Effect of Color on Aggression: Refinement of Methodologies

Denise-Marie V. Ilagan & Girard Vincent S. Aquino

(2004)

Kalyaan College

An experiment was conducted to test the separate effects of saturation and brightness on emotional responses, particularly their aggressive tendencies. Five groups of 10 participants were given a colored questionnaire with the same hue, but different levels of saturation and brightness. Results showed no significant differences among the tests scores, which suggest that the two variables controlled might not have affected the emotional responses of the participants to the questions provided.

Colors have a major influence on humans since time immemorial. From prehistoric art, the clothes that we wear everyday, to the color-coded symbols we use at work. Color is what helps us appreciate the world, as we perceive it.

Over the past decades, psychologists have been trying to understand the dynamics of color and its function particularly on human endeavors. This knowledge about colors, among other things is believed to be helpful for therapies. The detection of emotional and other mental disorders through the use of color is an accepted tool in psychology, the Rorschach being the primary example (Weller & Livingston, 1988). Therefore, if we are to properly manipulate the utilization of colors, we have to understand current and existing investigations done on the effect of colors on human responses, and more importantly, distinguish what these researches have overlooked to having a more thorough conclusion. Shortcomings of past researches are what we hope to point and be aware of. This awareness of the gap provides us the opportunity for a more refined methodology to improve experimental designs.

Characteristics of Color

A problem with researches on color is their failure to control specific variables or characteristics innate in colors such as hue, brightness and saturation (Geiineau, 1981, in Valdez & Mehrabian, 1994). Hue is a color’s attribute that distinguishes it from other colors (e.g. red from blue; orange from yellow). Brightness is the lightness of a color. It is determined by how much white or black is present. For example, yellow, having more white, would appear lighter than yellow, having more black. Saturation on the other hand, is determined by how much gray is present. A color may be of the same hue, but have different levels of brightness or saturation. An example of this discrepancy would be the difference of pastel colors from neon colors. Researchers from different experiments might use the same hue to replicate an experiment, but unknowingly use a different saturation or brightness level, which will confound the entire procedure. Each hue, red for instance, may differ from other reds in the characteristics mentioned. A pastel red, having a low saturation or chroma level, would visibly differ in appearance from a neon red color, which has a higher saturation. Neon colors have a higher level of saturation, meaning there is less gray present in these colors than in pastels. However, being neon does not necessarily mean being bright. For this reason, researches that failed to control these variables are confounded. Most researches have been primarily interested on hue (Guilford, 1933), therefore the probable effects of a colors brightness (value) and saturation (chroma) level have been taken for granted.

Weller & Livingston (1988) used blue, pink and white colored paper in their experiment, failing however, to control, or even account for the colors’ brightness and saturation levels. Valdez and Mehrabian (1994) controlled these two other characteristics of color and found a significant difference in their results. Guilford (1933) reported that saturation level affects the preference of men and women. Miss Washburn (1911), as cited in Guilford (1933), found that lighter colors, from 36 colors varying in saturation and brightness, were preferred over darker ones. It is therefore important to control these variables to attribute results only to one particular characteristic of color.

We agree with Guilford that even if there is a mass of repeated results in the matter of color preference or color and affect, most researches fail to control the factors of lightness, and saturation when drawing conclusions about the effects of hue. The neglect or failure of researches to control these psychological dimensions may suggest one good reason for the inconsistencies of previous studies. From this, we plan to conduct an experiment, which aims to better experiments regarding colors, especially on how to accurately manipulate experiments on colors. Thus, our plan for the experiment would be to control for the factors of saturation and brightness, not hue.

Experimental Procedure

A sample of 50 Kalayaan College students was randomly and equally assigned into five different groups. Group 1 was given a colored questionnaire with a low level of saturation and a constant level of brightness. Group 2 was given a questionnaire that had a high level of saturation and a constant level of brightness. The first two groups were designed to test the effects of saturation on the emotional responses of the participants. Group 3 had a low level of brightness and group 4 had a high level. Groups 3 & 4 were in turn, designed to test the effects of brightness on the responses of the participants. Both groups had a constant level of saturation to attribute the probable difference only to the brightness of the colored questionnaire. A fifth group was given a white questionnaire that serve as the control group to have a basis for comparison.

Each questionnaire had the same set of questions printed using black ink. The options that were provided were the same with all five groups. The questionnaires had eight different scenarios designed to place the participants in a situation wherein they would have the option to act aggressively or not. They were given three options to choose from. Each letter for each number had a corresponding point, which ranges from 5 aggressive points, which represents the lowest level of aggressiveness, to a maximum of 15 aggressive points, which in turn represents a high level of aggressiveness. The total number of aggressive points would then be summated per group to determine the group’s level of aggressiveness. Participants were given an instruction to answer the questions on the questionnaire with the first answer that comes to their mind.

Results

A general trend among the scores could be seen as shown in Table 1. Group 1, which was presented a questionnaire with a low level of saturation and a constant level of brightness accumulated a score of 730 points according to our standardized aggression scale. Group 2, which had a colored questionnaire with a high level of saturation and a constant level of brightness only differed from group 1 by 10 points scoring a total of 740. Group 3, which was given a questionnaire with a low level of brightness, scored a total of 785 points, while group 4 with the high brightness questionnaire scored a total of 800 points. The control group was able to score an expected median total score of 755 points.

Table 1.1

We did a separate analysis for the two independent variables (i.e. saturation and brightness) because the two variables were not manipulated at the same time on the same questionnaire, therefore not allowing it to have a combined effect on the dependent variable. By computing for the F-ratio of the saturation and brightness groups separately, we would be able to determine whether each had its own effect on the dependent variable.

The F-ratio obtained for the first ANOVA, which was an analysis of groups 1, 2 & 5 (control) to test saturation was 0.139. The critical values for the first analysis of variance, is F (2, 27) = 3.35, p < 0.05. This means that saturation has no significant effect on aggressive responses among our participants, thus we fail to reject our null hypothesis. The second ANOVA, an analysis of groups 3, 4 & 5, design to test the brightness of the color was 0.485. The critical values for this second analysis of variance is F (2, 27) = 3.35, p < 0.05. This means that brightness has no significant effect on aggressive responses among our participants, thus we also fail to reject our null hypothesis on brightness.

Qualitative Data Analysis

From the responses we gathered in the manipulation check, when asked what they thought the experiment was about, the most common answer was that the experiment was about temper, patience and personality. This common response among the participants is significant because it might have affected their answering the questions, which were designed to make them choose between an aggressive response and one that is not. To add, all the participants said that the scenarios presented were easy to imagine, which suggests that they might have been able to place themselves in the given situations. Although most of the participants said that all of the options were realistic, and could very well be their reactions if they were actually put in the same situation. Some of them would say that they would react otherwise, and that option was not provided in the choices forcing to choose the closest possible answer. Most of them were able to relay a similar real life experience with those provided in the questionnaire. Among those who relayed a similar experience, all chose the option similar to their actual reaction, which would suggest that there is no discrepancy between what they actually did, and with what they answered in the questionnaire.

The participants were asked whether they paid attention to the color of the questionnaire or not. 19 out of 50 participants said that they paid attention to or noticed the color of the questionnaires. Among the 19 participants, 3 said they felt the color might have affected their response to the questions. This suggests that most of the participants, those who did not notice the colored background of the questionnaire, did not perceive the color of the questionnaire as part of the experiment.

Conclusion

In the analysis of the data from the experiment, we did not find a significant effect of the saturation and brightness on the aggressive scores of the participants. Although there is a general trend among the results, no significant evidence could support our hypothesis unless a study on a larger sample is conducted. We therefore fail to reject our null hypothesis due to the insignificance of our results, but we hope to be able to test the hypothesis on a larger sample, since we believe that the insignificance of the results that we gathered were attributable to the limited number of samples utilized.

Back to Index

References:

Guilford, J.P. Affective Value of Color as a Function of Hue, Tint & Chroma. Journal of

Experimental Psychology, 17, 342-370, 1933.

Guilford, J. "There is System in Color Preferences" Journal of Optical Society of America, vol.

30, 455-459, 1940.

Lazreg, Cecilia and Mullet, E. “Judging the Pleasantness of Form-Color Combinations”

American Journal of Psychology”, Vol. 114, No. 4, 511-533, 2001.

Norman, Ralph and S. William. "Color and Affect: A Review and Semantic Evaluation"

Journal of General Psychology, vol. 46, 185-223, 1952.

Valdez, Patricia & Mehrabian, A. “Effects of Color on Emotions” Journal of Experimental

Psychology”, Vol. 123, No. 4, 394-409, 1994.

Weller, Leonard and Livingston, R. "Effect of Color of Questionnaire on Emotional

Responses" Journal of General Psychology, vol. 4, 433-440, 1988.

Wexner, Lois. "The Degree to Which Colors (Hues) Are Associated with Mood-Tones" Journal

of Applied Psychology, vol. 38, 432-435, 1954.