colors, and men associated positive feelings with bright colors, though men did not associate negative feelings with dark colors as much as women did (Hemphill, 1996). These findings were limited by the small sample size of forty participants and gender comparison was not the central focus of Hemphill’s investigation.
More recent research conducted by Carruthers, Morris, Tarrier, and Whorwell (2010) focused on mood-color associations with the use of new color wheel instrument they developed specifically for their study. They discussed the shortcomings of projective color tests such as “the Color Pyramid test, the Rorschach Inkblot test, the Luscher Color test, the Lowenfeld Mosaic test, and the Stroop test,” and sighted the lack of a valid tools for measuring emotional states (Caruthers et.al., 2010). In their study, three participant groups (healthy, anxious, and depressed) were exposed to a total of thirty-eight colors in a color-wheel (eight main colors with four variations of each and four shades of gray) and chose a color they were most drawn to, a favorite color, and which color best represented their recent mood (Caruthers et.al., 2010). Participants from all three groups showed similar responses on what colors were their favorites and most drawn to, but in comparison to the healthy group, the anxious and depressed groups had significantly different responses to mood-color association and most often chose variations of gray to describe their mood (Caruthers et.al., 2010).
Color perception and how it relates to mood has and continues to interest researchers. The intuitive belief that color preference is related to and descriptive of emotional states has been supported by empirical research (real world variables and the complexity of visual perception notwithstanding). Investigations into color psychology have made great progress over the years. In the last two decades especially, psychologists have corrected for prior methodological mistakes and shown strong connections between color, moods, and emotions. Color-mood experimentation has taken many different approaches focused on a variety of associations. Despite early experimental design problems that resulted in inconclusive data, researchers have corrected for past mistakes and