In the present study, we used two Stroop-word categorization tasks to test the automatic activation of red-female association in both conceptual and perceptual word processing. Experiment 1 showed that red font color facilitated categorization of feminine words and deteriorated categorization of masculine words, compared with green and gray font colors. Experiment 2 showed that the discrimination of red font color in low saturation levels could be enhanced by feminine words. Those results indicated an automatically activated association between red and female concept, which could accelerate word processing for both conceptual gender categorizations and perceptual color discriminations.
In Experiment 1, when categorizing masculine and feminine words, font colors were automatically processed. The learned associations between color and gender concept could be activated implicitly, interrupting the gendered words processing. When the gendered words and font colors were congruently associated, the categorization performance could have been accelerated, otherwise, it could have been impeded. Participants responded faster at categorizing feminine words presented in red font colors than in the other colors, whereas responded slower at categorizing masculine words presented in red font colors than in the other colors. Thus, a congruency effect of red-female association could have been activated, thereby biasing the gender categorization. Those results provided further evidence for the effect of red on female gender perception. For instance, a red colored body was more easily to be judged as female than male (Chen, Nakamura, & Watanabe, Under review).
Meanwhile, the congruency effect of red-female associations could also modulate the perceptual color discrimination. In Experiment 2, when discriminating low-saturated red font colors, the word meaning was automatically processed, congruent red-concept could also accelerate font color processing (Richter & Zwaan, 2009). Our participants responded faster and made fewer errors for discriminating red font colors in feminine words than masculine words. This enhanced discrimination performance for red font colors by feminine words indicated a congruent effect of red-female associations. Thus, feminine words could be readily to be perceived as red more than masculine words. Those results provided further evidence for the automaticity red-female associations, which could be strong enough to interrupt the perceptual color processing. Previous studies suggested that learned higher-order cognition could routinely “penetrate” early low-level perception. For instance, visual perceptions can be influenced by language representations, knowledge, emotions, motivations, and other such states (Collins & Olson, 2014; Bannert & Bartels, 2013; Bubl et al., 2010; Chen & Watanabe, 2021). Notably, limited congruency effect of green-male associations was observed. It might be related to that masculine concept was associated with blue colors more than green colors (Cunningham & Macrae, 2011; Chen et al., 2020; Jonauskaite et al., 2021). Future studies are in need to address the blue/green-male associations and testing the automaticity of those color-gender associations.
The robust red-female associations may stem from statistical learning with social gender categorizations. The mere exposure to co-occurrences of gender and color information in the social environment could provide a basis, which became deeply embedded through statistical learning and language development. For instance, the gender-stereotyped colors of toys and clothing since the early development, that parents commonly dress their baby girls in red/pink and their baby boys in blue/green colors (Auster & Mansbach, 2012; Cohen, 2013; Picariello et al., 1990; LoBue & DeLoache, 2011; Davis et al., 2021), and the widely using of red in female clothing and cosmetics in later life (Frank, 1990), may help to construct the red-female associations. Moreover, gender differences in color preferences, that female tend to prefer reddish colors (Saito, 1996; Hurlbert & Ling, 2007), may also lead to those associations. Furthermore, red colors are commonly used to represent love and romance (e.g., red-heart symbol for valentine’s day), and wearing red can increase sexual attractiveness (Elliot et al., 2010; Elliot & Pazda, 2012; Kramer & Mulgrew, 2018). The semantics associated with red and feminine concept may also contribute to red-female associations. These findings, together with other observed associations between red and different concepts (e.g., dominance, danger, anger), strengthen the nurture component of color associations (Fetterman et al., 2012; Moller et al., 2009; Soriano & Valenzuela, 2009; Pravossoudovitch et al., 2014; Mentzel et al., 2017). Future studies may explore cultural differences in those color-gender associations. For instance, in traditional Chinese culture, male in a higher official position was often dressed in red and female in green in some social occasions (e.g., “red-men green-women”), which may lead to different color-gender associations. Future studies may also explore whether color-gender associations can be established and/or mitigated by training. For instance, through repeatedly learning phrase with new color-gender associations and testing the strength of those associations to show the learning effect on the formation of color-gender associations. These future studies are valuable in revealing the nature of color-gender associations and further getting rid of gender-stereotypes in relation to colors.
One limitation of the current study is that the trial numbers for each saturation levels of red and green font colors in Experiment 2 were not equal (with 536, 544, 514, 529, 515 trials for each green font colors separately, and 528, 526, 530, 534, 524 for each red font colors separately). In data analysis, we used a mixed-effect modeling approach, including the color saturation as a random effect to control for this effect. Future study should set the color saturation levels as six instead of five. Moreover, it could be interesting to fit the psychometric function with well-controlled saturation levels of colors to show the effect of color-gender associations on color perception. Furthermore, future studies may examine how the color dimensions (i.e., lightness/saturation/hue) interacting with gender conceptions, to determine the precise variant of reddish color that associate with female concept.
In summary, we observed an automatically activated red-female association in both conceptual gendered word categorization and perceptual font color discrimination. Those results added to a growing body of evidence for learned color associations, pointing to the gender signaling value of red in social categorization. We propose that future studies examining how learned color-gender associations activating high-level semantic knowledge and low-level perceptual color processing are necessary to understand the learning process the stereotypes/associations are formed.