Abstract
Psychology and behavioral sciences lack diversity in their participant samples. In visual perception, more specifically, common practice assumes that the processes studied are fundamental and universal. In contrast, cultural psychology has accumulated evidence of cultural variability in visual perception. In face processing, for instance, this cultural variability may sabotage intercultural relations. Policies aim to increase diversity in research, supporting cultural psychology, and to increase awareness among professional workforces, as well as the general population, concerning how cultural variability may influence their interpretation of another's behavior.
Keywords
Social Media
Psychology research often lacks diverse samples, assuming universal processes. However, cultural studies reveal systematic variability, especially in face processing, impacting intercultural relations. Policies should promote diversity, support cultural psychology, and raise awareness.
Key Points
Behavioral sciences, including visual perception, lack culturally diverse participant samples.
Abundant evidence indicates cultural variability in visual perception, even for processes typically assumed to be fundamental and universal.
Face processing—at the core of human social interactions—shows cultural variability, which may undermine intergroup relations.
Policies should aim to increase sample diversity in behavioral sciences. Research in cultural psychology, specifically designed to measure and understand cultural variability, should be disseminated more widely.
Policies should aim to increase awareness about the impact of culture on psychological functioning.
Introduction
Psychological science aims to understand the psychological functioning of all humans, not just a narrow subset. However, a seminal article (Arnett, 2008) revealed that research published in major APA journals between 2003 and 2007 focused on only 12% of the global population. This trend has remained stable over time, reflecting sample demographics from research as far back as 1988 and continuing through to 2018 (Thalmayer et al., 2021). Specifically, about 95% of study participants in major APA journals are from the United States, other English-speaking countries (Canada, United Kingdom, Australia, New Zealand), or Europe. Consequently, only 5% of these studies include participants from Asia, Africa, Latin America, or the Middle East, leaving 88% of the world's population underrepresented in psychological research.
The underrepresentation of a large portion of the world's population in psychological research is problematic because the psychological processes studied are not universally consistent. Numerous studies in cultural psychology have shown that many psychological processes are significantly influenced by the people's cultural environment of origin. Culture impacts motivation, decision-making, memory, emotions, attention, and even basic visual processes (Gutchess & Sekuler, 2019; Henrich et al., 2010; Wang, 2016). The acronym “WEIRD” describes samples typically from Western, Educated, Industrialized, Rich, and Democratic societies, who sometimes display unique results when compared to participants from other regions of the world (Henrich et al., 2010).
Cognitive psychology or visual perception has a long tradition of methods that extract the participant from the real world and puts them in a highly controlled environment (i.e., the lab) that goes back to the birth of scientific psychology (Arnett, 2008). This controlled practice links to studying fundamental, universal processes that are exempt from environmental influences (Wang, 2016). Nevertheless, even fundamental processes of visual perception can vary as a function of cultural environment origin (Gutchess & Sekuler, 2019). This article briefly overviews research demonstrating this cultural influence. Then, a focus on face processing explores effects of cultural upbringing and the implications for everyday life. Additionally, the review highlights the critical issue of stimulus diversity in face-processing studies. Finally, recommendations address both the broader problem of underrepresentation in psychology and specific challenges posed by cultural variations in face processing.
A Definition of Culture
Defining culture is no small feat, sociology and anthropology each offering dozens of definitions. The present article uses a broad, largely agreed-upon definition: Culture refers to individuals from the same geographical area who share similar ideas, customs, and practices (Heine, 2020). As such, the term culture may also refer to large groups of individuals sharing a context, such as the Western culture or East Asian culture.
One of the main disadvantages of using such a broad definition is that many studies report “cultural differences” without strong evidence that the differences observed are truly cultural (Matsumoto & Hwang, 2017). More specifically, verifying cultural variability in a process commonly compares individuals from Western and East Asian cultures because much evidence shows that they differ on multiple cultural variables (systems of values, social norms). Any difference between the two groups often leads researchers to assume that culture is the cause. But of course, something else, not cultural, may explain the difference. The two groups may differ in their exposure to a formal education system, their socioeconomic level, or even the physical environment (e.g., rural, urban). Such results still have a high value for the argument. The observation that individuals from different geographical regions exhibit distinct psychological processes, regardless of whether the underlying cause is cultural, underscores the importance of diversifying research samples. Investigating the source of this variability may foster a more comprehensive understanding of the process under study.
Cultural Variations in Visual Perception: A Brief Overview
Among the cultural and environmental influences on cognitive mechanisms, the most fundamental processes relate to perception. Evidence of differences in visual perception dates back to the early 20th century, demonstrating that individuals from some non-Western cultures do not experience well-known visual illusions, such as the Müller-Lyer (Rivers, 1905; Segall et al., 1963). Theoretically, exposure to different social systems may lead to cultural differences in how humans deploy attention and perceive the world (Nisbett et al., 2001). Specifically, Ancient Greek philosophy, which emphasizes individual power and freedom and the idea that different categories of objects are governed by their own physical laws, contrasts with Chinese philosophy, which emphasizes the relationship of individual agency to their group and collective agency. According to this theory, exposure to Greek versus Chinese philosophy is associated with differences in attentional focus. Cultures emphasizing relationships between individuals and their group increase attention to context, while cultures viewing individuals as governed by their own laws perceive individuals as independent of their context.
Attention to Background
One set of studies shows that individuals from cultures influenced by Chinese philosophy pay more attention to the background of visual scenes than those from cultures influenced by Greek philosophy. In a landmark study on the impact of culture on visual cognition, researchers asked participants to describe animated underwater scenes (Masuda & Nisbett, 2001). Participants’ memory was then tested by presenting objects (e.g., a fish, a frog) that were either part of the original scenes or not, with the objects shown against either the same or a different background. Japanese participants were more impaired than Americans at recognizing objects when presented with a new background. The authors concluded that Japanese participants pay more attention to context than Americans. This pattern was replicated in a study recording eye movements, showing that Chinese participants made more fixations on the background of visual scenes compared to Americans (Chua et al., 2005). Many studies further supported the idea that East Asians attend to the background of visual scenes more than Westerners. For instance, in a change-blindness paradigm East Asians detected background changes more easily than Westerners (Masuda & Nisbett, 2006). In a task evaluating the intensity of emotions expressed by a foreground cartoon character, emotions expressed by the background characters had more impact on the evaluations of East Asians than of Westerners (Masuda et al., 2022).
Breadth of Attentional Deployment; Global Versus Local Perception
Cultural differences also manifest in the breadth of attentional deployment. Another set of studies indicates that the Chinese group deploys their attention more broadly and/or has a more global perception than Westerners. In a color change detection task, Chinese participants were more efficient at detecting changes farther from the center of a visual display, while American participants were better at detecting changes closer to the center (Boduroglu et al., 2009). Deploying spatial attention over a larger area is linked to decreased resolution in extracted visual information (Balz & Hock, 1997). Studies leveraging this phenomenon have shown that the spatial resolution for processing visual information is lower for Chinese participants compared to Americans and Canadians, both in simple target detection and face processing (Boduroglu & Shah, 2017; Tardif et al., 2017). This difference in lower-level perception is also reflected in memory studies. Westerner participants encode (or retrieve) visual stimuli in memory with greater detail specificity compared to East Asians (Leger & Gutchess, 2021).
Similarly, East Asians tend to rely on more “global” attentional strategies, while Westerners use more “local” attentional strategies. This has been demonstrated using Navon hierarchical stimuli, which consist of large shapes made up of smaller shapes (e.g., a large H made of small E's). East Asians typically perform better than Westerners when asked to find a target at the global level; the opposite is found when they are asked to find a target at the local level (McKone et al., 2010). Additionally, when making subjective decisions about the similarity between a target image and two other images—one sharing the same global shape and the other sharing the same local shapes—East Asians show a greater preference for the global shapes compared to Westerners (Caparos et al., 2012).
The tendency of East Asians to rely on more global attentional strategies than Westerners is congruent with the aforementioned theory of social exposure. However, some evidence argues against this theory (Caparos et al., 2012). The Himbas, a remote African culture, prioritizes collective well-being over individual well-being; they display a clear preference for local information—even more so than Westerners. The Himbas’ local preference decreases as a function of their exposure to an urban environment; perhaps one's exposition to a visually cluttered environment influences preferential attentional strategies (see also Miyamoto et al., 2006). This highlights the importance of testing experimental paradigms on samples from diverse cultural backgrounds.
Other lower-level mechanisms have been proposed to explain cultural differences observed in visual perception (Gutchess & Sekuler, 2019). Pinpointing the mechanism underlying cultural differences remains a critical aim, pivotal for understanding the factors shaping visual perception.
Cultural Variations in Face Processing
Faces are considered a special object category for several reasons. They are a rich source of social information, conveying details about identity, age, gender, race, and affective and mental states. Faces are also among the most frequently encountered stimuli for newborns (Sugden et al., 2014). Even adults spend about 12 min per waking hour viewing faces (Oruc et al., 2019). Due to their social importance and humans’ exposure to them, humans may develop perceptual expertise for faces (Gauthier & Nelson, 2001).
Since the configuration of faces encountered during development is similar worldwide, and given that face processing reflects visual expertise, face processing should be governed by universal mechanisms. However, cultural upbringing shapes face processing in ways similar to the previously described effects.
Face Identification Across Cultures
When processing a face, an observer's eyes fixate on different areas to extract information needed for recognition. This pattern of eye movements is highly systematic within an individual but differs across individuals (Mehoudar et al., 2014). When looking at the average eye fixations across individuals, an inverted triangular pattern emerges, revealing a high density of fixations on the eyes and mouth (Yarbus, 1965). This average pattern changes as a function of the cultural environment. East Asians fixate more on the center of a face and less on the eyes and mouth compared with Westerners (Blais et al., 2008). This cultural difference in fixations is congruent with East Asians’ broader deployment of attention just described. In fact, while fixating on the center of a face, East Asians still extract, like Westerners, the information contained in the eye and mouth areas, although more peripherally (Caldara et al., 2010). Moreover, East Asians rely on visual information of lower spatial resolution (i.e., coarser, less detailed information) than Westerners, as expected if they deploy their attention over a larger area (Tardif et al., 2017).
In vision sciences, spatial resolution is typically measured using spatial frequencies (SFs), which represent a measure of luminance changes within a specific unit of distance. For example, in a face, low SFs code coarse visual information, such as the face contour or the localization of the eyes, nose, and mouth. Higher SFs code finer visual information, such as the shape of the iris or nostrils, wrinkles, or eyelashes. SFs are considered to be low-level information, processed in the earliest steps of the visual stream (DeValois & DeValois, 1991). Top-down processes, occurring later in the visual stream, may also influence SF processing (Kauffmann et al., 2015). The finding that culture shapes SF processing in faces suggests that early steps of visual information extraction are influenced by one's environment. Cultural differences in SF use emerge as early as 34 ms after stimulus onset, indicating their influence on early stages of visual processing (Estéphan et al., 2018).
Facial Expressions Across Cultures
Cultural differences have also been observed in the way facial expressions of emotions are visually processed. As for face identification, eye fixation patterns displayed by East Asians and Westerners differ during the processing of facial expressions of emotions (Jack et al., 2009). More specifically, while Westerners display an inverted triangular fixation pattern with a similar density on the eye and mouth areas, East Asians mostly fixate on the eye area, with fewer fixations on the mouth. Cultural differences observed in the visual processing of facial expressions may be driven by differences in the way emotions are displayed (Jack et al., 2012). While East Asians expect most emotional signals to occur in the eye area, Westerners expect they will occur in both the eye and mouth areas (Jack et al., 2012). These cultural differences in the way emotions are signaled have been proposed to emerge following exposure to different social norms prescribing which emotions are appropriate to express in different contexts and the intensity with which they should be expressed (Ekman, 1971; Matsumoto et al., 2008a). Collectivistic cultures report being, on average, less expressive (i.e., to suppress emotional expression more) compared to individualistic cultures. Moreover, the more individualistic a culture is, the more they report expressing positive emotions (Matsumoto et al., 2008a). The socially accepted intensity with which emotions might be expressed also varies from culture to culture and depends on the specific emotion expressed (Matsumoto et al., 2008b). This variability has been theorized to be due to differences in beliefs and values regarding emotions (Tsai, 2007). While the Chinese value calm affects, Americans value higher arousal positive emotions, such as excitement. Consequently, faces displaying calm smiles are perceived more positively by Chinese than Americans, and vice versa for faces displaying excited smiles (Tsai et al., 2019).
Variability in socially accepted expressions may impact the ability of different cultures to discriminate subtle differences in affective states. When expressing pain, East Asian cultures display an all-or-none pattern of expression. More specifically, they report suppressing the expression until the pain is too intense (Tung & Li, 2015). Accordingly, Chinese participants expect pain facial expressions to be displayed with a higher intensity than Canadian participants (Saumure et al., 2023). Moreover, Chinese participants need more visual information to reach a similar accuracy than Canadians when they are asked to decide which of two faces expresses the most pain, suggesting that this task is more difficult for them (Saumure et al., 2023). These cultural differences in expectations about emotional expression intensity can significantly impact intercultural interactions and healthcare quality.
Studies examining cultural variations in face and emotion processing are still limited by a lack of sample diversity. Studies have almost exclusively compared Western and East Asian participants, and those that have expanded their cultural samples often rely solely on self-report measures.
Diversity in the Face Stimuli Used for Research
The field of face processing also suffers from a lack of diversity in the stimuli used (Dildine & Atlas, 2019 make a similar argument for pain facial expressions). To our knowledge, no study precisely quantifies the proportion of articles using only White faces, but the majority of studies likely fall into this category. Most studies including faces with other racial profiles typically aim at either studying cultural differences in face processing or the so-called other-race effect (ORE). This research reveals that the processing of other-race faces may involve different visual mechanisms, showing the importance of diversifying stimulus samples to better understand face processing in general.
ORE in Face Identification
The number and social significance of faces encountered make people experts in face processing. However, this expertise rarely extends to faces of other races. Most observers struggle to memorize and identify other-race faces—a phenomenon known as the ORE (Malpass & Kravitz, 1969). White participants have difficulty recognizing Black faces (e.g., Hugenberg et al., 2007), while Black participants struggle with White faces. Similarly, Asians perform poorly with White faces and vice versa (e.g., Blais et al., 2008). The frequency and quality of contact with other-race individuals are key factors in recognizing other-race faces (Hancock & Rhodes, 2008).
The ORE has real-world consequences in social and legal settings. In the US and Canada, non-White individuals are more likely to be wrongly accused due to misidentification (Jackiw et al., 2008). Despite extensive research and multiple theories, understanding of the ORE remains fragmented, with many questions about its perceptual, cognitive, social, and neurological origins.
ORE in Facial Expression Recognition
Similar to face identification, observers struggle to recognize facial expressions from individuals of another race or culture (Elfenbein & Ambady, 2002). This effect may in part be explained by cultural differences in the way facial expressions convey emotional signals. If the configuration of a facial expression or its intensity does not overlap well with someone's expectation of the facial expression that should be displayed in a given context, the recognition accuracy should decrease (Gosselin & Schyns, 2002).
Nevertheless, the ORE for facial expressions is also, in part, caused by the perceivers’ perceptual or cognitive processes. White and Black observers differ in detecting pain facial expressions displayed by Black individuals (Lin et al., 2024). Participants need a higher intensity of pain expressions to detect pain in Black avatars compared to White ones, despite identical identity and expression features, differing only in skin color (Mende-Siedlecki et al., 2019). This bias also affects treatment recommendations; Black individuals receive less medication for the same reported pain level compared to White participants.
Similar biases occur with other facial expressions. In a racialized anger bias, Black children are more often falsely perceived as angry compared to White peers (Halberstadt et al., 2022). Altered facial expression processing for other-race faces may contribute to racial inequalities in healthcare and society.
Policy Implications and Recommendations
Abundant evidence shows (1) a critical lack of diversity in psychology research samples but (2) considerable cultural variability in multiple psychological processes. This problem has several implications for the reliability and generalizability of research results. The reader is invited to consult also several key articles (e.g., Arnett, 2008; Gutchess & Rajaram, 2023; Henrich et al., 2010; Snibbe, 2003; Thalmayer et al., 2021; Thomas et al., 2023; Wang, 2016).
Research: Increase Diversity in Human Samples
First, research institutions, funding agencies, and editorial boards of scientific journals may encourage researchers to diversify their samples by asking them to (1) report the sociodemographic profile of their participants and (2) indicate their efforts to reach a diverse pool of participants.
However, given cultural variability, simply increasing diversity in samples may decrease the reliability of results if the data is averaged together. For instance, if spatial resolution with which faces are processed was tested using a sample of participants composed of both Chinese, relying on visual information of low spatial resolution, and Canadians, relying on visual information of high spatial resolution, the group average would appear as relying on information of middle range spatial resolution. This would neither represent Canadian nor Chinese tunings to visual information. Even more problematic, if the different spatial resolutions are linked with different attentional strategies, relying on this average tuning to understand visual information extraction will hinder our comprehension of the mechanisms at play. If the sample size is sufficient, running analyses using the race or cultural group of participants as covariates will help shortcut this issue.
Given that sample size is often insufficient for such analyses, a second recommendation is for formal institutions to request researchers, whenever possible and appropriate, to (1) make their data publicly available via open science platforms and (2) include information on the race and/or cultural group of each participant. In the long term, this will enable meta-analyses to verify the impact of these variables on the pattern of results.
Third, testing sufficiently large and diverse samples requires time and resources and is not always achievable. For instance, research groups relying on neuroimaging techniques may not always have access to the required equipment in multiple cultural sites. Most researchers are aware of this limitation and typically address it in their publications’ discussion sections. However, these remarks tend to be easily missed by readers, who assume that the results apply universally. To avoid this confusion, journals should ask researchers to specify the relevant demographic information in the most commonly read parts of the article, namely, the title and/or abstract. This will facilitate nuance, not only for researchers and readers, but also in meta-analyses, which will help behavioral sciences yield more reliable results.
Finally, research designed to understand culture's impact on psychological processes should be actively promoted. Doing research in cultural psychology comes with many challenges, often disadvantaging researchers who seek to reach a diverse sample of participants (Henrich et al., 2010; Snibbe, 2003). Reaching participants from diverse regions of the world and designing tools and procedures well adapted to diverse cultural backgrounds take time and expertise. Moreover, cultural psychology is still perceived by many as a parallel branch of psychology, making it harder to publish in some higher-impact journals (Wang, 2016).
Research: Increase Diversity of Researchers
Another challenge arising from increasing sample diversity is ensuring methodological equivalence in research. Methodological equivalence refers to the principle that research methods should be functionally comparable across different cultural settings. It ensures that the tools and procedures used in a study are understood and interpreted in the same way by participants, regardless of their background. This is crucial for maintaining the validity and reliability of research, as it helps to ensure that the outcome measured reflects the phenomenon being studied rather than discrepancies across participants in the way the methodology was understood (Heine, 2020).
Multiple steps help to respect methodological equivalence. One of them is to increase the diversity in the research task force (see Dupree & Boykin, 2021, for recommendations on that matter). A research team composed of individuals from diverse backgrounds will offer multiple points of view regarding methodology, theory, and conceptualization surrounding the research project. Moreover, if institutions, such as funding agencies, ask researchers to diversify their samples, they should also ask them to explain how they will adapt their research design. Some funding agencies have started to require researchers to indicate whether their sample will be diverse and to explain how they will achieve that diversity. However, they typically do not explicitly ask researchers to explain how they will adapt their research designs to ensure methodological equivalence, and the space allotted to that information is insufficient. Moreover, some researchers do not have any training in cultural psychology and might not be aware of the methodological equivalence challenge.
Education and Training: Courses in Cultural Psychology
Incorporating a cultural psychology course in the curricula of various programs in behavioral sciences may help future researchers understand the importance of relying on diverse samples, giving them the necessary competence to create well-adapted research designs. A better understanding of cultural differences in psychological functioning would also greatly enrich the curricula of various professions, including health, education, and law enforcement. Providing students in these fields with comprehensive training in cultural psychology could thus help reduce societal inequalities. Such a course on cultural psychology should however be extensive, for example a 12 to 15 weeks course, since it has been demonstrated that shorter diversity training may end up with negative outcomes (Devine et al., 2012; Dobbin & Kalev, 2018). Moreover, such a course should remain optional, as research has shown that diversity training is more efficient when it is voluntarily chosen (Dobbin & Kalev, 2018; Legault et al., 2007).
Education: Training on Cultural Differences in Face Processing for Professionals
Interventions and training to educate the population about racism have become mandatory in several institutions. However, some racial biases are only minimally impacted by racism. For example, the ORE explained earlier is not correlated with racism (Ferguson et al., 2001; see however, Trawiński et al., 2021). Similarly, the association between racism and the alteration at detecting pain facial expressions in Black individuals is either very small or nonexistent (Lin et al., 2024). Therefore, factors other than racism, such as visual perception mechanisms, may also lead to inequalities. Professionals across different spheres of society must be aware of these factors to address and mitigate their impact effectively.
The face is a social stimulus of high importance, and the way people interpret its cues shapes interactions. Face processing studies have critical implications for society. The pain of Black individuals may be underestimated in part because their facial expressions go undetected. An East Asian patient in the Western health system may not be treated adequately because they suppress their pain expression. Inversely, a Western patient in an East Asian health system may not be treated adequately because their pain expression is not intense enough to be detected. An East Asian child displaying a calm smile may be perceived less positively by their Western peers than if their smile were full-blown. A Black individual displaying a neutral expression may be misinterpreted as angry or threatening by others. And so on. This knowledge has implications for health, education, and law enforcement. The development of specific training modules on face processing would help raise awareness of cultural differences in face perception and the consequences of these differences on the interpretation of other people's behavior.
In the same vein, knowledge of cultural differences in face processing could inform engineers and developers of artificial intelligence (AI) tools for automated face or emotion recognition. Biases in AI models have been widely documented, and many originate from deficits in their training diets. For instance, as suggested by Dildine and Atlas (2019), should a biased AI model of pain recognition becomes widely used despite failing to recognize pain in other races or cultures accurately, the consequences could be disastrous. Encouraging diversity in research on face processing will help better understand cultural specificities in emotion expression and help prevent new biases from emerging in automated face processing tools.
Concluding Remarks
The accumulated evidence of cultural variability in psychological processes is compelling. Even visual perception processes, which are at the root of human contact with the outside world, are influenced by the environment in which one is born. Together, these findings argue against the common practice of relying on homogeneous samples of participants and generalizing conclusions to humanity as a whole. Policymakers can put various measures in place to help researchers modify their practices. Measures to support researchers in translating their findings to specific professional settings or the general population must also be implemented. Increasing awareness about findings with implications for the quality of intercultural interactions, such as those in the face processing literature, may help reduce inequalities. In doing so, researchers will nevertheless need to remain cautious about how to convey the information, as the potential for misinterpretation is real (Gutchess & Rajaram, 2023).
Acknowledgement
We would like to thank professor Susan T. Fiske for her insightful comments on the manuscript.
Footnotes
Declaration of Conflicting Interests
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/or publication of this article: This work was supported by a Canada Research Chair in Cognitive and Social Vision (grant number 950-232282).
