Abstract
This study explored whether vertical position affects social categorization of the rich and the poor. Experiment 1 used high- and low-income occupations as stimuli, and found participants categorized high-income occupations faster when they were presented in the top vertical position compared to the bottom vertical position. In Experiment 2, participants responded using either the “up” or “down” key to categorize high- and low-income occupations, and responded faster to high-income occupations with the “up” key and low-income occupations with the “down” key. In Experiment 3, names identified as belonging to either rich or poor individuals were presented at the top or bottom of a screen, and the results were the same as in Experiments 1 and 2. These findings suggest that social categorization based on wealth involved perceptual simulations of vertical position, and that vertical position affects the social categorization of the rich and the poor.
Introduction
Embodied theories argue that thinking involves perceptual simulation, and higher-level concepts require much more perceptual content than basic cognition (Barsalou, 1999). Previous research has found that social-category knowledge, such as male or female and Republican or Democrat, is grounded in perceptual simulation and sensorimotor activity (Slepian, 2015; Slepian, Weisbuch, Rule, & Ambady, 2011; Slepian, Rule, & Ambady, 2012; Zhang, Li, Eskine, & Zuo, 2014). To extend the current literature, the present study examined whether cognitive representations of the rich and the poor, which is one of the most important dimensions of social categorization (Harvey & Bourhis, 2013; Leahy, 1981), involved the perceptual simulation of vertical position. Specifically, we investigated if cognitive representations of the rich and the poor involved perceptual simulations of vertical position, and if different vertical positions could affect the social categorization of individuals as either rich or poor.
The association between the rich (poor) and high (low) vertical position
Perception of others based on wealth is an important and common social categorization dimension (Harvey & Bourhis, 2013; Leahy, 1981). Even children between the ages of 3 to 5 can understand the terms “poor” and “rich,” and can classify people as such (Ramsey, 1991). Furthermore, children between the ages of 6 to 11 begin to develop beliefs that the poor and the rich have differences regarding certain traits, such as intelligence and motivation (Leahy, 1981). Many studies have found that individuals associate the rich with positive valences and the poor with negative valences (Bjornsdottir & Rule, 2017; Piff & Moskowitz, 2018). For example, Cozzarelli, Wilkinson, and Tagler (2001) found that the rich are associated with positive behaviors and traits (such as working harder, healthier, more intelligent, friendlier, more ethical, and more responsible). The poor are associated with negative behaviors and traits (such as stupid, lazier, dirtier, and more violent). Bjornsdottir and Rule (2017) found that individuals classify faces that look good, smarter, and healthier as faces of the rich. Piff and Moskowitz (2018) found that people tend to associate the rich with satisfaction and pride, while the poor are associated with fear. A recent study by Bjornsdottir and Rule (2020) also found that negative emotions such as sadness, disgust, and anger are associated with the poor, while positive emotions, such as happiness, are associated with the rich.
Research in the field of embodied cognition has found that abstract positive and negative valences are associated with high and low vertical positions, respectively. Compared with low position, high vertical position is associated with positive valence (such as good and positive traits) (Meier, Hauser, Robinson, Friesen, & Schjeldahl, 2007; Meier & Robinson, 2004). For example, Meier and Robinson (2004) found that judgment of positive words was faster when words were presented in the up rather than presented in the down position, whereas judgment of negative words was faster when words appeared in the down rather than appeared in the up position. Since the rich are associated with positive valence, and positive valence is associated with high position, the rich should be associated with high vertical position. Similarly, since the poor are associated with negative valence and negative valence is associated with low position, the poor should be associated with low vertical position. Based on the above research, we proposed that the rich are associated with the higher vertical position, and the poor are associated with the lower vertical position. We hypothesized that categorization based on wealth may involve perceptual simulation of vertical position, and can be inferred by the perception of differences in vertical positions.
The embodied bases of social categorization
Social categorization is one of most important processes of social cognition, and our hypotheses are compatible with past work on embodied bases of social categorization, such as gender categorization. Slepian et al. (2011) first found that representations of gender are grounded in the sensorimotor experience of different levels of hardness, and men are associated with “tough,” while women are associated with “tender.” In their study, participants were required to continuously squeeze either a soft or hard ball while categorizing sex-ambiguous faces. The findings showed that more faces were categorized as female by participants squeezing the soft ball, and more faces were categorized as male by participants squeezing the hard ball (Slepian et al., 2011). On the basis of this research, Zhang et al. (2014) further showed that the process of gender categorization utilizes perceptual simulations of vertical position. They found that responses to female faces were faster when the female faces were in the “down” rather than “up” position, and vice versa for male face categorization (Zhang et al., 2014). Recently, Lamer, Weisbuch, and Sweeny (2017) also found that vertical position can influence visual perceptions of gender, as participants perceived faces presented in lower positions as more feminine.
Aside from social categorizations of gender, the categorization of political ideology has also been shown to involve sensorimotor processes. Slepian (2015) instructed participants to categorize target faces as either “Republican” or “Democrat,” with half the trials including priming with a hard object and the other half including priming with a soft object. The social categorization process of Democrat or Republican was shown to be affected by sensorimotor experience, as the results indicated that Democrats were associated with softness, while Republicans were associated with hardness.
Overview of present study
In the present study, we examined how vertical position affects the social categorization based on wealth. We hypothesized that participants would categorize the rich faster when presented in a higher, rather than lower, vertical position. Conversely, the categorization of the poor would be faster when presented in a lower, rather than higher, vertical position. One may ask whether social power judgment and social categorization on wealth are the same construct in different labels. We think they are not, for several reasons. First, it should be pointed out that previous research has examined the association between the power judgment and vertical positions (Schubert, 2005; Schubert, Waldzus, & Giessner, 2009). For example, Schubert (2005) also indicated that judgments of power can be affected by the target’s perceived vertical position, and response times to the power were faster when it was presented at the top of a screen, compared to at the bottom. This research only showed that vertical positions can influence power judgment, rather than the process of social categorization of the rich and the poor. Second, social power and wealth are not equivalent concepts. Social power means the ability to influence others or control others’ outcomes (Fiske, 1993). Wealth, on the other hand, is the possession of a large amount of money. Although some people with social power are wealthy, individuals with power are not necessarily wealthy, and wealthy people do not necessarily possess social power.
We designed three experiments to test our hypotheses. In Experiment 1, high- and low-income occupations were presented at either the top or bottom of a screen to explore whether vertical position affected their categorization. In Experiment 2, we again chose high- and low-income occupations as stimuli; however, participants responded with either the “up” or “down” key. In Experiment 3, we used names described to participants as belonging to either rich and poor individuals as stimuli, and further explored whether vertical position affected categorization of the rich and the poor. This research was reviewed and approved by the committee for the protection of subjects at the authors’ university. Written consent was also obtained from each participant before the experiment according to the established guidelines of the committee.
Experiment 1
Methods
Participants and design
The participants were 60 Chinese college students (11 males), ages ranged from 17 to 28 years, with an average age of 21.87 (SD = 2.09) years. The experiment assistant randomly asked the college students met on the university campus to participate in the experiment. After the experiment, each participant received a mug as a reward. The experiment had a 2 (occupational type: high-income occupation vs. low-income occupation) × 2 (vertical position: top vs. bottom) repeated measures design. Sample size (n=57) was determined using G*Power, α = .01 for Type I error, a power of 0.85, and an effect size of ηp2=0.11 (based on previous research; Schubert, 2005). We recruited 60 subjects to avoid an insufficient sample due to invalid responses. The same method was used in Experiment 2 and Experiment 3 to recruit participants.
Materials
The experimental materials were the names of 16 high-income occupations and 16 low-income occupations. To select the occupations to use as stimuli, we first chose 78 occupations through an internet search. Then, 19 college students who did not participate in the main experiment rated each occupation on a 5-point scale for income (low income, “1” to high income, “5”) and familiarity (low level of familiarity, “1” to high level of familiarity, “5”). We then chose the 16 occupations with the highest incomes (M = 4.21; SD = 0.38) and the 16 with the lowest incomes (M = 2.17; SD = 0.24). Scores for the high-income occupations were significantly higher than the low-income occupations, t(15) = 20.47, p < 0.001. The familiarity of all occupations was above 3 (M = 3.73; SD = 0.20).
Procedure
Participants completed the task in a quiet room. A fixation cross was first presented for 300 ms at the center of a computer screen. A subsequent fixation cross was presented either 3 cm above (or 3 cm below) the central fixation cross for 300 ms. A third fixation cross was presented for 300 ms either 6 cm above or 6 cm below the central fixation cross, in the same vertical direction as the second cross. Names of occupations then appeared either 8 cm above or 8 cm below the central fixation cross for 2000 ms, or until participants pressed a response key. All fixation crosses and occupations were centered horizontally on the computer screen. Participants were asked to judge the income level of each occupation by pressing a key on the keyboard to signify either “high-income” (F/J keys) or “low-income” (J/F keys) as quickly and accurately as possible. The keypress of reaction was counterbalanced between participants. If the reaction time was longer than 2000 ms, the phrase “please be quicker” was presented for 700 ms. The methods used for stimuli presentation and participant responses were based on previous research (Meier & Robinson, 2004). The experimental program was created using E-Prime 2.0, and participants used an Asus laptop (the screen size was 14 inches) to complete the experiment. The participants were about 60 cm away from the screen.
Experiment 1 included 128 trials. Each occupation appeared four times. High-income occupations appeared 32 times at the top of the screen and 32 times at the bottom of the screen, as did the low-income occupations. Before the formal experiment, participants completed a practice experiment of 24 trials. The stimuli from the practice experiment were not used in the formal experiment.
Results and discussion
Mean categorization reaction times were the dependent variables. Reaction times outside 2.5 standard deviations (152, 1.98% of the trials) were excluded from data analysis, as were trials in which errors were committed (945, 12.30% of the trials). Latencies were then log-transformed to normalize their distribution; however, untransformed mean reaction times are reported for ease of interpretation.
The transformed latencies were submitted to a 2 (occupational type: high-income vs. low-income) × 2 (vertical position: top vs. bottom) ANOVA. The analysis revealed no main effects of occupational type, F(1,59) = 3.69, p = 0.06, η2 = 0.06, or vertical position, F(1,59) = 0.75, p = 0.39, η2 = 0.01. As we expected, the interaction between occupational type and vertical position was significant, F(1,59) = 19.117, p < 0.001, η2 = 0.25. Further simple effects analysis demonstrated that responses to high-income occupations in the higher position (M = 593.30, SD=68.58) were significantly faster than in the lower position (M = 607.04, SD = 64.04), F(1,59) = 13.91, p < 0.001, η2 = 0.19. In contrast, responses to low-income occupations in the higher position (M = 617.70, SD = 66.39) were slower than in the lower position (M = 607.63, SD = 59.03), and the difference was significant, F(1,59) = 5.08, p = 0.03, η2 = 0.08.
These results provided initial support for our hypothesis that social categorization based on wealth involves the perceptual simulation of vertical position. Wealth is associated with a higher vertical position, and poverty is associated with a lower vertical position. If the social category was consistent with the perceptual simulation (e.g., high-income occupations were presented at the top of screen), the process of social categorization was facilitated. If the social category was incongruent with the perceptual simulation (e.g., high-income occupations were presented at the bottom of screen), the process of social categorization was hindered. To verify the results of Experiment 1, according to previous research, Experiment 2 used response buttons with different vertical positions (“up” or “down”) to manipulate the process of perceptual simulation (Tang, Zhou, Zhang, & Zhu, 2018). We predicted that responses to high-income occupations would be faster using the “up” key, whereas responses to low-income occupations would be faster using the “down” key.
Experiment 2
Methods
Participants and design
The participants were 60 Chinese college students (4 males). Ages ranged from 18 to 28 years, with an average age of 20.75 (SD = 2.43) years. After the experiment, each participant received a mug as a reward. The experiment had a 2 (occupational type: high-income vs. low-income) × 2 (position of response key: up vs. down) repeated measures design.
Materials
The materials were the same as in Experiment 1.
Procedure
Participants were asked to sit in front of a computer screen to complete the experiment. A fixation cross first appeared at the center of the screen for 300 ms. Then, the name of a high- or low-income occupation appeared in the same position for 2000 ms, or until the participant recorded a response. Participants were required to judge whether each occupation had a high income or low income by pressing the “up” (↑) or “down” (↓) key on the keyboard (on the keyboard, ↑ key is directly above ↓ key) as quickly and accurately as possible. If response times were longer than 2000 ms, the phrase “please be quicker” was presented for 700 ms. The experimental program was created using E-Prime 2.0, and participants used an Asus laptop (the screen size was 14 inches) to complete the experiment. The participants were about 60 cm away from the screen.
Experiment 2 included 128 trials, divided into four blocks. According to the relationship between occupational type and response key, there were two block types: consistent and inconsistent. In the consistent block, participants were asked to press the “up” key for high-income occupations and the “down” key for low-income occupations. In the inconsistent block, participants were asked to press the “down” key for high-income occupations and the “up” key for low-income occupations. To eliminate the influence of block order on the results, the order of the blocks was counterbalanced among participants. For half the participants, the order of the blocks was inconsistent, consistent, inconsistent, and consistent, while for the other half, the order of the blocks was consistent, inconsistent, consistent, and inconsistent. Each occupation was presented four times. The high-income occupations were assigned to the “up” key twice and to the “down” key twice, as were the low-income occupations. The first block consisted of 12 practice trials and 32 formal trials, and the other three blocks consisted of 6 practice trials and 32 formal trials.
Results and discussion
Mean categorization response times served as dependent variables. Reaction times outside 2.5 standard deviations (155, 2.02% of the trials) were excluded from data analysis, as were trials in which errors were committed (832, 10.83% of the trials). Latencies were then log-transformed to normalize their distribution; however, untransformed reaction times are reported for ease of interpretation.
The transformed latencies were submitted to a 2 (occupational type: high-income vs. low-income) × 2 (key position: up vs. down) ANOVA. The analysis revealed the main effect of occupational type was significant, F(1,59) = 9.64, p = 0.003, η2 = 0.14, and the main effect of key position was also significant, F(1,59) = 56.51, p < 0.001, η2 = 0.49. As we expected, the interaction between occupational type and key position was significant, F(1, 59) = 36.71, p < 0.001, η2 = 0.38. Further simple effect analysis showed that participants responded significantly faster to the high-income occupations with the “up” key (M = 708.21, SD = 97.40) than with the “down” key (M = 807.39, SD = 149.96), F(1,59) = 74.82, p < 0.001, η2 = 0.56. In contrast, responses to low-income occupations with the “up” key (M = 794.31, SD = 148.97) were slower than with the “down” key (M = 765.68, SD = 112.78), and the difference was significant, F(1, 59) = 4.76, p = 0.03, η2 = 0.08.
As with the results of Experiment 1, Experiment 2 also found that representation of social categories of rich and poor are embodied with vertical position. In Experiments 1 and 2, we used high-income occupations, such as doctor, professor, and so forth. Previous studies have found that judgments of high or low social power are influenced by vertical position metaphors (Schubert et al., 2009). Therefore, we considered that the results of Experiments 1 and 2 may have been influenced by the perception of social power attached to high-income occupations. Thus, in Experiment 3, we investigated this issue by directly using names identified as belonging to either rich or poor individuals, which did not indicate any information relevant to social power, to verify the results of Experiments 1 and 2.
Experiment 3
Methods
Participants and design
The participants were 73 Chinese college students (23 males). Ages ranged from 17 to 26 years, with an average age of 20.25 (SD = 1.84) years. After the experiment, each participant received a mug as a reward. The experiment had a 2 (name type: name of the rich vs. name of the poor) × 2 (vertical position: top vs. bottom) repeated measures design.
Materials
Thirty full Chinese names (e.g., “张建国”; 15 typical male names and 15 typical female names) used in a previous study were selected as stimuli (Zhang et al., 2014).
Procedure
Participants arrived at the laboratory individually and were asked to sit in front of a computer screen. Participants were told that names in red or blue font would appear successively on the screen, and names in red font were rich/poor individual’s names, and names in blue font were poor/rich individual’s names (the match between font color and rich/poor was counterbalanced between participants). The task was to determine whether each name was of a poor individual or a rich individual, and to press corresponding keys (“the rich” key or “the poor” key).
A fixation cross was first presented for 300 ms at the center of a computer screen. A subsequent fixation cross was presented either 3 cm above or 3 cm below the central fixation cross for 300 ms. A third fixation cross was presented for 300 ms either 6 cm above or 6 cm below the central fixation cross, in the same vertical direction as the second cross. Names in red or blue font then appeared either 8 cm above or 8cm below the central fixation cross for 2000 ms, or until participants pressed a response key. Participants were asked to quickly and accurately judge whether the name was of a rich or a poor individual, and to press “the rich” key (F/J keys) or “the poor” key (J/F keys). The keypress was counterbalanced between participants. The experimental program was created using E-Prime 2.0, and participants used an Asus laptop (the screen size was 14 inches) to complete the experiment. The participants were about 60 cm away from the screen.
In Experiment 3, each name was presented four times, for 120 trials in total. Fifteen red names were presented at the top of the screen twice and at the bottom of the screen twice, as were 15 blue names. Before the formal experiment, participants were asked to complete 24 practice trials. To ensure participants had established the association between font color and rich or poor names, an accuracy rate of 100% needed to be reached before participants could begin the formal experiment. The interviews after the experiment indicated that all the participants reported judging the names based on the wealth activated by the color rather than the color per se.
Results and discussion
Mean categorization reaction times were the dependent variables. Reaction times outside 2.5 standard deviations (140, 1.60% of the trials) were excluded from data analysis, as were inaccurate responses (363, 4.14% of the trials). Reaction times were then log-transformed to normalize their distribution; however, untransformed means are reported for ease of interpretation.
The transformed latencies were submitted to a 2 (name type: names of the rich vs. names of the poor) × 2 (vertical position: top vs. bottom) ANOVA. The analysis revealed no main effects of name type, F(1,72) = 0.54, p = 0.47, η2 < 0.01, or vertical position F(1,72) = 0.09, p = 0.77, η2 = 0.001. As expected, the interaction of name type × vertical position was significant, F(1,72) = 9.19, p= 0.003, η2 = 0.11. Simple effects analysis demonstrated that responses to names of the rich presented at the top of the screen (M = 455.79, SD = 51.82) were significantly faster than responses to names of the rich presented at the bottom of the screen (M = 462.22, SD = 50.88), F(1,72) = 5.78, p = 0.02, η2 = 0.07. In contrast, responses to names of the poor presented at the top of the screen (M = 459.77, SD = 53.73) were significantly slower than responses to names of the poor presented at the bottom of the screen (M = 453.56, SD =47.42), F(1,72) = 4.89, p = 0.03, η2 = 0.06.
General discussion
In three experiments, we demonstrated that the representation of social category (e.g., the rich and the poor) involved perceptual simulations of vertical position, and that vertical position can affect social categorization of the rich and the poor. Experiment 1 found that responses to high-income occupations presented at the top of a screen were significantly faster than responses to high-income occupations presented at the bottom of a screen, while the reverse was found for low-income occupations. In Experiment 2, we used “up” and “down” keys to manipulate vertical position perceptual simulations, and received the same results as in Experiment 1. In Experiment 3, we used names identified as belonging to either rich or poor individuals as stimuli, and received the same results as in Experiments 1 and 2. Our findings were in line with previous studies showing that representation of social categories involves proprioception (Slepian, 2015; Slepian et al., 2012; Slepian et al., 2011; Zhang et al., 2014), and we extended these prior findings to another important social category: the rich and the poor.
Our findings provided new evidence that early perceptual processes contribute to social categorization (Cloutier, Mason, & Macrae, 2005; Macrae, Quinn, Mason, & Quadflieg, 2005). Although previous research demonstrated that facial cues, such as blurred or quickly presented faces, can influence the speed of social categorization (Cloutier et al., 2005), the present study indicated that vertical position may also contribute to the process of social categorization. Just as visual facial cues (such as gender, race, and age) have a ubiquitous effect on categorical thinking regarding other people, different vertical positions could have an important influence on social categorization based on wealth. For example, an individual’s vertical position (e.g., sitting on a high or low stool) might influence impressions others have of him or her.
Barsalou (1999) indicated that the schematization of experiences is a vital source of information in the development of perceptual symbols. In the process of individual perception, wealth is a very common and important dimension of social categorization in everyday life (Harvey & Bourhis, 2013; Leahy, 1981; Ramsey, 1991). Extensive studies have found that both adults and children associate positive valence (such as positive emotions) with the rich, and associate negative valence (such as negative emotions) with the poor (Bjornsdottir & Rule, 2017; Piff & Moskowitz, 2018; Baldus & Tribe, 1978; Sigelman, 2012), while researchers focused on embodied cognition have found that abstract positive valence is relevant with high vertical positions, and abstract negative valence is relevant with low vertical positions (Meier, Hauser, Robinson, Friesen, & Schjeldahl, 2007; Meier & Robinson, 2004). Thus, we proposed vertical positions may become the embodied base of the representation of the rich and the poor, and the rich may be associated with the high vertical position and the poor may be associated with low vertical position. The association between the rich/poor and different vertical positions may give rise to mental representations of the rich and the poor; thus, the process of social categorization based on wealth can be influenced by the perceptual experience of the vertical position.
Given the important, foundational role of social categorization in the process of social cognition, such as stereotyping and discrimination (Zhang, Li, Sun, & Zuo, 2018), the findings of the present study have several implications. For example, categorization based on wealth is an important dimension of individual perception (Harvey & Bourhis, 2013; Leahy, 1981), and can lead to a variety of outcomes, such as stereotypes of the poor and the rich. Our findings raise the possibility that perceptual simulations of vertical position may provide a foundation for the representations of social categories of the rich and the poor. Thus, further research could examine whether perceptual simulations of vertical position affects stereotyping based on whether one is rich or poor. Prior studies demonstrated that a higher vertical position was associated with good and a lower vertical position was associated with bad (Meier, Robinson, & Caven, 2008). Similarly, we can infer that the process of social categorization of the rich (the poor) involved the perceptual simulation associated with a higher (lower) vertical position which is related to a positive (negative) valence. Thus, perceptual simulation involving social categorization may play an important role in the formation of negative stereotypes toward the poor. Additionally, the present study showed the association between social categorization of the poor and the rich as well as the perceptual simulation of vertical position in Chinese culture. Future research should be conducted to verify whether similar associations exist in other cultural contexts.
Previous research has shown that vertical position metaphors affect power judgments (Schubert, 2005; Schubert et al., 2009). The present study is different from those previous studies in the following aspects. First, in our study, participants were asked to categorize the poor and the rich, which had no explicit association with power judgments in our experiments. In addition, social power and wealth are not exactly equivalent concepts. For instance, a wealthy person does not necessarily have social power, and social power is not always accompanied by wealth. Second, in Experiment 3, we used names identified as belonging to either rich or poor individuals, which did not indicate any information regarding social power, and again found categorization of the poor was affected by vertical position perceptual simulations. However, we should point out that there is some degree of overlap between the two concepts of power and wealth and we did not empirically tease apart power and wealth in the present study, which is a limitation. Thus, future research must clarify whether the relationship between social categorization based on wealth and perceptual simulations of vertical position is due to the conflation between power and wealth.
It should be noted that there are both conceptual and empirical critiques on the research of metaphor-focused embodied cognition (Lakens, 2014; Papesh, 2015; Skulmowski, & Rey, 2018; Wilson, & Golonka, 2013). Thus, future research should further verify the conclusion of this research. For example, using children as participants, future research could explore the development of the association between social categorization based on wealth and metaphors of vertical position. Based on ERP technology, future research could verify the conclusions of this study by testing whether incongruent target stimuli (the social category is incongruent with the perceptual simulation, e.g., high-income occupations presented at the bottom of the screen) would induce larger ERP amplitude.
Overall, the present study demonstrated that the process of social categorization based on wealth involved perceptual simulations of vertical position.
Ethical compliance section
Compliance with ethical standards
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.
Declaration of conflicting interest
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and publication of this article: This work was supported by the National Natural Science Foundation of China (Nos. 31760287, 31400902), and the Fundamental Research Funds for the Central Universities (No. GK202103134).
Informed consent
Informed consent was obtained from all individual adult participants included in the study.