Skip to content
Frontiers
Search

Original Research

Frontiers in Social Psychology ·

Volume 4 - 2026 | https://doi.org/10.3389/frsps.2026.1678079

The effects of skin tone and prototypicality on the dynamic racial categorization and stereotyping of Latine Americans

  • Gabriel Camacho1*corresponding author
  • Mora A. Reinka2
  • Cassidy J. Burt3
  1. 1Department of Psychology, John Jay College of Criminal Justice, New York, NY, United States
  2. 2Department of Psychology, Ursinus College, Collegeville, PA, United States
  3. 3Independent Scholar, Queens, NY, United States
Article metrics0 viewsView details
PDFCiteMetrics

Abstract

Latine people are phenotypically diverse and encompass a wide range of skin tones; however, psychological research on person perception has devoted comparatively little attention to the social-cognitive processes through which skin tone and other phenotypic cues shape the racial categorization and stereotyping of Latine individuals. The present work uses computer mouse tracking to explore how skin tone and Latine prototypicality—the degree to which an individual's physical appearance aligns with stereotypical features associated with Latine identity—affect the dynamic processes of racial categorization and stereotyping of Latine Americans. We examine how Latine individuals who appear typical (medium skin tone with prototypical features), atypical (light skin with non-prototypical features), or ambiguous (light skin with prototypical features) affect categorizations as “White American” or “Latino American” (Study 1; N = 139), and as “American” or “Foreign” (Study 2; N = 206). Results revealed that both typical and ambiguous Latine targets were more likely to be categorized as Latino and foreign, compared to atypical targets. Skin tone and prototypicality both influenced the mouse trajectories leading to these judgments. Trajectories for typical and atypical targets were fast and direct, suggesting low cognitive conflict, whereas ambiguous targets elicited more complex and indirect movements, indicating greater indecision. These findings demonstrate that both skin tone and overall Latine prototypicality affect the dynamic processes through which the largest racial minority group in the U.S.—Latine Americans—is racially and stereotypically categorized.

Keywords

Introduction

Latine people are phenotypically diverse and possess a wide range of skin tones, which have been shown to influence both their lived experiences and the ways in which they self-identify (Golash-Boza and Darity, 2008; Irizarry et al., 2022; Telles, 2014). Yet psychological research on person perception has given comparatively little attention to the social-cognitive processes through which skin tone and other phenotypic cues shape the ways in which Latine people are racially categorized and stereotyped (Harris et al., 2020). The limited studies that do exist often rely on facial-database methods, in which Latine prototypicality—the degree to which an individual's physical appearance aligns with stereotypical features associated with Latine identity—is explicitly rated by samples of online participants (Chen et al., 2021; Ma et al., 2015). Other research examining the cognitive processes underlying stereotyping often uses implicit measures, such as the Implicit Association Test, which evaluates the strength of automatic associations between different social groups (e.g., Latine targets) and positive or negative attributes (March and Graham, 2015; Uhlmann et al., 2002). While informative, these approaches primarily capture the outcomes of racial categorization and stereotyping, offering limited insight into the dynamic, real-time processes through which such judgments unfold. Understanding these moment-to-moment processes is important because they reveal how phenotypic cues begin to influence racial judgments long before a final categorization is made. In the present work, we address this gap by using computer mouse tracking to examine how individuals continuously and dynamically categorize Latine faces as a function of their skin tone and overall Latine prototypicality.

Mouse tracking is a widely used method for assessing continuous and dynamic cognitive processes by analyzing computer mouse trajectories during binary choice tasks. In a typical task, participants click a “Start” button at the bottom center of the screen, after which an image appears (e.g., a Black man; Freeman et al., 2010). They are then asked to quickly categorize the image by selecting one of two categories displayed at the top left (“White”) or top right (“Black”) corners of the screen. The temporal movements of the mouse along the x and y axes are analyzed to assess decision-making behavior and cognitive processes, with the mouse's movements theorized to serve as a valid index of participants' decision-making processes (Freeman, 2018). Research has shown that while typical exemplars of a category tend to be categorized directly into the appropriate group (e.g., a prototypical Black target categorized as “Black”), ambiguous exemplars (e.g., a biracial Black–White target) often initially diverge toward the “White” category before ultimately being categorized as “Black.” This divergence is significant because it highlights the complexity of decision-making, involving both lower-level sensory information (physical cues) and higher-order social cognition (stereotypical associations) processes (Freeman and Ambady, 2011; Freeman et al., 2020). Although mouse tracking has been extensively used to document racial categorization and stereotyping (e.g., Hehman et al., 2014; Johnson et al., 2012; March, 2022; Stolier and Freeman, 2017; Wojnowicz et al., 2009), no study to our knowledge has yet applied this method to explore perceptions of Latine individuals.

One potential reason that the faces of Latine Americans have been underrepresented as stimuli in mouse tracking studies is the difficulty in distinguishing typical from atypical Latine exemplars. Latine Americans are often viewed as racially ambiguous (Chen et al., 2021), with few facial features that reliably set them apart (Ma et al., 2018). However, they are still phenotypically identifiable (Wilkins et al., 2010). Indeed, medium skin tone is the strongest predictor of Latine prototypicality on both implicit and explicit measures (Feliciano, 2016; Norman et al., 2024; Camacho and Reinka, 2026). As a result, in a mouse tracking paradigm, light-skinned Latine faces may be miscategorized as White (atypical target), whereas those with medium skin tones are more likely to be correctly categorized as Latine (typical target). Moreover, because skin tone plays a critical role in racial categorization (Stepanova and Strube, 2012), particularly during early-stage processes like those assessed in computer mouse tracking studies (Freeman and Ambady, 2011), it is likely that these mouse trajectories will show relatively quick reaction times and minimal deviation from an idealized straight line. In other words, medium skin tone represents the prototypical Latine exemplar and will likely function as such in a mouse tracking paradigm in which Latine faces are shown.

The complexity of racial categorization is likely to be reflected in participants' mouse trajectories when they encounter ambiguous targets displaying physical traits associated with both categories. Latine individuals, for example, who have the light skin tone typical of White Americans but also display physical or cultural traits associated with Latine Americans, may be more affected by this because it activates multiple competing racial categories at once. A review of Latine prototypicality scores in widely used face databases reveals that it is not uncommon for light-skinned individuals to be rated as prototypically Latine (Chen et al., 2021; Ma et al., 2015). This is likely due to other features in the images such as lip shape (Feliciano, 2016), eye color (Ma et al., 2018), and hairstyle (MacLin and Malpass, 2001), which contribute to an overall prototypical Latine appearance. While skin tone may be a primary cue in the racial categorization of Latine Americans (Norman et al., 2024), the overall prototypicality of a Latine target is also likely to influence categorization. Distinguishing between light-skinned, prototypical, and non-prototypical Latine targets can be complex. Therefore, we predict that mouse trajectories for ambiguous exemplars (light-skinned, prototypical Latine targets) will exhibit greater deviations in mouse trajectories, reflecting the simultaneous activation of competing racial categories: White American and Latine American.

The skin tone and overall prototypicality of Latine Americans should also influence decision-making processes tied to perceived nationality: American vs. foreign. Specifically, Latine Americans are often perceived as perpetual foreigners (Camacho and Abouras, 2025; Cheryan and Monin, 2005; Devos et al., 2010), facing stereotypes that depict them as undocumented immigrants, poor English speakers, and unassimilated into American society (Zou and Cheryan, 2017). These stereotypes of foreignness are more frequently applied to Latine Americans with darker skin tones compared to those with lighter skin tones (Wilton et al., 2013). However, according to the theory of racial phenotypicality bias (Maddox, 2004), individuals who appear more prototypical of a racial group are more likely to be perceived through the lens of group-based stereotypes. This suggests that both ambiguous targets (light-skinned/prototypical Latine Americans) and typical targets (medium-skinned Latine Americans) could be categorized as foreigners, as they both exhibit prototypical Latine traits, albeit in different ways. Nonetheless, while both groups would likely be categorized as foreign rather than American, the cognitive processes leading to this categorization may differ. Specifically, light-skinned/prototypical Latine individuals may generate confusion due to their shared traits with both White and Latine individuals. This could result in greater variability in their categorization—moving between being seen as American or foreign—while medium-skinned Latine individuals would likely be more readily and directly categorized with stereotypical associations of foreignness. We test this possibility in the current research.

Present research

Previous studies have shed light on how skin tone and overall prototypicality influence both explicit and implicit forms of racial categorization and stereotyping of Latine Americans (Chen et al., 2021; Camacho and Reinka, 2026; Uhlmann et al., 2002). However, there is a lack of research exploring the continuous and dynamic processes involved in these categorizations, particularly as they apply to Latine targets. The present work addresses this gap by examining the temporal mouse trajectories involved in decisions about whether Latine targets are categorized as “White American” or “Latino American” (Study 1) and as “American” or “Foreign” (Study 2).1

In each study, Latine targets varied in prototypicality: typical of the Latine category (medium skin tone and prototypical), atypical (light skin and non-prototypical), or ambiguous (light skin but prototypical). We hypothesize that typical Latine targets (medium skin tone and high prototypicality) will be quickly and directly associated with being Latine American rather than White American (Study 1) and with being Foreign rather than American (Study 2). We predict the opposite pattern for atypical targets (light skin tone and low prototypicality). We further hypothesize that ambiguous targets (light skin tone but high prototypicality) will be categorized as “Latino American” (Study 1) and “Foreign” (Study 2) but will elicit the greatest categorization conflict, resulting in greater deviations in mouse trajectories across both studies.

Given the well-documented racial ambiguity of Latine faces and evidence that few specific facial features reliably predict Latine prototypicality (Ma et al., 2018), we selected faces using both objective measures of skin luminance (lighter vs. darker) and previously established Latine prototypicality ratings (Chen et al., 2021; Ma et al., 2015). This approach allows us to examine how these perceptual dimensions shape categorization, with particular attention to light-skinned Latine faces. We test the hypothesis that skin tone plays a primary role in rapid judgments of Latine group membership, while also anticipating that even prototypically appearing light-skinned Latine individuals will still be categorized as Latine and stereotyped as “foreign,” though the cognitive pathways to these judgments may be less direct. Mouse tracking enables us to capture both explicit categorization decisions (“White American” vs. “Latino American” and “American” vs. “Foreign”) and the implicit cognitive dynamics (mouse trajectories) underlying them. Together, these studies aim to determine whether rapid-response paradigms like mouse tracking can effectively capture the social perception of a racially ambiguous group such as Latine Americans.

Both studies received approval from the college's Institutional Review Board, and all participants provided informed electronic consent. All data and syntax for both studies, as well as preregistration for Study 2, are available here: https://osf.io/jkwue/?view_only=7388465ff6da4f049f313b0a8de0a96d

Study 1

The goal of Study 1 was to examine the real-time processes by which Latine faces are racially categorized using a binary-choice, computer mouse tracking task. Participants were asked to rapidly categorize a randomized set of faces representing typical (medium skin tone and prototypical), atypical (light skin and non-prototypical), and ambiguous (light skin with prototypical Latine traits) appearances of the Latine category. Building on previous research (Camacho and Reinka, 2026), we hypothesized that typical Latine faces would most frequently be categorized correctly as Latine American, while atypical faces would likely be miscategorized as White American. We also predicted these categorizations to be made with relatively quick response times and minimal deviation from an idealized straight line in the mouse trajectory. In contrast, we predicted that faces with ambiguous Latine features—those with light skin but still displaying Latine traits—would be the most challenging to categorize, resulting in longer reaction times and greater deviation in mouse trajectories, reflecting the added complexity of their racial categorization.

Method

Participants

A total of 139 undergraduate students attending either a Hispanic Serving Institution (n = 105) or a private liberal arts college (n = 34) in the Northeast United States2 participated in the study for course credit (M*age* = 19.71, SD*age* = 4.11). Most participants self-identified as female (n = 101), followed by male (n = 34), and non-binary/third gender (n = 3), with 1 person preferring not to say. Participants self-identified as Hispanic/Latino (n = 56), White/European American (n = 30), Black/African American (n = 23), Asian/Asian American (n = 16), multiracial (n = 10), or another racial identity (n = 4). The sample size was determined by the number of participants available in the subject pools during the academic semester at both colleges. A sensitivity power analysis using G*Power (Faul et al., 2007) for a within subjects, repeated measures ANOVA showed that Study 1 had enough power to detect a small effect, f = 0.12.

Materials and procedure

When participants arrived at the laboratory, they were seated at a laptop where an informed consent form was displayed. Each laptop had the trackpad disabled and instead used a corded computer mouse. Before each laboratory session, researchers ensured the mice cords were tangle-free and could be moved unincumbered. The cursor speed was set at the midline of the speed-range on each laptop.

After consenting, participants completed a mouse tracking task presented using MouseTracker (v. 2.83; Freeman and Ambady, 2010) in which they were asked to categorize twenty-five faces as either “White American” or “Latino American”. Face stimuli were colored photographs taken from the shoulders up, retrieved from the Chicago Face Database (Ma et al., 2015) and the American Multiracial Faces Database (Chen et al., 2021). Twenty of the 25 people in the photos self-identified as Latine: ten of the images showed light skin men and women (i.e., atypical Latine) and ten of the images showed medium skin men and women (i.e., typical Latine), as determined by the researchers and prior ratings of skin luminance scores (Ma et al., 2015). The remaining five images presented faces of light skin multiracial individuals who were deemed to appear phenotypically Latine by previous testing (i.e., ambiguous targets; Chen et al., 2021), but their specific racial backgrounds were not recorded.

MouseTracker records not just the participants' categorization of each image, but the trajectory of their computer mouse as they make their decision (Freeman and Ambady, 2010). Relevant to our investigation, the program records each trial's area under the curve (AUC; computation of the area between the actual mouse trajectory and an idealized straight line between start and end points), maximum deviation (MD; furthest point of mouse travel, measured perpendicularly from the idealized straight line), reaction time (RT; time taken to click the chosen category, in ms), and the number of x-flips (integer measure of how many times the participant's mouse reversed directions along the x-axis).3

As previously described, each trial begins when the participant clicks a “Start” button, centered on the bottom of the screen. For all participants, the button labeled “White American” was presented on the top left, and the button labeled “Latino American” was presented on the top right corner of the screen. Participants were instructed to start moving their mouse as soon as a face appeared, even if they had yet to make up their mind; this allows the software to track the decision-making process in real-time. If participants moved their mouse too slowly (> 750 ms; M*initiation*= 242.66 ms), they received an error message. If a participant took too long to decide (> 2000 ms) they received a time-out warning. The timeout warning latency was the default in the MouseTracker example programming (Freeman and Ambady, 2010); the initiation cutoff was set to be slightly faster than the default (1000 ms) to discourage pre-movement processing. Unrelated practice photos were presented before the task began to ensure participants understood the process. All trial images were presented in a randomized order, and participants saw each of the 25 photographs once.

After completing the task, participants answered several demographic questions and ranked their skin tone (1 = Very Light, 7 = Very Dark; M = 3.45, SD = 1.20), as well as how their skin tone compares to others of their racial/ethnic group (1 = Much lighter, 5 = Much darker; M = 2.53, SD = 0.792). Lastly, participants were presented with the NIS Skin Color Scale (Massey and Martin, 2003), a ten-point scale that shows a range of skin tones, represented by digitally altered hands, ranging from lightest (1) to darkest (10). They were asked to indicate what they believed to be the skin tone of the “average” Latine American (M = 3.68; SD = 1.02) and the “average” White American (M = 1.60; SD = 0.64). These items were included as exploratory measures; however, none significantly moderated the mouse tracking results and are therefore excluded from further analysis.

Analytic plan

All analyses were conducted in RStudio (v. 2024.12.1.563; Posit team, 2025). The three different facial categories (typical Latine, atypical Latine, and ambiguous) were used as the independent variable in all analyses; ambiguous (light skin, prototypical) faces were used as the reference group. Decision outcomes (e.g., “White American” vs. “Latino American”) were analyzed in a logistic regression framework using the Generalized Linear Models (“glm”) function in base R (R Core Team, 2016). Mouse tracking trajectory results (AUC, MD, RT, and x-flips) were analyzed using the Linear and Nonlinear Mixed Effects Models (“nlme”) package (v. 3.1-168; Pinheiro et al., 2025). AUC, MD, and RT were standardized for easier interpretation in the models but left unstandardized for descriptive statistics and easier visual interpretation in figures. The participant was included as a random intercept to account for the dependency of the data within each person. Any trials with RTs greater than 3 SDs from the mean were removed before analysis (trial n = 67 from 36 participants). Descriptive statistics and correlations between all mouse tracking outcomes (AUC, MD, RT, and x-flips) are presented in Table 1.

Table 1. Descriptive statistics and correlations between mouse tracking outcomes, Studies 1 and 2.

Variables**M (SD)**1234
Study 1
1. AUC0.746 (1.27)-
2. MD0.388 (0.460)0.897-
3. RT1540.29 (661.80)0.3510.283-
4. X-flips7.41 (3.47)0.1670.1750.107-
Study 2
1. AUC1.01 (13.31)-
2. MD0.941 (2.590)0.296-
3. RT1311.04 (859.83)0.0860.119-
4. X-flips1.89 (2.01)0.0580.1630.442-

AUC, Area Under the Curve; MD, Maximum Deviation; RT, Reaction Time. All correlations are significant at p < 0.001.

Results

Decision outcomes by facial category

First, we examined the differences in decision outcome by facial type. The typical Latine faces were more often categorized as “Latino American” (trial n = 1339, 97.0%) than “White American” (trial n = 42, 3.0%), whereas the atypical Latine faces showed the opposite pattern (“Latino American” trial n = 274, 20.4%; “White American” trial n = 1068, 79.6%). The ambiguous faces were more often categorized as “Latino American” (trial n = 580, 84.7%) than “White American” (trial n = 105, 15.3%); however, the difference was not as stark as among the typical Latine faces.

To test if these differences were significant, we ran a logistic regression. Compared to the null model, the inclusion of the facial category significantly improved model fit, 2(2) = 2118.4, p < 0.002. As expected, the atypical Latine faces were much less likely to be categorized as “Latino American” than the ambiguous faces [OR = 0.046, p < 0.001, 95% CI (0.036, 0.059)]. In contrast, the typical Latine faces were much more likely to be categorized as “Latino American” compared to the ambiguous faces [OR = 5.772, p < 0.001, 95% CI (4.014, 8.443)].

Mouse tracking trajectory analyses

To begin, we examined how facial category (typical, atypical, and ambiguous) affected participants' mouse trajectories, as measured by AUC. There were significant differences between facial categories: participants showed larger AUCs to the ambiguous faces (M = 0.915, SD = 1.453) compared to both the typical (M = 0.613, SD = 1.126) and atypical (M = 0.796, SD = 1.294) Latine faces, indicating that the ambiguous faces elicited the most trajectory divergence (see Figure 1a and Table 2). That is, participants swung their mouse further to one answer before deciding on another when viewing the ambiguous faces. Comparatively, the typical and atypical faces elicited more straightforward decisional processes.

Study 1 mouse tracking results measuring **(a)** AUC, **(b)** MD, **(c)** RT, and **(d)** X-flips. Error bars represent standard error.
Figure 1. Study 1 mouse tracking results measuring (a) AUC, (b) MD, (c) RT, and (d) X-flips. Error bars represent standard error.

Table 2. Mouse tracking results, studies 1 and 2.

AUCMDRTX-Flips
Fixed effects*b**SE**p**d**b**SE**p**d**b**SE**p**d**b**SE**p**d*
Study 1
Typical Latine faces−0.2410.044< 0.0010.088−0.2240.044< 0.0010.222−0.2650.038< 0.0010.275−0.5000.146< 0.0010.140
Atypical Latine faces−0.0990.0440.0260.242−0.0620.0440.1600.056−0.1270.038< 0.0010.140−0.3430.1470.0190.095
Random effectsVarianceVarianceVarianceVariance
Participant0.3170.1250.3452.281
Residual0.9450.8520.6139.732
Study 2
Typical Latine faces0.0170.0400.6720.0240.0450.0400.2560.041−0.0730.0300.0160.067−0.1390.0620.0240.068
Atypical Latine FACES0.0250.0400.5280.0240.0130.0400.7360.021−0.0860.0300.0040.089−0.2030.0620.0010.100
Random effectsVarianceVarianceVarianceVariance
Participant0.0110.0390.4711.658
Residual0.9870.9600.5562.334

Reference group is “Ambiguous” faces in both studies. AUC, Area Under the Curve; MD, Maximum Deviation; RT, Reaction Time.

The test of MD also showed a significant difference between the ambiguous (M = 0.440, SD = 0.495) and typical (M = 0.338, SD = 0.438) Latine faces, such that participants' mouse movements deviated more from an idealized straight line when categorizing the ambiguous faces, further showing the indecision participants felt to the ambiguous faces. Here, though, there was no significant difference between the ambiguous and the atypical (M = 0.413, SD = 0.460) Latine faces (see Figure 1b and Table 2). In examining MD, participants swung their mouse equally far to one decision choice before opting for the other when viewing the atypical (e.g., light skin, non-prototypical) and ambiguous (e.g., light skin, prototypical) faces.

In predicting RTs, we see that participants were significantly faster to categorize both the typical (M = 1474.784, SD = 600.875) and atypical (M = 1552.120, SD = 696.126) faces compared to the ambiguous faces (M = 1649.165, SD = 694.477; see Figure 1c and Table 2), suggesting the ambiguous stimuli required more cognitive processing time before ultimately reaching a decision.

Finally, we examined x-flips, the horizontal trajectory reversals. We once again see a difference by facial category, where the ambiguous faces (M = 7.734, SD = 3.556) elicited the most x-flips compared to either the typical (M = 7.248, SD = 3.426) or atypical Latine faces (M = 7.402, SD = 3.453). Similar to the other mouse tracking outcomes (AUC, MD, and RT), it appears as if the ambiguous facial stimuli elicit more decisional complexity in their attempts to categorize these faces as either “White,” or “Latino American,” as compared to the a/typical Latine faces (see Figure 1d and Table 2).

Discussion

As predicted, participants robustly and correctly categorized the typical Latine faces as “Latino American,” and miscategorized the atypical Latine faces as “White American.” Participants were also more likely to categorize the ambiguous faces as “Latino American,” albeit to a lesser degree than the medium skin (typical) faces. This is consistent with the pre-tested prototypicality of each type of stimuli. Further, the mouse-tracking results show that participants generally struggled to categorize the ambiguous stimuli: faces featuring the skin tone typical of White Americans combined with facial features typical of Latine Americans. We see this via greater mouse trajectory deviation (AUC and MD), in increased cognitive processing time (RT), and increased decisional complexity (x-flips).

According to the theory of racial phenotypicality bias (Maddox, 2004), individuals who appear prototypical of their racial group are more likely to be perceived through the lens of group-based stereotypes. In Study 2, we aim to extend these findings by examining whether prototypically appearing Latine individuals (both typical and ambiguous in appearance) are more likely to be categorized as foreign compared to atypical Latine individuals. Additionally, we explore whether the mouse trajectories associated with these stereotypical categorizations mirror the patterns observed in Study 1, where medium skin tone was the easiest to categorize and ambiguous-looking Latine targets generated the greatest decision complexity.

Study 2

The goal of Study 2 was to examine the real-time processes by which Latine faces are categorized as either American or Foreign. Specifically, participants were asked to rapidly categorize a randomized set of faces—representing typical (medium skin tone, prototypical), atypical (light skin, non-prototypical), and ambiguous (light skin, prototypical) Latine appearances—as either “American” or “Foreign.” Consistent with prior research (Camacho and Reinka, 2026), we predicted that typical Latine faces would be categorized as “Foreign,” while atypical faces would more likely be categorized as “American.” As in Study 1, we predicted that these categorizations would be made with relatively quick response times and minimal deviation from an idealized straight line in the mouse trajectory. However, we anticipated that faces with ambiguous Latine features—those that appear both White and Latine—would be more difficult to categorize, resulting in greater deviation from clear categorization as “Foreign,” a category that aligns with the stereotype of Latine individuals as perpetual foreigners (Zou and Cheryan, 2017).

Method

Participants

Participants were U.S. adults recruited via Connect by Cloud Research who were monetarily compensated for completing the study. After exclusions for incomplete data (n = 22), poor data quality (n = 11), or technical issues (n = 3), the final sample yielded N = 206 individuals (M*age* = 38.80, SD*age* = 11.44). Participants self-identified as male (n = 109) and female (n = 96), with 1 person preferring not to say. Participants self-identified as Hispanic/Latino (n = 26), White/European American (n = 147), Black/African American (n = 13), Asian/Asian American (n = 15), Native American or Alaska Native (n = 1), or multiracial (n = 4). The sample size for Study 2 was determined by a desire to obtain an effect size of f = 0.10. A sensitivity analysis using G*Power (Faul et al., 2007) for a within-subjects, repeated measures ANOVA indicated that this study had enough power to detect a small effect, f = 0.098.4

Materials and procedure

Unlike Study 1, Study 2 was conducted completely online via Qualtrics using a mouse-tracking integration JavaScript code as outlined in Mathur and Reichling (2019). This code produces the same outcome variables of interest as the desktop version of MouseTracker (v. 2.84; Freeman and Ambady, 2010): AUC, MD, RT, and x-flips; however, there are some differences inherent in the online presentation of stimuli.

Participants were required to complete the task on a personal computer (not a phone or tablet). If a participant completed the task via touchscreen, there would be no recorded mouse trajectory, and they would be excluded from analysis. Similarly, bots can only make direct selections (clicks); they cannot move a mouse, and therefore, bot activity would be excluded in the same manner. In addition, the JavaScript code allows several warnings to be displayed as needed. At the beginning of the first practice trial, participants were shown a warning if their screen size was too small to see the entirety of the experimental setup. In these cases, they were asked to expand the screen before continuing. Unlike the desktop version by MouseTracker (Freeman and Ambady, 2010), the trial begins when the image loads. If participants took too long (> 700 ms; M*initiation*= 682.10) to start moving their cursor after the image appeared, they were shown an alert to begin earlier; conversely, if participants moved their cursor while the image was loading, they were told they started too early. If participants took too long (> 5000 ms) to select a response, they were shown an alert reminding them to respond faster. See Table 3 for the proportion of trials that triggered each alert. The allowable initiation time, as well as the cut-off time, were the defaults in the task programming (Mathur and Reichling, 2019), and give more leeway than desktop programming due to the variability of image load times in online data collection. In the online translation of this task, participants click a “Next” button, centered at the bottom of the experimental screen, which not only moves them onto the next trial but also provides a consistent cursor starting point. This is analogous to the “Start” button used for the same purpose on the desktop version of the mouse tracking task.

Table 3. Percentage trials with alerts shown, study 2.

Alert type% trials
None60.87%
Started too early15.00%
Started too late21.34%
Surpassed trial time limit2.80%

Despite these idiosyncrasies, the experimental setup was similar to Study 1. Participants were again asked to categorize faces into binary choices. However, participants were asked to categorize the faces as “American” (presented on the top left corner of the screen) or “Foreign” (presented on the top right corner of the screen). The face stimuli, too, were slightly altered. Rather than combining different stimulus sources as in Study 1, we used the faces of 18 self-identified Latine men and women from the Chicago Face Database (Ma et al., 2015). Based on the data provided in the Database, six faces were medium-skinned (luminance M = 144.50, SD = 10.12) and viewed as prototypically Latine (Latine probability M = 78.18%, SD = 5.71%; “typical”), while another six were light-skinned (luminance M = 167.00, SD = 7.73), but viewed as prototypically White (Latine probability M = 14.43%, SD = 9.02%; White probability M = 76.42%, SD = 10.75%; “atypical”). Instead of self-identified multiracial faces as in Study 1, here we used six self-identified Latine targets who were light-skinned (luminance M = 154.92, SD = 6.48) but viewed as prototypically Latine (Latine probability M = 79.22%, SD = 11.68%; “ambiguous”). Moreover, the equal number of stimuli across categories allows for consistent base-rates, and therefore more consistent priors for the participants, compared to Study 1. The stimuli presentation order was randomized, and participants saw each image once.

After the mouse tracking task, participants were asked the same demographic questions as in Study 1. They were also presented with questions about their skin tone (M = 2.50, SD = 1.13) as well as how their skin tone compared to others of their same racial or ethnic group (M = 2.55, SD = 0.761). Finally, participants again completed ratings of the average White American's skin tone (M = 1.77, SD = 0.799) and average Latine American's skin tone (M = 3.68, SD = 1.05) using the ten-point NIS Skin Color Scale (Massey and Martin, 2003). Similar to Study 1, these exploratory variables are not considered further.

Analytic plan

Results from Study 2 were analyzed in RStudio (v. 2024.12.1.563; Posit team, 2025) in much the same way as in Study 1. Any trials with RTs greater than 3 SDs from the mean were removed before analysis (n*trials* = 24 from 9 participants). The three image categories (ambiguous, typical, and atypical) were used as the independent variable in all analyses, with the ambiguous category used as the reference group. Logistic regressions were used to test the decision outcomes (e.g., “Foreign” vs. “American”) using the “glm” function in base R (R Core Team, 2016). Mouse trajectory results were again analyzed in the “nlme” package (v. 3.1-168; Pinheiro et al., 2025) with participant allowed as a random intercept. Finally, AUC, MD, and RT were again standardized for easier interpretation in the models but left unstandardized for descriptive statistics and figures. Descriptive statistics and correlations of all mouse tracking outcomes are presented in Table 1.

Results

Decision outcomes by facial category

To begin, we examined the decision outcomes (“Foreign” vs. “American”) by facial stimuli type. As expected, the typical Latine faces were more frequently categorized as “Foreign” (trial n = 910, 74.0%) than “American” (trial n = 320, 26.0%), and the atypical Latine faces were more often deemed to be “American” (trial n = 982, 80.0%) than “Foreign” (trial n = 245, 20.0%). Participants were more evenly split in their categorization of the ambiguous faces (“Foreign” trial n = 684, 55.7%; “American” trial n = 544, 44.7%). Testing these counts with a logistic regression showed that the inclusion of facial category did improve the model fit above and beyond the null model, X2(2) = 785.09, p < 0.001. Compared to the ambiguous faces, the typical Latine faces were much more likely to be categorized as “Foreign” than “American,” OR = 2.26, p < 0.001, 95% CI (1.909, 2.683), whereas the atypical Latine faces were much less likely to be categorized as “Foreign” (as opposed to “American”) than the ambiguous faces, OR = 0.198, p < 0.001, 95% CI (.166, 0.237).

Mouse tracking trajectory analyses

In examining AUC, there were no significant differences between facial types (ambiguous: M = 0.820, SD = 7.547; typical: M = 1.052, SD = 11.133; atypical: M = 1.158, SD = 18.726; see Figure 2a and Table 2). MD similarly yielded no differences by stimuli type (ambiguous: M = 0.890, SD = 1.371; typical: M = 1.008, SD = 3.853; atypical: M = 0.824, SD = 1.842; see Figure 2b and Table 2). In this context of national categorization, we do not see any difference in mouse trajectory deviation when participants categorize the ambiguous faces vs. the typical or atypical Latine faces.

Study 2 mouse tracking results measuring **(a)** AUC, **(b)** MD, **(c)** RT, and **(d)** X-flips. Error bars represent standard error.
Figure 2. Study 2 mouse tracking results measuring (a) AUC, (b) MD, (c) RT, and (d) X-flips. Error bars represent standard error.

However, when examining RT, we did see significant effects: participants took longer to categorize the ambiguous (M = 1356.200, SD = 885.311) faces than either the typical (M = 1297.377, SD = 858.939) or atypical Latine faces (M = 1279.546, SD = 833.271; see Figure 2c and Table 2). This suggests that participants spent more time cognitively processing how to categorize the ambiguous (light skin, prototypically Latine) faces than the comparatively easier task of categorizing the typical or atypical Latine faces. Similarly, participants demonstrated significantly more x-flips when they categorized the ambiguous faces (M = 1.973, SD = 2.142) than either the typical (M = 1.835, SD = 1.909) or atypical faces (M = 1.769, SD = 1.961; see Figure 2d and Table 2). This suggests that, in line with our predictions, participants showed the most indecision by moving their mouse back and forth between the outcome options when they saw the ambiguous stimuli.

Discussion

As predicted, participants categorized the typical Latine faces (medium skin tone and prototypical) as more stereotypically “Foreign” and the atypical Latine (light skin and non-prototypical) faces as “American.” However, their responses were more mixed when categorizing the ambiguous (light skin and prototypical) faces. Although these faces were still more frequently categorized as “Foreign” (684 trials) than “American” (544 trials), the gap was much smaller than for the typical Latine faces. Consistent with Study 1, these findings highlight the complexity of rapidly categorizing Latine faces that display physical characteristics associated with both White and Latine Americans.

The mouse-tracking results of Study 2, however, show a mixed picture. Specifically, as expected, participants took longer and generated more x-flips when categorizing the faces of ambiguous targets than a/typical targets. Contrary to our prediction, however, the more direct measures of mouse trajectory deviation—AUC and MD—failed to show significant differences between the facial stimuli. The discrepancy in outcomes between Studies 1 and 2 could be due to several factors, including more erratic mouse trajectories when completing the study online (Study 2) vs. in person (Study 1), or the overall difficulty of categorizing nationality (Study 2) rather than race (Study 1). Nevertheless, given that we do see differences in RT and x-flips in Study 2, we can reasonably surmise that participants did struggle more to decide how to categorize the ambiguous faces than the a/typical Latine faces, but that indecision was not captured in AUC and MD in this experimental paradigm.

General discussion

The present research examined the social cognitive processes underlying the racial categorization and stereotyping of Latine faces. Both typical (medium skin tone, prototypical) and ambiguous (light skin, prototypical) Latine individuals were more likely to be categorized as Latino American and labeled as “Foreign,” compared to atypical (light skin, non-prototypical) Latine individuals. Our findings also revealed that both skin tone and Latine prototypicality uniquely influence the dynamic cognitive processes involved in these categorizations. Specifically, as predicted, mouse trajectories for typical and atypical Latine targets showed quick reaction times with minimal deviation from an idealized straight line. In contrast, mouse trajectories for ambiguous Latine targets exhibited greater complexity, leading to more significant deviations during racial categorization and greater indecision in national categorization. These results highlight that both skin tone and Latine prototypicality are key predictors of how Latine individuals are categorized. Furthermore, even when light-skinned and medium-skinned Latine individuals are matched for high prototypicality, skin tone continues to affect the cognitive processes involved in their categorization.

These findings align with the dynamic interactive theory of person construal and are consistent with traditional mouse-tracking research. Specifically, prior work has shown that ambiguous physical appearance—such as possessing features characteristic of multiple racial categories—creates cognitive competition that complicates final categorization decisions (e.g., Freeman et al., 2020). In our work, this was evident in the categorization of light-skinned, prototypically Latine-appearing individuals, as indicated by higher AUC, longer reaction times, and more x-flips. Importantly, we provide the first evidence of this effect in perceptions of Latine targets, a group that is both racially ambiguous (Chen et al., 2021) and growing in numerical representation in the United States (U.S. Census Bureau, 2023). Our findings demonstrate that, despite the racial ambiguity of Latine Americans as a whole, medium skin tone serves as the most direct and reliable phenotypic cue for categorizing someone as Latine. In addition, while past studies have shown that light-skinned Latine individuals are more likely to be categorized as White than Latine in implicit binary choice outcomes (Camacho and Reinka, 2026), our findings add nuance by revealing that even light-skinned, prototypically-appearing Latine individuals are still predominantly categorized as Latine, though this categorization requires more cognitive effort, highlighting the complexity of these judgments.

Our findings are also consistent with the theory of racial phenotypicality bias. As posited by the theory, we find that people who appear more prototypical of a racial group are more likely to be perceived through the lens of group-based stereotypes (Maddox, 2004). Specifically, we found that both medium-skin, prototypically appearing Latine individuals and light-skinned, prototypically appearing Latine individuals were significantly more likely to be categorized as “Foreign”—a stereotype commonly associated with Latine individuals—compared to light-skinned, non-prototypically appearing Latine individuals. Novel to our study, however, is the finding that the dynamic pathways by which these categorizations are made differ for ambiguous Latine targets compared to either typical or atypical targets. Participants took longer and reversed directions more along the x-axis when categorizing these faces, demonstrating indecision and delay in categorizing them as either stereotypical (“Foreign”) or counter-stereotypical (“American”). Notably, AUC and MD were not statistically significant in Study 2, which may suggest that the complexity involved in stereotypically categorizing ambiguous Latine faces is more effectively captured by subtler measures of decisional conflict—such as x-flips—rather than by overall deviation from a direct response trajectory, which is what AUC reflects (Freeman et al., 2010). Regardless, our findings demonstrate that faces that prior research (and Study 1) showed to be the most prototypical (medium skin tone) are also dynamically categorized as the most stereotypical, while ambiguous Latine faces, though largely categorized in the same manner, require more cognitive effort.

Our research has several limitations that affect the generalizability of our findings (Simons et al., 2017). Although we observed significant differences in the dynamic racial categorization and stereotyping of Latine individuals based on skin tone and prototypicality, we were unable to identify the specific facial features that make light-skinned individuals appear prototypically Latine, limiting our ability to manipulate Latine faces beyond skin tone and average prototypicality. Additionally, while using real faces—rather than the morphed faces commonly used in mouse-tracking research (Chwe et al., 2024)—enhances ecological validity, it introduces uncontrolled variability in previously identified Latine prototypicality features such as lip shape.

We also did not employ a fully crossed design, as we were unable to include medium-skinned Latine individuals who varied in prototypicality, preventing us from disentangling the effects of skin tone and prototypicality for medium-tone individuals and restricting such conclusions only to light-skinned targets. Moreover, prior work shows that impressions depend on both target and perceiver characteristics (Hehman et al., 2017), suggesting that racial and national categorizations may be shaped by participant factors such as prior exposure to Latine individuals—factors we did not measure. In addition, individuals may also appear prototypically Latine for reasons beyond physical appearance (e.g., name, Spanish proficiency; Garcia and Abascal, 2016; Sanchez et al., 2012), underscoring the need for future research to examine how non-phenotypic features contribute to prototypicality judgments and stereotypical categorizations.

Finally, in Study 2, participants could use either a trackpad or an external mouse depending on their personal computing setup. Because we could not assess which device was used, reaction time variability likely differed across participants. However, given our within-participants design, this variability should be greater between participants than within participants, and our inclusion of participant as a random effect should largely account for these differences in the statistical model.

Despite these limitations, our findings have important implications for understanding how Latine Americans are racially categorized and stereotyped in various contexts. For example, both medium- and light-skinned Latine individuals who appeared prototypically Latine were categorized as Latine American (Study 1) and stereotyped as foreigners (Study 2). This suggests that prototypical targets are likely to experience similar forms of explicit stereotyping and bias due to their prototypicality (cf. Maddox, 2004). However, in contexts where stereotyping and bias are automatically activated or occur within limited time frames, medium-skinned Latine individuals may be more susceptible to these biases. As shown in the current study, their racial categorization and stereotyping transpired more quickly and directly. Therefore, in situations like the shooting paradigm—where groups stereotyped as criminals are more likely to be erroneously shot in a computer-simulated game (Correll et al., 2014)—medium-skinned Latine individuals may be more likely to experience this bias than their equally prototypical, but lighter-skinned, Latine counterparts. Indeed, skin tone is an important predictor of stronger implicit bias in split-second “shoot/don't shoot” decisions, with dark-skinned Black targets more likely to be mistakenly identified as threats compared to light-skinned Black targets (Kahn and Davies, 2011). This is particularly relevant for Latine Americans given that they are also stereotyped as criminals (Zou and Cheryan, 2017), in addition to stereotypes of foreignness, as measured here.

Conclusion

Despite being the largest racial or ethnic minority group in the United States (U.S. Census Bureau, 2023), Latine Americans remain underrepresented in person-perception studies examining the role of skin tone and phenotypic features in their stereotypicality (Maddox et al., 2022). We aimed to address this limitation by exploring the continuous and dynamic cognitive processes through which Latine Americans are racially (White vs. Latine) and nationally (American vs. Foreign) categorized. Our findings align with previous studies on the explicit and implicit categorization and stereotyping of Latine individuals: medium skin tones are perceived as more prototypically Latine and stereotypically foreign compared to lighter skin tones (Camacho and Reinka, 2026; Chirco and Buchanan, 2021; Wilton et al., 2013). However, we demonstrate, for the first time, the unique influences of prototypicality and skin tone on the categorization and stereotyping of light-skinned Latine individuals. Specifically, we find that ambiguous-appearing Latine targets experience similar categorizations compared to those with medium skin tones, but the cognitive pathways to such categorizations are intermixed with the possibility of being considered White and American—possibilities not as readily ascribed to their medium-skinned counterparts. These findings underscore the need for further exploration of the cognitive processes through which Latine Americans are perceived and the consequences of such processes.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Aisha Phoenix, King's College London, United Kingdom

Reviewed by: David S. March, Florida State University, United States

René Rejón, The University of Melbourne, Australia

Footnotes

Throughout the manuscript, we use the term Latine as a non-binary alternative to Latino. However, we retain the term Latino when referencing the study itself, as this is the terminology that participants were presented with.

These two institutions were chosen based on convenience (i.e., they are the home institutions of the first two authors).

MouseTracker also provides y-flips (the number of vertical directional reverses); however, we do not examine this in the current manuscript. Y-flips are more pertinent to choice paradigms with more than two options (e.g., that place options in all four corners of the screen with the starting point in the middle); our setup will create more decisional conflict horizontally between the two presented options (e.g., x-flips), and therefore we see the y-flips as irrelevant.

In our preregistration, we said we would conduct simple ANOVAs instead of the linear mixed effects model. We feel that being able to account for the dependency of within-participant data is necessary to more appropriately test our hypotheses and therefore justified this deviation. The general pattern of results holds, though, if conducted in an ANOVA framework, as is reported in the Supplemental Material on OSF.

Statements

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This study was supported by an Open Access Publication Award provided to GC by the John Jay College of Criminal Justice/CUNY.

Data availability

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://osf.io/jkwue/?view_only=7388465ff6da4f049f313b0a8de0a96d.

Ethics approval

The studies involving humans were approved by the Institutional Review Boards at the City University of New York and Ursinus College. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Author contributions

GC: Data curation, Writing – review & editing, Methodology, Writing – original draft, Conceptualization, Investigation. MR: Investigation, Data curation, Writing – original draft, Formal analysis, Methodology, Writing – review & editing. CB: Visualization, Formal analysis, Writing – review & editing.

Acknowledgements

We would like to thank our research assistants who helped administer and supervise Study 1 and provided feedback on Study 2: Jaliela Rivera, Danielle Key, Daniah Dailey, Imani Thomas, Brandon Dolores, Avenlea Johnson, and Hailie Davis.

References

  1. Camacho G., Abouras A. (2025). I, too, am America: displaying national symbols on clothing increases the perceived ethnic and civic nationalism of Latinx Americans. J. Soc. Psychol., 10, 1 – 13. doi: doi:10.1080/00224545.2025.2503007
  2. Camacho G., Reinka M. A. (2026). The effect of skin tone on the implicit categorization and stereotyping of Latine Americans. Soc. Cogn.
  3. Chen J. M., Norman J. B., Nam Y. (2021). Broadening the stimulus set: introducing the American Multiracial Faces Database. Behav. Res. Methods 53, 371 – 389. doi: doi:10.3758/s13428-020-01447-8
  4. Cheryan S., Monin B. (2005). Where are you really from?: Asian Americans and identity denial. J. Pers. Soc. Psychol. 89, 717 – 730. doi: doi:10.1037/0022-3514.89.5.717
  5. Chirco P., Buchanan T. M. (2021). The (un)making of a citizen: the role of skin tone in perceived immigrant status and support for hostile immigration policy. Transl. Issues Psychol. Sci. 7, 378 – 391. doi: doi:10.1037/tps0000283
  6. Chwe J. A. H., Lick D. J., Freeman J. B. (2024). Reflexive activation of monoracial categories during multiracial categorization. Pers. Soc. Psychol. Bull. 52, 161 – 175. doi: doi:10.1177/01461672241273194
  7. Correll J., Hudson S. M., Guillermo S., Ma D. S. (2014). The police officer's dilemma: a decade of research on racial bias in the decision to shoot. Soc. Personal. Psychol. Compass 8, 201 – 213. doi: doi:10.1111/spc3.12099
  8. Devos T., Gavin K., Quintana F. J. (2010). Say “adios” to the American dream? The interplay between ethnic and national identity among Latino and Caucasian Americans. Cult. Divers. Ethnic Minor. Psychol. 16, 37 – 49. doi: doi:10.1037/a0015868
  9. Faul F., Erdfelder E., Lang A. G., Buchner A. (2007). G * Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav. Res. Methods 39, 175 – 191. doi: doi:10.3758/BF03193146
  10. Feliciano C. (2016). Shades of race: how phenotype and observer characteristics shape racial classification. Am. Behav. Sci. 60, 390 – 419. doi: doi:10.1177/0002764215613401
  11. Freeman J. B. (2018). Doing psychological science by hand. Curr. Dir. Psychol. Sci. 27, 315 – 323. doi: doi:10.1177/0963721417746793
  12. Freeman J. B., Ambady N. (2010). MouseTracker: software for studying real-time mental processing using a computer mouse-tracking method. Behav. Res. Methods 42, 226 – 241. doi: doi:10.3758/BRM.42.1.226
  13. Freeman J. B., Ambady N. (2011). A dynamic interactive theory of person construal. Psychol. Rev. 118, 247 – 279. doi: doi:10.1037/a0022327
  14. Freeman J. B., Pauker K., Apfelbaum E. P., Ambady N. (2010). Continuous dynamics in the real-time perception of race. J. Exp. Soc. Psychol. 46, 179 – 185. doi: doi:10.1016/j.jesp.2009.10.002
  15. Freeman J. B., Stolier R. M., Brooks J. A. (2020). ”Dynamic interactive theory as a domain-general account of social perception,“ in Advances in Experimental Social Psychology (San Diego, CA: Elsevier Academic Press), 237 – 287.
  16. Garcia D., Abascal M. (2016). Colored perceptions: racially distinctive names and assessments of skin color. Am. Behav. Sci. 60, 420 – 441. doi: doi:10.1177/0002764215613395
  17. Golash-Boza T., Darity W. Jr. (2008). Latino racial choices: the effects of skin colour and discrimination on Latinos and Latinas racial self-identifications. Ethn. Racial Stud. 31, 899 – 934. doi: doi:10.1080/01419870701568858
  18. Harris K., Armenta A. D., Reyna C., Zárate M. A. (2020). ”Latinx stereotypes: myths and realities in the twenty-first century,“ in Stereotypes: The Incidence and Impacts of Bias, eds J. T. Nadler and E. C. Voyles (Santa Barbara, CA: Praeger), 128 – 143.
  19. Hehman E., Ingbretsen Z. A., Freeman J. B. (2014). The neural basis of stereotypic impact on multiple social categorization. NeuroImage 101, 704 – 711. doi: doi:10.1016/j.neuroimage.2014.07.056
  20. Hehman E., Sutherland C. A. M., Flake J. K., Slepian M. L. (2017). The unique contributions of perceiver and target characteristics in person perception. J. Pers. Soc. Psychol. 113, 513 – 529. doi: doi:10.1037/pspa0000090
  21. Irizarry Y., Monk E. P., Cobb R. J. (2022). Race-shifting in the United States: Latinxs, skin tone, and ethnoracial alignments. Sociol. Race Ethn. 9, 37 – 55. doi: doi:10.1177/23326492221114813
  22. Johnson K. L., Freeman J. B., Pauker K. (2012). Race is gendered: how covarying phenotypes and stereotypes bias sex categorization. J. Pers. Soc. Psychol. 102, 116 – 131. doi: doi:10.1037/a0025335
  23. Kahn K. B., Davies P. G. (2011). Differentially dangerous? Phenotypic racial stereotypicality increases implicit bias among ingroup and outgroup members. Group Process. Intergroup Relat. 14, 569 – 580. doi: doi:10.1177/1368430210374609
  24. Ma D. S., Correll J., Wittenbrink B. (2015). The Chicago face database: a free stimulus set of faces and norming data. Behav. Res. Methods 47, 1122 – 1135. doi: doi:10.3758/s13428-014-0532-5
  25. Ma D. S., Koltai K., McManus R. M., Bernhardt A., Correll J., Wittenbrink B. (2018). Race signaling features: identifying markers of racial prototypicality among Asians, Blacks, Latinos, and Whites. Soc. Cogn. 36, 603 – 625. doi: doi:10.1521/soco.2018.36.6.603
  26. MacLin O. H., Malpass R. S. (2001). Racial categorization of faces: the ambiguous race face effect. Psychol. Public Policy Law 7, 98 – 118. doi: doi:10.1037/1076-8971.7.1.98
  27. Maddox K. B. (2004). Perspectives on racial phenotypicality bias. Pers. Soc. Psychol. Rev. 8, 383 – 401. doi: doi:10.1207/s15327957pspr0804_4
  28. Maddox K. B., Perry J. M., Pagan J. (2022). Cues and categories: revisiting paths to racial phenotypicality bias. Soc. Personal. Psychol. Compass 16: e12699. doi: doi:10.1111/spc3.12699
  29. March D. S. (2022). Perceiving a danger within: Black Americans associate Black men with physical threat. Soc. Psychol. Personal. Sci. 14, 942 – 951. doi: doi:10.1177/19485506221142970
  30. March D. S., Graham R. (2015). Exploring implicit ingroup and outgroup bias toward Hispanics. Group Process. Intergroup Relat. 18, 89 – 103. doi: doi:10.1177/1368430214542256
  31. Massey D. S., Martin J. A. (2003). The NIS Skin Color Scale. Princeton, NJ: Office of Population Research, Princeton University.
  32. Mathur M. B., Reichling D. B. (2019). Open-source software for mouse-tracking in Qualtrics to measure category competition. Behav. Res. Methods 51, 1987 – 1997. doi: doi:10.3758/s13428-019-01258-6
  33. Norman J. B., Lingl D. C., Hehman E., Chen J. M. (2024). Race in the eye of the beholder: decomposing perceiver-and individual-level variation in perceived racial prototypicality. J. Exp. Soc. Psychol. 115: 104667. doi: doi:10.1016/j.jesp.2024.104667
  34. Pinheiro J., Bates D. R Core Team. (2025). nlme: linear and nonlinear mixed effects models. R package version 3.1-168. Available online at: (Accessed September 5, 2025).
  35. Posit team (2025). RStudio: Integrated Development Environment for R. Boston, MA: Posit Software, PBC. Available online at: (Accessed September 5, 2025).
  36. R Core Team (2016). R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing. Available online at: (Accessed September 5, 2025).
  37. Sanchez D. T., Chavez G., Good J. J., Wilton L. S. (2012). The language of acceptance: Spanish proficiency and perceived intragroup rejection among Latinos. J. Cross Cult. Psychol. 43, 1019 – 1033. doi: doi:10.1177/0022022111416979
  38. Simons D. J., Shoda Y., Lindsay D. S. (2017). Constraints on generality (COG): a proposed addition to all empirical papers. Perspect. Psychol. Sci. 12, 1123 – 1128. doi: doi:10.1177/1745691617708630
  39. Stepanova E. V., Strube M. J. (2012). The role of skin color and facial physiognomy in racial categorization: moderation by implicit racial attitudes. J. Exp. Soc. Psychol. 48, 867 – 878. doi: doi:10.1016/j.jesp.2012.02.019
  40. Stolier R. M., Freeman J. B. (2017). A neural mechanism of social categorization. J. Neurosci. 37, 5711 – 5721. doi: doi:10.1523/JNEUROSCI.3334-16.2017
  41. Telles E. E. (2014). Pigmentocracies: Ethnicity, Race, and Color in Latin America. Chapel Hill, NC: UNC Press Books.
  42. U.S. Census Bureau (2023). Hispanic Heritage Month: 2023. Available online at: (Accessed August 7, 2025).
  43. Uhlmann E., Dasgupta N., Elgueta A., Greenwald A. G., Swanson J. (2002). Subgroup prejudice based on skin color among Hispanics in the United States and Latin America. Soc. Cogn. 20, 198 – 226. doi: doi:10.1521/soco.20.3.198.21104
  44. Wilkins C. L., Kaiser C. R., Rieck H. (2010). Detecting racial identification: the role of phenotypic prototypicality. J. Exp. Soc. Psychol. 46, 1029 – 1034. doi: doi:10.1016/j.jesp.2010.05.017
  45. Wilton L. S., Sanchez D. T., Chavez G. F. (2013). Speaking the language of diversity: Spanish fluency, White ancestry, and skin color in the distribution of diversity awards to Latinos. Basic Appl. Soc. Psychol. 35, 346 – 359. doi: doi:10.1080/01973533.2013.803969
  46. Wojnowicz M. T., Ferguson M. J., Dale R., Spivey M. J. (2009). The self-organization of explicit attitudes. Psychol. Sci. 20, 1428 – 1435. doi: doi:10.1111/j.1467-9280.2009.02448.x
  47. Zou L. X., Cheryan S. (2017). Two axes of subordination: a new model of racial position. J. Pers. Soc. Psychol. 112, 696 – 717. doi: doi:10.1037/pspa0000080

Citation

Camacho, G., et al. (2026). The effects of skin tone and prototypicality on the dynamic racial categorization and stereotyping of Latine Americans. Frontiers in Social Psychology, Volume 4. https://doi.org/
Received
Accepted
Published
Volume
4 - 2026

Copyright

Copyright © 2026 Camacho, Reinka and Burt.

License: https://creativecommons.org/licenses/by/4.0/

*Correspondence: Gabriel Camacho

Article metrics

Views
0
Downloads
0
Citations
0
Shares
0

Related articles

Original ResearchHow does belief in conspiracy theories affect loneliness and connectedness? A longitudinal investigationMatt N. Williams, John R. Kerr, Mathew D. Marques, Haneen Abu-Obeid, Stephen R. HillOriginal Research“They got what they deserved”: situational justification and schadenfreude toward an LGBTQIA+ organizationIvan Giuseppe Cammarata, Costanza Scaffidi Abbate, Martina Macrillò, Stefano BocaOriginal ResearchAnother one bites the dust? Threatened meat-eater identity and veg * n conspiracy beliefsDominika Adamczyk, Natalia Korczak, Marta Marchlewska