State of the Science: Introduction to Implicit Bias Review 2018-2020

          The racial and ethnic classifications utilized in Kirwan’s 2018-2020 State of the Science reports adhere to the guidelines established by the 1997 Office of Management and Budget (OMB, 1997). It is also important to acknowledge that approaches to defining racial or ethnic categories have shifted and changed over time. For example, the US Census has allowed respondents to mark more than one racial category in the 2000, 2010 and 2020 census efforts, which was previously forbidden. Our approach to reviewing the scholarship published from 2018-2020 utilizes the OMB and Census approaches that were applicable during this timeframe. As aggregators of the research, we cannot control the racial and ethnic classifications employed by the original authors. This means not all racial and ethnic groups are studied or mentioned in recent publications, which itself is a research finding that emerges from this effort. That said, we are committed to ensuring consistency and coherence in the terminology used throughout our reports to enhance reader comprehension and facilitate a more accessible reading experience.

American Indian or Alaskan Native. An individual who has origins in any of the indigenous peoples of North America and South America (including Central America) and maintains their cultural identity through tribal associations or attachment with their community. 

Asian. An individual with origins in any of the original populations of the Far East and Southeast Asia, or the Indian subcontinent including, for instance, Cambodia, China, India, Japan, Korea, Malaysia, Malaysia, Pakistan, the Philippine Islands, Thailand, and Vietnam.

Black. An individual with origins in any of the Black racial groups of Africa.

Hispanic or Latino. An individual with a cultural or ancestral background linked to Mexican, Puerto Rican, Cuban, Dominican, Central or South American, or other Spanish culture or origin, irrespective of their race.   

Native Hawaiian or Other Pacific Islander. An individual having origins in any of the indigenous peoples of Hawaii, Guam, Samoa, or other Pacific Islands.

White. An individual with origins connected to the original peoples of Europe, North Africa, or the Middle East.

          Implicit bias is defined as the set of attitudes or stereotypes that affect our cognition, affect and behavior in unconscious and automatic manner (Blair, 2002; Rudman, 2004). Implicit biases can be either positive or negative, and are activated involuntarily without an individual’s intention, awareness, or control. Implicit biases are pervasive; every individual is susceptible to them, including children and those with avowed commitments to impartiality (Nosek et al., 2007; Rutland et al., 2005; Greenwald et al., 1998; Kang et al., 2011).  

           Both nature and nurture contribute to how our implicit biases are formed. Renowned scholar Jerry Kang contends that, “even if nature provides the broad cognitive canvas, nurture paints the detailed pictures – regarding who is inside and outside, what attributes they have, and who counts as friend or foe” (Kang, 2012, p. 134). Indeed, our implicit biases are often developed as a result of mental associations formed by the direct and indirect messages we receive from our social environment, including those from parents, siblings, school environment, and the media. These messages are often about different groups of people, where certain identity groups are paired with certain characteristics (Nosek et al., 2007).  For example a long line of research has documented media programming excessively portrays Blacks as criminals (e.g., Dixon & Linz, 2000; Gilens 2000; Gilliam 2000). After constantly being exposed to such messages, we may begin to associate a group unconsciously and automatically with certain characteristics, regardless of whether these associations actually reflect the truth of reality (Hancock 2004; Mendelberg 2001; Winter 2008).  Implicit forms of bias also differ from more explicit forms, such as beliefs and preferences, which individuals are generally consciously aware of and can, if they choose to, identify and communicate to others (Daumeyer et al., 2019; Dovidio & Gaertner, 2010).  

          There are a few key types of mental associations that are frequently studied in conjunction with implicit bias. Infographic below illustrates both how they relate to each other and how they relate to implicit bias. Where it’s helpful we also provide some examples of each concept. 

Note. From Gawronski, B., De Houwer, J., & Sherman, J. W. (2020). Twenty-five years research using implicit measures. Social Cognition, 38(Supplement), 1-25. https://doi.org/10.1521/soco.2020.38.supp.s1 

          In general, many researchers have concluded that implicit and explicit measures capture distinct but related mental constructs (Nosek et al., 2007). The primary difference between implicit bias and other forms of bias revolves around the degree of consciousness. Explicit biases can be consciously registered and reported (Amodio & Mendoza, 2010). Implicit biases, however, arise outside of one’s conscious awareness and are not always congruent with an individual’s consciously held beliefs or attitudes (Graham & Lowery, 2004; Kang et al., 2012; Reskin, 2005). They may not necessarily reflect stances that one would explicitly endorse. While implicit and explicit bias are distinct concepts, they share similar impact: both mental constructs affect one’s actions, decisions and understanding.   

          In attempts to understand the differences between implicit and explicit bias, a number of other moderating factors have been identified to explain the divergent results captured by implicit and explicit measures (Rudman, 2004), including motivation to report explicit attitudes that correspond to one’s implicit attitudes (Dunton & Fazio, 1997; Fazio et al., 1995; Nier, 2005), the psychometric properties of the specific measurement techniques (Cunningham et al., 2001; Greenwald et al., 2003) and social desirability concerns (Dunton & Fazio, 1997; Nier, 2005; Nosek & Banaji, 2002). 

            Rudman's (2004) work elucidates five factors that are more likely to influence the development of implicit biases than explicit ones. The first factor pertains to early experiences that may lay the foundation for our implicit attitudes, whereas explicit attitudes are influenced more by recent events. Furthermore, affective experiences tend to have more impact on implicit orientations than self-reports. She also listed cultural biases and balance principles as key factors that shape implicit biases. While the above-mentioned four factors were already established in the literature, Rudman adds a fifth factor for consideration, namely the self. According to Rudman (2004), it may be challenging to have implicit associations that are detached from the self, whereas controlled evaluations may enable more objective responses. If the self is a primary cause of implicit orientations, it is likely because individuals tend not to view themselves impartially, and this partisanship can then shape evaluations of other objects that are (or are not) related to the self. 

         The difficulties associated with measuring implicit bias arise in part due to its automatic and unconscious nature. Conventional self-report measures of bias rely on introspective and conscious reports regarding one's experiences and beliefs, which are indirect methods for assessing implicit bias. Additionally, impression management, or social desirability concerns, can undermine the validity of self-report measures. It is of particular concern when the target of measure is related to bias, as the desire to be perceived positively can lead people to alter their self-reported beliefs and attitudes to appear more impartial (Amodio & Devine, 2009; Dovidio  et al., 1997; Fazio et al., 1995; Greenwald & Nosek, 2001; Greenwald et al., 2009; Nier, 2005; Nosek et al., 2007). This inclination towards impression management can compromise the validity of self-reports, especially when they are questioned on socially or politically sensitive issues such as intergroup or interracial behaviors (Dovidio et al., 2009; Greenwald & Nosek, 2001; Greenwald et al., 2009). Due to these limitations, self-reports are generally considered insufficient for capturing all dimensions of individual prejudice (Tinkler, 2012). 

The Evaluative Priming Task (EPT)  

          The development of the EPT was guided by the notion that attitudes are represented in memory as associations of varying strength for object evaluation (Fazio, 2007). This view suggests that encountering an attitude object may automatically activate its associated evaluation via the spread of activation to the extent that the association between the two is sufficiently robust. For example, Fazio et al. (1995) conducted a study that explored the effects of priming participants with photographs of black and white undergraduates. During the study, participants were presented with an evaluative adjective, such as "pleasant" or "awful," and their primary task was to indicate the connotation of the adjective as quickly as possible. A photo was briefly presented prior to each target adjective, with participants instructed to attend to the faces for later identification. The study was presented to participants as a cover story in which judgment of word meaning was portrayed as an automatic skill that should not be disturbed by the additional task. The results of this study indicated that Black and White faces had different effects on participants' response time in indicating the connotation of the target adjective. Compared to white faces, black faces facilitated the response to negative adjectives and disrupted the response to positive adjectives. This pattern suggested that, on average, negativity was automatically activated by the black primes (Fazio& Olson, 2003). In this vein, the EPT was seen as an "unobtrusive" measure of attitudes, capturing unintended expressions of attitudes that are difficult to regulate, or that individuals may be unwilling to reveal on explicit self-report measures (Gawronski et al., 2020).  

The Implicit Association Task (IAT)  

          Unlike the focus on unintentional expressions of attitudes that are challenging to regulate, the creation of the IAT was guided by studies on implicit memory. This approach proposed that memories of past experiences could impact responses even when they were inaccessible to introspection (Gawronski et al., 2020). In practice, the IAT gauges the degree of association between a target concept and an attribute dimension by measuring the time it takes for participants to utilize two response keys, each with a dual meaning. The participants' task is to classify stimuli as they appear on the screen, such as categorizing names (e.g., "Latonya" or "Betsy") as typical of blacks or whites. For example, race is the target concept, and the keys are labeled "black" and "white" in Greenwald et al.’s (1998) IAT on racial attitudes. Participants then categorize a range of words with clearly valenced meanings (e.g., "poison" or "gift") as pleasant or unpleasant, which constitutes the attribute dimension. 

        During the critical phase of the study, the two categorization tasks are combined. Participants perform this combined task twice - once with one response key signifying Black/pleasant and the other labeled White/unpleasant, and once with one key indicating Black/unpleasant and the other White/pleasant - in a counterbalanced order. The main question is which response mapping participants find easier to use. In the Greenwald et al. (1998) study, participants were much quicker at responding when Black was paired with unpleasant than when Black was paired with pleasant. On average, participants found it easier to associate the target concept Black with the attribute unpleasant than with the attribute pleasant, which can be interpreted as indication of implicit racial bias.  

            Over the past decades, at least 20 different instruments (see Table 1) have been developed to measure implicit bias. Among these instruments, the IAT (#11) and the Single Category Implicit Association Test (SC-IAT; #19) are the most common measures used for the assessment of implicit biases and self-concepts. Most recently the IAT was found to have a higher accuracy of predicting behavioral choices over the SC-IAT, regardless of the scoring procedure (Epifania et al., 2020). 

          The target of negative implicit racial bias is not limited to Blacks. Two IAT instruments were designed to measure the implicit associations of ethnic-national groups. Specifically, they focus on determining whether Asian Americans or native Americans are more easily associated with the attribute of Foreign versus White Americans with American. Nosek and colleagues (2007) found that, on average, participants demonstrated a higher ease in associating White faces with the concept of “American,” and Asian or Native American faces with the notion of “Foreign,” rather than the reverse. This association, referred to as the Asian stereotype, was observed in 61% of the sample, while the Native American stereotype was observed in 57% of the sample.

             When using the IAT instrument, Nosek and colleagues (2007) found that this measure detected stronger bias in people than the self-report measures. Specifically, they used a statistic called “Cohen’s d” to measure how far from a “neutral” standpoint people’s racial attitudes were in their nationally representative sample. While explicit self-report measures showed that people only had a small to medium level of preference for White (d = 0.31), the race attitude IAT for the same respondents revealed a much higher level of implicit preference for White (d = 0.86, more than twice the self-report score). If we were to translate these findings into a “normal” statistical distribution, we would predict that as many as 54.4% of individuals may explicitly admit to having a preference for White individuals. However, Nosek and colleagues’ empirical findings document a substantially larger proportion, 74.5%, who actually hold this preference for White individuals at an implicit level. 

          Research on racial attitudes conducted by Fazio and colleagues has identified three distinct types of individuals, characterized by the automatic processes they engage in and their subsequent efforts to counter or control those evaluations (Fazio et al., 1995). The first group consists of individuals who are non-prejudiced; these individuals do not exhibit the activation of negative attitudes towards a particular racial group. The second group encompasses individuals who are genuinely prejudiced, as they experience negative automatic associations but make no attempts to counter or control the expression of those associations. The third group comprises individuals who may experience negative automatic associations, but they are motivated to actively counter such sentiments. Individuals in this last group, who seek to mitigate the role of implicit racial biases in their overt actions or behaviors can pursue intentional actions at either the personal (internal) or the situational (external) level.  Each type of influence has been documented to shape the conditions under which implicit racial biases are likely to affect behavior. 

Personal Factor 1: The Role of Motivation 

            Motivated individuals have the capacity to exert influence over the automatic operation of stereotypes and biases (Blair, 2002). Dunton and Fazio (1997) investigated the influence of motivation on controlling biased reactions and identified two sources of motivation: a concern with appearing prejudiced and a desire to avoid conflicts or disagreements about one's thoughts or positions. Their findings suggested that individuals with low motivation to control biased reactions were more likely to provide accurate self-reports of racial attitudes. Devine et al. (2002) further explored the impact of motivation on implicit and explicit biases, concluding that explicit biases were moderated by internal motivation, while implicit biases were influenced by the interaction of both internal and external motivation. 

          Devine et al.’s finding is consonant with that of Irene Blair. Blair (2002) identified two individual (internal) level interventions and one situational (external) intervention that can potentially mitigate implicit bias. The internal factors involve examining the manipulations of the perceiver's motivations, goals, and strategies during the process in which implicit bias arises. On the other hand, the external factor identified by Blair focuses on changes in the context surrounding the stimulus or variations in the attributes of group members. 

          Individuals who possess high motivation have the ability to influence the automatic activation of stereotypes and bias (Blair, 2002). When an individual’s self-image is threatened, they may automatically activate negative stereotypes in order to enhance their own appearance or discredit someone they dislike (Sinclair & Kunda, 1999; Spencer et al., 1998). Conversely, they can also automatically suppress negative stereotypes and activate positive ones when it benefits their self-image (Sinclair & Kunda, 1999). Automatic stereotypes and bias can be influenced by the social demands of a situation and the nature of one's relationship with others. For example, White individuals may moderate their automatic bias when interacting with a Black person, particularly if they are in a subordinate position (Lowery et al., 2001; Richeson & Ambady, 2001). They may also adjust their automatic stereotypes if those stereotypes conflict with social norms (Sechrist et al., 2001). 

          The concept of overcorrection, where individuals compensate for their biases by displaying exaggeratedly positive or negative responses that conceal their true attitudes, has also been investigated. Dunton and Fazio (1997) suggested that highly motivated individuals who aim to control prejudiced reactions might overcompensate for their automatically activated negativity, resulting in more positive judgments than those with activated positivity. 

Personal Factor 2: Promotion of Counter Stereotypes         

          It is argued that individuals who perceive stereotypes have the ability to actively foster counter-stereotypic associations. By doing so, they can potentially undermine the influence of implicit stereotypes in the process of information processing. Promotion of counter stereotypes can be achieved through tasks such as priming of counterstereotype expectancy, mental imagery, and exposure to counterstereotypic group members (Blair & Banaji, 1996; Blair et al., 2001; Dasgupta & Greenwald, 2001). To examine the effect of counterstereotype expectancy on implicit bias, Irene Blair and Mahzarin Banaji (1996) conducted a sequential priming task, which assesses automatic associations by measuring how a briefly presented stimulus (prime) influences the response time to a subsequent stimulus (target). Half of the participants were instructed to anticipate prime-target trials that aligned with stereotypes, while the other half were informed to expect counterstereotypic trials. However, in reality, all participants experienced both stereotypic and counterstereotypic trials, with the anticipated trial type appearing only 63% of the time. The results of this study revealed that the counterstereotype expectancy significantly lowered levels of automatic stereotypes than the stereotype expectancy. 

          The effect of mental imagery was examined through tests targeting implicit gender stereotypes. Blair and colleagues (2001) instructed the participants to spend around 5 minutes to create a mental image of a strong woman, which is counterstereotypic to the traditional stereotype, and then complete an automatic gender stereotype test. Participants who engaged in the counterstereotypic mental imagery exhibited substantially weaker automatic stereotypes compared to participants in the control groups, which included those who attempted to suppress their stereotypes on their own.  

            The exposure-to-counterstereotypic-group-members mitigation strategy takes a somewhat different approach. A group of participants were exposed to either a) admired African Americans and disliked White Americans, b) disliked African Americans and admired White Americans, or c) nonracial control stimuli. Following the exposure, the participants took IAT automatic racial bias measures both immediately and 24 hours later (Dasgupta & Greenwald, 2001). The study found that exposure to positive Black group members reduced automatic bias against Blacks compared to exposure to negative group members or to control stimuli. It is also worth noting that this effect continued to be significant after 24 hours. While Dasgupta and Greenwald (2001) believe that this effect is likely to diminish after an extended hiatus, these findings remain significant as they challenge the assumption that automatic racial attitudes, ingrained through long-term socialization, are unchangeable.    

         One potential concern the researchers emphasized about their results regards the possibility that when individuals are exposed to strongly positive and negative exemplars of a particular stereotype, they might categorize these exemplars as “exceptions to the rule” and adjust their attitudinal response accordingly before expressing it. This observation aligns with studies on stereotypic belief revisions, which demonstrate that when confronted with evidence, explicitly held stereotypes can remain firm if the new evidence or exemplars are perceived as atypical (Bodenhausen et al., 1995; Kunda & Oleson, 1995; Rothbart & John, 1985; Weber & Crocker, 1983). As such, instead of revising the original stereotype, new cognitive categories are created to accommodate these counterstereotypic exemplars, a phenomenon known as subtyping. However, Dasgupta and Greenwald (2001) contend that this effect is less likely to manifest when individuals do not have ample cognitive resources to reflect on and recategorize these counterstereotypic exemplars. This perspective is supported by the findings of Dasgupta and Greenwald’s study, where exposure to counterstereotypes reduced implicit, but not explicit, stereotypes against Black individuals. The researchers therefore cautioned that care must be taken to ensure that this type of exposure doesn't label the counterstereotypic exemplars as mere exceptions, which would inadvertently perpetuate the original stereotype. 

Personal Factor 3: Group Membership and Identity 

           Perceptions of group membership and identity are another personal factor that can influence one’s implicit bias. Categories such as race, gender, and age can be perceived almost instantaneously (Cikara & Van Bavel, 2014; Ito & Bartholow, 2009). Ingroups, whether they exist in real life or are artificially created for an experiment, can result in attitudes and behaviors of favoritism (Pinter & Greenwald, 2011; vanKnippenberg & Amodio, 2014). Simon and Gutsell (2020) interrogated this phenomenon utilizing the minimal group paradigm developed by Tajfel, Billig, Bundy and Flament (1971). This paradigm allows researchers to create experimental social categories that participants have never encountered before, enabling control over individual experiences, motives, and opinions. Participants in this study were evenly assigned to two minimal groups with arbitrarily created racial names.  Researchers found that knowledge of arbitrary group membership led participants to humanize and show implicit bias favoring their ingroup members of the minimal group, despite the cross-cutting presence of race. This finding demonstrates that even within minimal groups, the consequences of implicit racial bias can be overcome through recategorization.  

          To further comprehend the underlying dynamics of this effect, Fritzlen and colleagues (2020) conducted two studies exploring the boundary conditions of the reclassification effort. They first examined whether learning of an identity-implying similarity with an outgroup could influence once’s implicit bias towards that outgroup.  One of the identity-implying similarities they used in the study was genetic overlap with an outgroup. Their first study found that White participants who were made to believe that they had greater than average genetic overlap with African Americans showed decreases in implicit bias against that racial group. In a follow-up study, the researchers explored whether this effect holds in the context of sexual orientation identity. In this study, heterosexual participants were led to believe that they exhibited same-sex attraction and their implicit bias against homosexual individuals were measured. The findings showed that heterosexual participants’ implicit biases against homosexual individuals were reduced only if respondents believed that sexual orientation was biologically determined. Thus, Fritzlen and colleagues (2020) concluded that learning of an identity-implying similarity with an outgroup can reduce implicit bias if that group membership is perceived as immutable.  

           These findings hint at a broader implication that aligns with the intergroup contact theory, which suggests that direct contact between members of different social or cultural groups can be effective in reducing prejudice and conflicts between groups (e.g., Allport & Kramer, 1946; Lee & Humphrey, 1968; Pettigrew & Tropp, 2006). For instance, Kephart (1957) found that White police officers in who had worked with Black colleagues differed substantially from other White policemen on the force. They had fewer objections to teaming with a Black partner and taking orders from qualified Black officers. Combined with the evidence for minimal group reclassification effect on implicit bias (Frizlen et al., 2020), this suggests that enhancing awareness of shared group membership or identities, even arbitrary ones, could serve as a strategic approach to fostering positive intergroup relations and reducing implicit biases. 

            Furthermore, having positive intergroup contact has been shown to have potential positive influence over implicit bias as well in both virtual and real-life settings. Early behavioral studies found that people are better at recognizing faces from their own race compared to faces from other racial targets (Meissner & Brigham, 2001), a phenomenon dubbed as “the own race bias.”  Following this line of research, Farmer and colleagues (2020) investigated whether positive intergroup contact can moderate the implicit bias in face perception using fMRI. In their study, White participants with a varied range of contact experiences with Blacks were first selected through a pre-task questionnaire. During the fMRI scan the selected participants completed an individuation task (favorite vegetables) and a social categorization task (above or under 25 years old) while viewing the faces of both Black and White strangers. The fMRI results showed that more frequent positive contact with Black individuals increased activation for Black faces in the social categorization task. This finding shows that quality of intergroup contact is also a key factor in processing out-group faces and reducing the own-race face bias. Improving the quality of intergroup contact is something individuals who are motivated to mitigate their biases have control over in their personal lives.   

Contextual Factor 1: Configuration of the Cues 

           One important area for mitigation efforts focuses on how individuals interpret cues provided to them. Two types of studies have ensued: impression formation and perspective taking. In the classic impression formation studies conducted by Asch (1946), participants were given the task of forming impressions of an individual based on a set of traits such as skillfulness, industriousness, and intelligence. The presence of the trait "warm" or "cold" on the list had a significant impact on the impressions formed by the participants, leading to notable differences in their evaluations. Researchers argued that such fluidity in meaning may also be observed in automatic processes such as implicit stereotypes and biases (Blair, 2002). For instance, Macrae et al. (1995) conducted a study that highlighted how even a slight contextual change can exert a notable influence on automatic stereotypes. In their experiments, all participants were exposed to a Chinese woman, and subsequent measurements of automatic stereotypes toward both Chinese individuals and women were taken using a lexical decision task. In one condition, the Chinese woman was observed applying makeup, while in another condition, she was seen using chopsticks. As hypothesized, Macrae et al. discovered that participants who witnessed the woman applying makeup exhibited faster responses to traits stereotypically associated with women and slower responses to traits stereotypically associated with Chinese individuals. Conversely, participants who observed her using chopsticks displayed the opposite pattern of responses. Despite the stimulus remaining the same, a minor alteration in the context resulted in a substantial change in the automatic stereotypes evoked by her presence. Similarly, the automatic attitudes triggered by the presence of the same Black individual can vary depending on the setting (Wittenbrink et al., 2001). For example, when a person is encountered on a city street, different automatic attitudes may emerge compared to when they are encountered inside a church. In general, city streets are often associated with more negative perceptions compared to the interior of a church, and similarly, Black faces tend to be perceived more negatively than White faces. It is important to note, however, that these qualities do not independently impact the automatic process. Instead, their effects are multiplicative, meaning that when a Black face is encountered in a city street context, the automatic negativity becomes particularly strong.  

            An individual’s capacity to take perspectives and to show empathic concerns are also significant predictors of their implicit bias (Patané et al., 2020). Researchers have theorized that bilingualism may enhance perspective-taking at early developmental stages, thus reducing the expression of social bias.  Being raised with more than one native language is often considered an advantage as it diversifies the range of linguistic, social, and cultural experiences available to an individual (Singh et al., 2020).  Singh and colleagues compared monolingual and bilingual children to examine whether bilingual individuals are as susceptible to racial bias as their monolingual counterparts. Results showed that bilingual individuals displayed less implicit bias against Blacks.  Additionally, it is possible that bilingualism predisposes individuals to more frequent social encounters with outgroups (something we discussed in association with  personal factor #2), which may also reduce implicit bias against members of an outgroup.   

           Kang and Falk (2020) explored perspective-taking via an fMRI study to investigate the relationship between individuals’ tendencies to engage brain regions that support perspective-taking and their implicit biases towards a stigmatized group. Their findings revealed that an intervention task that exercised one’s ability to make inferences about others’ mental states was able to reduce implicit bias. Therefore, the researchers argued that intervention strategies designed to enhance social cognitive processing could be effective in reducing people’s implicit biases against stigmatized individuals. Patané and colleagues explored a related perspective-taking question using immersive virtual reality (VR) technology to explore whether interventions tailored to increase perpective-taking can help to reduce implicit racial bias and the use of VR technology to design interventions to reduce implicit racial bias. The researchers had a group of White study participants embody a Black virtual body. They compared those who engaged in a cooperative activity with another Black avatar to those who engaged in a neutral activity. The results revealed that those who engaged in a cooperative activity with a Black avatar had lower IAT scores after the social interaction than those who engaged in a neutral activity with a Black avatar. 

           Collectively, these findings indicate that automatic perception relies on the configuration of various stimulus cues, and even small changes can lead to remarkably different outcomes. It highlights the sensitivity of automatic processes to contextual factors and the potential for subtle modifications to have a profound impact on perception (Blair, 2002). 

Contextual Factor #2: Time 

           Time is a recognized factor that plays a role in decision-making and the ability of individuals to regulate their reactions. Studies have indicated that time pressures can create conditions in which implicit attitudes become evident, even when individuals make explicit efforts to control them (Kruglanski & Freund, 1983; Payne, 2006; Sanbonmatsu & Fazio, 1990). This suggests that under time constraints, implicit attitudes may manifest themselves despite individuals' explicit attempts to exert control over them (Bertrand et al., 2005). This phenomenon could have important implications for professionals such as police officers and other authorities who must make split-second decisions in their line of work. For some of these professionals who already harbor implicit biases, making quick judgment may not necessarily alter their biases but rather make these biases evident through behavioral manifestations. Highlighting this, Payne (2005) conducted a study to examine participants' implicit associations related to weapons and race. The findings revealed that individuals with more negative implicit attitudes towards Blacks exhibited a stronger bias associating weapon with them. In a subsequent analysis, Payne (2006) emphasized that a crucial component of intentional response control gets overshadowed when the focus is solely on automatic bias.  

           Rapid judgments do not alter individuals' biases but instead allow these biases to manifest as overt behavioral errors when time is limited. For example, Gallo et al. (2013) found that the time pressure on a White referee before making a decision to punish a football player affected their level of discrimination against players who are foreign and non-White, signaling an increase in implicit racial bias. Time pressure also shapes medical decisions, potentially to the detriment of racial minority patients. Stepanikova (2012) found that when family physicians and general internists are under high time pressure, their implicit bias against Black and Hispanic patients results in less serious diagnoses for both groups and a reduced likelihood of specialist referrals for Black patients.  

Contextual Factor #3: Cognitive “Busyness” 

           According to Reskin (2005), any factors that tax our attention, such as multiple demands, complex tasks, or time pressures, increase the likelihood of us resorting to stereotyping. Gilbert and Hixon (1991) investigated the influence of cognitive busyness on the activation and application of stereotypes. Their findings showed that cognitive busyness reduces the likelihood of stereotype activation. However, if a stereotype is activated, cognitive busyness increases the probability of applying that stereotype to the individual(s) in question. Payne (2006) conducted an experimental study to examine the relationship between bias and cognitive overload, a state in which an individual's working memory is overwhelmed by more information or tasks than it can effectively handle at a given moment. Payne found that cognitive overload leads to increased bias, which he attributed to individuals’ diminished control over their responses due to lack of cognitive resources. Furthermore, Bertrand et al. (2005) identified three conditions conducive to the emergence of implicit bias, including lack of attention to a task, time constraints or cognitive overload and ambiguity. In the health care setting, cognitive load may stem from stressful mental tasks, such as decision-making for multiple patients, as well as environmental factors such as overcrowding. Supporting the perspective put forward by Bertrand et al., Johnson and colleagues (2016) found that overcrowding and patient load indeed contributed to increases in physicians’ implicit racial bias. Additionally, despite unfounded concerns about non-White individuals being more likely to abuse opioids, Burgess et al. (2014) reported that male physicians were less than half as likely to prescribe opioids for pain management to Black patients compared to White patients when under high cognitive load. Similar evidence was found in the field of sports as well. The ambiguity of situations regarding infringements in football was found to affect the referee’s level of discrimination, which strongly points to the influence of cognitive busyness on the presence of implicit bias (Gallo et al., 2013).