Emotional Responses to Racial Violence: Analyzing Sentiments and Emotions Among Black Women in Missouri

Abstract

This study examines the emotional responses of Black women in Missouri regarding incidents of racial violence in the United States. Grounded in an analysis of self-reported emotions, this study explores how Black women (n = 384, M_age = 37) express their emotional experiences in response to racial violence. Utilizing the Multiple Affect Adjective Checklist-Revised (MAACL-R), sentiment analysis was used to assess the overall emotional tone of participants’ responses, while emotion analysis was used to identify specific emotions expressed. The findings highlight the complexities of Black women’s emotional responses, considering factors such as coping mechanisms, racial identity beliefs, spirituality and religiosity, and resilience and strength. By applying computational methods to analyze these emotions, this study reveals how racial violence shapes sentiment and emotional expression patterns. Furthermore, it highlights the significance of acknowledging the complex ways Black women navigate and process racial violence.

1. Introduction

Racial violence, which is defined as acts of aggression, discrimination, or harm based on an individual’s race, remains a persistent and deep-seated issue in the United States [1] (pp. 599–619). In particular, Black communities have suffered greatly with the impacts of both psychological and emotional distress [2]. According to the FBI, over 60% of reported hate crimes in 2022 were race-based, with Black individuals making up the largest group of victims [3]. Moreover, 75% of Black adults reported experiencing racial discrimination either regularly (13%) or occasionally (62%) [4]. As a result, these ongoing experiences of racial violence can result in heightened risks for anxiety, depression, and post-traumatic stress disorder [5] (pp. 1152–1173).

While these trends affect the collective Black community, Black women’s experiences are shaped by intersecting forms of oppression [6] (pp. 132–133). The concept of gendered racism, which is the discrimination and oppression experienced by individuals based on the intersection of their race and gender, captures the unique challenges faced by Black women [7,8] (see [7], pp. 314–343). These experiences manifest through microaggressions, stereotyping, exclusion, and heightened social scrutiny, which can have a clear emotional and psychological impact [7] (pp. 314–343). Additionally, Black women often carry the dual burden of enduring racial violence while emotionally supporting others in their communities [8].

This study explores how Black women in Missouri express their emotional responses to incidents of racial violence in America and how broader sociopolitical and cultural contexts shape those responses. Using the Multiple Affect Adjective Checklist-Revised (MAACL-R) to assess self-reported responses and applying computational models such as sentiment and emotional analysis using RoBERTa, this research explores emotional expression patterns in response to racial violence incidents.

Building on the literature that applies sentiment and emotion analysis to racial and social justice discourse [9,10,11,12,13] (see [9], pp. 165–190; [11], pp. 51–59; [12], pp. 87–100), this study contributes a nuanced perspective on how racial violence affects Black women’s emotional states. By centering Black women’s voices and utilizing psychometric and computational tools, this work advances intersectional understandings of emotional well-being in the face of systemic racism.

1.1. Racial Violence in the Lives of Black Women

1.1.1. Racial Violence and Black Women

For Black women, racial violence is often exacerbated by gender-based discrimination. These experiences are shaped by gendered racism, a term used to describe the discrimination and oppression experienced by individuals based on the intersection of their race and gender [7,14] (see [7], pp. 314–343). Black women face unique vulnerabilities, including hypersexualization, maternal stereotyping, and criminalization, particularly in media and public discourse [6] (pp. 132–133). These intersecting oppressions show up in both everyday encounters and high-profile incidents of violence, such as the deaths of Sonya Massey, Breonna Taylor, Atatiana Jefferson, Sandra Bland, and Michelle Cusseaux [15,16]. These incidents expose the disproportionate risk Black women face within systems of policing and justice [15,16].

In addition to direct harm, Black women often assume the emotional labor of shielding their families and communities from the psychological impact of racism while at the same time being expected to display strength and resilience [8]. These expectations can make it hard for Black women to show vulnerability, often leading to hidden emotions or their distress being overlooked [8]. Despite the central role that Black women play in sustaining their communities, their emotional responses to racial violence remain underexplored in empirical research.

1.1.2. Racial Violence and Sentiment Analysis

Sentiment analysis, a Natural Language Processing (NLP) technique used to assess emotional tone in text, has gained traction in the study of public responses to racial and social justice issues [17,18]. Prior research has employed sentiment analysis to examine reactions to events such as police shootings, social media discourse around Black Lives Matter, and racialized news coverage [9,11,18]. These studies reveal patterns of collective outrage, grief, resilience, and sometimes apathy or uncertainty.

However, much of this work has focused on public opinion at large, often overlooking how sentiment differs across marginalized groups. In particular, Black women’s sentiment in response to racial violence may differ in tone, intensity, and complexity due to their lived experiences and social positioning. For instance, expressions of pain or anger may coexist with emotional detachment, sarcasm, or coded language as a means of self-protection [8]. Incorporating sentiment analysis into the study of racial violence allows researchers to quantify emotional expression while also uncovering emotional patterns often missed in traditional analysis.

1.1.3. Racial Violence and Emotion Analysis

While sentiment analysis classifies text as positive, negative, or neutral, emotional analysis identifies specific emotions within textual data, such as fear, anger, sadness, or joy [17]. This approach offers a more nuanced view of emotional experiences, which is particularly useful for understanding complex reactions to racial trauma. Emotion analysis has been applied to examine emotions and the prominence of positivity in #BlackLivesMatter tweets, revealing that while anger and disgust were present, positive emotions such as hope and optimism were more prevalent in pro-BLM tweets and significantly associated with on-the-ground protests [19].

For Black women, emotion analysis is especially valuable in unpacking complex emotional responses shaped by trauma, strength narratives, and community solidarity. Their emotions may reflect not only personal pain, but also broader emotional states tied to cultural expectations, intergenerational resilience, and historical consciousness. This study places emotion at the center, acknowledging it as both a reaction and a form of resistance.

By combining self-reported emotional data with computational emotion analysis, this research captures both the internal and external aspects of Black women’s responses to racial violence. This dual approach allows for a deeper understanding of how emotions are experienced, expressed, and shaped by systemic forces.

1.2. Broadening the Metrics of Racial Violence

1.2.1. Racial Violence Measurement

Traditional approaches to measuring racial violence have relied primarily on self-report instruments designed to capture individuals’ lived experiences with racism, discrimination, and microaggressions. Instruments such as the Racial and Ethnic Microaggressions Scale (REMS) [20] (pp. 57–66), the Everyday Discrimination Scale (EDS) [21] (pp. 335–351), and the Index of Race-Related Stress (IRRS) [22] (pp. 149–167) assess the frequency, type, and perceived impact of racialized experiences. These scales have been critical in demonstrating how racial violence, both overt and covert, correlates with adverse physical and mental health outcomes [23] (pp. 407–440).

However, these tools often focus on the occurrence of discriminatory events rather than the emotional aspects of racial violence. They also rely on static data, which may not fully capture the real-time nature of racialized experiences. As such, there is a growing need to expand how racial violence is measured to not only account for its frequency but also for the complex emotional and psychological responses it evokes.

1.2.2. Racial Violence and Sentiment Analysis Measurement

Sentiment analysis offers an innovative approach to improve traditional survey-based measurement. By analyzing a text’s emotional valence (positive, negative, or neutral), sentiment analysis can uncover how individuals and communities respond effectively to racialized events [11]. For example, public responses to police shootings, racial profiling, and systemic injustice can be examined through large-scale analysis of social media posts, news commentary, and even open-ended survey responses.

Unlike static survey instruments, sentiment analysis allows for real-time, dynamic insights into public and personal sentiment, potentially revealing shifts in emotional tone as events unfold [24]. It also captures nuances often missed by Likert scales—for instance, identifying specific tones of anger, despair, or solidarity across different contexts [24]. By applying sentiment analysis to narratives or self-reported texts, researchers can detect patterns of emotional expression that reflect broader emotional climates and the psychosocial impact of racial violence [25,26] (see [26], pp. 110–121).

1.2.3. Racial Violence and Emotion Analysis Measurement

While sentiment analysis broadly classifies tone as positive, negative, or neutral, emotion analysis goes further by identifying specific emotional states such as fear, sadness, anger, joy, or disgust [17]. This approach is particularly valuable in measuring racial violence because it acknowledges the depth and complexity of emotional responses, especially among historically marginalized groups [19].

Emotion analysis has recently gained traction in studies of trauma, protest, collective grief, and social justice [19,27,28] (see [27], pp. 139–164; [28], pp. 69–89). In the context of racial violence, it can reveal how individuals experience not just generalized negativity, but specific affective responses like rage, exhaustion, or numbness, emotions often tied to trauma or resilience [29]. Incorporating emotion analysis into the measurement of racial violence allows researchers to examine not only whether someone is emotionally affected, but how they are affected and what those patterns suggest about coping, desensitization, or emotional resistance.

For Black women, whose responses are often shaped by intersecting systems of oppression, emotion analysis can highlight the unique emotional burden of gendered racism. By expanding measurement beyond frequency and severity into the emotional aspect, researchers can generate a richer, more humanizing understanding of the toll racial violence takes.

1.2.4. MAACL-R and Sentiment Analysis

The Multiple Affect Adjective Checklist-Revised (MAACL-R) is a well-established psychological instrument designed to assess emotional states such as anxiety, depression, hostility, positive affect, and sensation seeking [30]. Although traditionally used in clinical or experimental contexts, integrating MAACL-R data with sentiment analysis represents an innovative research approach.

This dual approach allows researchers to align self-reported emotional states with computationally derived sentiment scores, providing subjective and objective insights into emotional expression. For example, if a participant reports high hostility on the MAACL-R but uses emotionally neutral or positive language in their responses, that contrast may signal emotional suppression or masking. Conversely, alignment between negative MAACL-R scores and sentiment results could validate the emotional intensity reported.

Using sentiment analysis along with the MAACL-R integrates the qualitative and quantitative domains. Furthermore, it allows researchers to capture not just the presence of emotional distress but its expression in language. This integrative strategy strengthens the measurement of racial violence and offers a more layered understanding of emotional responses.

1.2.5. MAACL-R and Emotion Analysis

Incorporating emotion analysis alongside the MAACL-R further expands the complexity of emotional assessment. While the MAACL-R captures predefined categories of affect, emotion analysis can identify a wider range of context-dependent emotions directly from participant language [30,31] (see [31], pp. 327–332). This allows researchers to compare structured checklist responses and with emotional expressions, highlighting where they match or differ.

For instance, emotion analysis might detect expressions of grief, pride, or numbness, emotions not explicitly measured by the MAACL-R but deeply relevant to racial trauma and violence [32] (pp. 15129–15215). Analyzing this alongside checklist responses creates opportunities to explore how cultural norms, trauma responses, or sociopolitical contexts shape emotional expression.

This complex approach is particularly important in the study of Black women’s responses to racial violence, as it honors the complexity of their emotional experiences while also enhancing methodological rigor. By combining the MAACL-R with emotion analysis, researchers can more accurately measure the emotional and psychological impact of racial violence in ways that are both empirical and culturally grounded.

1.3. Sentiment Analysis and Racial Violence

Sentiment analysis has become a widely used technique in social science research for quantifying subjective emotional content in textual data [33]. It offers researchers a scalable way to examine emotional expression across diverse populations and platforms [34]. Early studies demonstrated how sentiment analysis could highlight public responses to societal issues, contributing to discussions around decision-making and social climate [35]. Similarly, sentiment analysis was applied to detect hate speech in tweets, offering insights into the prevalence and evolution of online hostility [36]. Advances in hate speech detection have further emphasized the role of sentiment modeling in understanding the linguistic patterns of racial animosity [26].

Historically, rule-based approaches like VADER (Valence Aware Dictionary for sEntiment Reasoning) have been popular for their ease of use and performance on short, informal texts [37] (pp. 216–225). These models rely on predefined lexicons and heuristics to infer sentiment from text and have been instrumental in tracking emotional shifts following mass trauma events, such as mass shootings [38]. However, while rule-based methods provide broad sentiment classifications, they often lack the contextual depth necessary to fully capture the nuanced emotional responses produced by complex social phenomena like racial violence, particularly among marginalized populations.

This study employs a transformer-based machine learning model, such as RoBERTa (a robustly optimized BERT variant), to address these limitations for sentiment analysis. These models surpass rule-based approaches by leveraging contextual embeddings and deep neural architectures that consider the text’s complete syntactic and semantic structure. As a result, they can more accurately interpret implicit sentiment, sarcasm, and culturally specific expressions, which are critical when analyzing emotionally charged narratives around racial trauma.

Applying RoBERTa-based sentiment classification to self-reported emotional responses from Black women in Missouri allows this study to uncover not only the emotional tone (positive, negative, or neutral) of their responses but also the subtle affective cues reflected in their reported emotions. This approach builds on recent research examining the psychological impact of racial violence, which highlights a range of emotional reactions from anger to sadness to numbness and resilience [2,39] (see [39], pp. 334–358). By incorporating machine learning models, this study captures the complex emotional tone of responses that might otherwise be oversimplified or misclassified by lexicon-based tools.

Furthermore, this approach enhances our understanding of how cultural, psychological, and contextual factors influence sentiment expression. Individuals may report positive or neutral sentiments about adverse events due to psychological adaptation or sociocultural expectations [40]. This suggests that sentiment is not merely a reaction but also a reflection of coping processes, meaning-making, and identity. Using transformer-based models enables this study to detect these complex expressions more reliably, advancing both the methodological and conceptual understanding of racial violence.

1.4. Emotion Analysis and Racial Violence

While sentiment analysis broadly categorizes emotional tone, emotion analysis provides a more granular understanding by identifying discrete emotions such as anger, fear, sadness, and joy [17]. This distinction is crucial in the study of racial violence, where emotional responses are often complex and deeply contextual. Emotion analysis allows researchers to uncover the specific emotional states that emerge in response to racialized experiences, shedding light on the psychological processes that underlie reactions such as grief, rage, fatigue, and hope [17].

Emerging work in Natural Language Processing (NLP) has applied emotion classification to textual data from social movements, protests, and responses to traumatic events [12,41]. However, few studies have applied this methodology to examine the emotional experiences of Black women in the context of racial violence, despite the longstanding calls in Black feminist and trauma literature to center their narratives and emotional realities [6,7] (see [6], pp. 132–133; [7], pp. 314–343).

This study addresses this gap by applying a fine-tuned emotion classification model based on RoBERTa to open-text responses collected through the Multiple Affect Adjective Checklist-Revised (MAACL-R). This model, where trained on emotion-labeled corpora, is capable of identifying nuanced emotional states from participant responses [42]. This enables a deeper examination of how Black women in Missouri convey their emotional responses to racial trauma and how those responses reflect broader patterns of coping, resistance, and emotional regulation.

Emotion analysis complements the MAACL-R by offering insights into how emotions are expressed in natural language, beyond structured checklists. While the MAACL-R provides validated self-reported measures of affect (e.g., anxiety, hostility, depression), machine learning-based emotion classification captures the expressed emotion in participants’ words, highlighting whether participants’ reported feelings match or differ from how their emotions are expressed [30,43]. This dual perspective offers a deeper and more realistic representation of emotional experience.

This study explores the emotional responses of Black women in Missouri to racial violence, a state marked by a long-standing history of racial injustice, systemic inequity, and police brutality [44,45] (see [45], pp. 421–441). By focusing on Black women, whose experiences are shaped by both racial and gendered oppression, this study offers a deeper understanding of how intersecting forms of oppression are emotionally processed within a highly racialized environment. Through applying sentiment and emotion analysis on MAACL-R responses, this study captures the affective patterns of participants’ experiences. It reveals not only the prevalence of negative emotions such as fear, disgust, and sadness but also the presence of resilience, hope, and strength. These insights contribute to a more comprehensive and culturally grounded understanding of how Black women emotionally navigate and interpret racial violence in Missouri. Guided by two central questions, what sentiments and emotions are expressed, and what psychosocial factors influence positive emotional responses, this research sheds light on the broader psychological impacts of racial violence on Black women in a region representative of national racial tensions.

1.5. Theorizing Emotional Reactions to Racial Violence

Understanding the emotional responses of Black women to racial violence requires a theoretical framework that acknowledge the historical, systemic, and intersectional nature of oppression. How Black women in Missouri respond emotionally to incidents of racial violence can be better understood through the lens of racial trauma theory [46] (pp. 13–105). Racial trauma theory offers a valuable approach for examining the emotional impact of racial violence on Black women in Missouri, as it helps highlight the range of emotions these women may experience and express in response to such events.

Racial Trauma Theory

Racial trauma theory, which is a part of the broader concept of trauma theory, refers to the idea that being exposed to racism over time can lead to psychological and emotional harm in individuals who experience it [47] (pp. 675–687). This trauma stems from both direct experiences, such as hate crimes or racial discrimination, or indirect experiences, such as witnessing acts of racism or experiencing racial microaggressions [48] (pp. 1–5). These experiences contribute to a compounded stress burden that can affect mental health, leading to symptoms such as anxiety, depression, hypervigilance, and post-traumatic stress disorder (PTSD) [29]. In contrast to other types of trauma, racial trauma is different because it is maintained by systems and structures that uphold racial inequity, making it persistent and impossible for many individuals to escape [49] (pp. 1849–1863).

Hate crimes, defined as crimes motivated by prejudice based on race, religion, sexual orientation, gender, and other identities, remain a persistent issue [50]. In 2023, the Federal Bureau of Investigation (FBI) reported approximately 11,862 hate crimes involving 13,829 offenses [51]. Of these, 64.5% were motivated by bias against race, ethnicity, or ancestry, making race the leading factor in hate crimes [51]. Black individuals were the most frequently targeted racial group, accounting for over half of all race-based crime victims [51]. These numbers highlight how widespread racial violence is and its disproportionate impact on Black communities.

Racial discrimination also plays a significant role in racial trauma. It refers to attitudes, behaviors, or policies that serve to either (1) keep physical distance between racially privileged and racial underprivileged groups or (2) ensure that individuals with marginalized racial identities stay on the margins of society [21]. Unlike overt acts of hate, racial discrimination is often systemic and subtle, which can be just as harmful over time. These include denial of opportunities, racial profiling, and discriminatory practices that reinforce feelings of isolation and otherness [52].

For Black women in Missouri, the combined impact of overt and systemic racism is worsened by the state’s history of racial violence and ongoing racial inequalities [53]. Understanding Racial Trauma Theory is crucial for examining the psychological impacts of racial trauma, particularly for Black women in Missouri, who face unique stressors and emotional responses from ongoing exposure to racism and violence in their communities. By situating racial trauma within the context of Black women’s lived experiences, this study aims to highlight the specific ways in which racism and violence manifest as chronic stressors that shape Black women’s emotional responses. Our study is guided by two key research questions: Our study is guided by two key research questions:

• As measured through self-reported data and computational analysis, what sentiments and emotions do Black women in Missouri express regarding incidents of racial violence?
• What contextual or psychosocial factors may account for the expression of positive sentiments and emotions regarding incidents of racial violence?

2. Materials and Methods

The data utilized in this study are drawn from the ongoing Black Families and Racial Justice Study (BFRJS), a three-year longitudinal project conducted by blind authors. Initiated in 2022, the BFRJS follows an initial cohort of roughly 700 Black families, including both parents and adolescents, across the state of Missouri. Designed with a strong community focus, the study includes participants from a variety of geographic settings, including rural, urban, and suburban areas. The sample is also socioeconomically diverse, encompassing families across different income levels and educational backgrounds. This diversity enables a more nuanced examination of how factors such as income, education, and geographic context shape family experiences and outcomes.

2.1. Participants

The present study examines the first wave of the BFRJS dataset. A total of 384 Black women were in wave 1 of the study. The median household income level for the women was, on average, USD 25,000 to USD 50,000 per year, and their average educational attainment fell between some college and an associate’s degree (M = 4.28, SD = 0.69).

2.2. Procedure

In the present study, the initial phase included Black women who participated by completing surveys either online or in person. Each session required approximately 45 min to one hour and included demographic questionnaires alongside surveys centered on their racialized experiences. Participants provided informed consent prior to participation. This study (#202112032) received approval from the Institutional Review Board (IRB). Recruitment occurred over a six-month period as part of a three-year research initiative. Recruitment materials were distributed through local institutions, social service organizations, and Black community-based spaces. Participants had the option to access the surveys electronically or request physical copies. Outreach efforts were intentionally community-centered, with the support of local community leaders to encourage engagement.

2.3. Measures

2.3.1. Demographics

Participants were asked to provide socio-demographic information, including age, educational attainment, and total household income. Descriptive statistics, including means and standard deviations for these variables, are presented in

Table 1. Descriptive statistics for participant demographic variables.

Demographic Variables	M	SD
Age	37.04	3.835
Education 1	4.28	0.693
Income	2.14	0.516

¹ Note. Education = associate’s degree or some college education. Income = USD 25,000 to USD 50,000 per year.

.

2.3.2. Racial Violence

In this study, participants were asked to reflect on how incidents of racial violence made them feel. Participants responded to the following prompt: “Below are some feelings that parents/people may feel in response to racial injustices. Choose the words that describe how you feel in response to racial violence against Black Americans (for example, police brutality, hate crimes, etc.). Racial violence can bring up a lot of different emotions—we want you to check all the words that describe your feelings?” Emotional responses to racial violence were assessed using the Multiple Affect Adjective Checklist-Revised (MAACL-R), a validated self-report measure commonly used to evaluate emotional states in response to external stressors [30]. This instrument includes three subscales capturing negative affect: anxiety, depression, and hostility, with internal consistency reliability (α) ranging from 0.70 to 0.92 [30].

2.3.3. Sentiment Analysis

Sentiment analysis was conducted using the pre-trained RoBERTa model “cardiffnlp/twitter-roberta-base-sentiment” via the Hugging Face Transformers library [54]. Text data from the MAACL-R responses were first cleaned and the relevant adjective column was selected. The RoBERTa sentiment pipeline was used to classify each response as positive, neutral, or negative. To evaluate model performance, a benchmark dataset was created by having research team members independently label the adjectives using a majority vote system. An initial evaluation of 48% accuracy and a class imbalance within the sentiment data prompted a 5-fold cross-validation procedure. For each fold, the data were split into stratified training and test sets. However, the model’s training was impacted by this imbalance, with neutral sentiments being severely represented. Although oversampling or data augmentation techniques were considered, they were not implemented to preserve the natural distribution of emotional responses. This decision reflects the real-world prevalence of highly negative sentiment in participants’ responses to racial violence.

The RoBERTa model was fine-tuned on the training set and evaluated on the test set using the Hugging Face Trainer API. Evaluation metrics, including accuracy, precision, recall, and F1-score, were calculated using scikit-learn version 1.3.2, and a confusion matrix was generated to assess model performance. Following cross-validation, the final model was retrained on the full labeled dataset and used to generate sentiment predictions and confidence scores for all responses.

2.3.4. Emotion Analysis

Emotion analysis was performed using the pre-trained model “bhadresh-savani/bert-base-go-emotion”, which is fine-tuned on the GoEmotions dataset [55]. The same cleaned text data used in the sentiment analysis were input into the model using the Hugging Face pipeline(“text-classification”, model=” bhadresh-savani/bert-base-go-emotion”) utility. Each observation yielded a list of emotion scores, and the emotion with the highest probability was stored for further analysis. This approach enabled multi-label classification of discrete emotions such as sadness, anger, or joy, and provided additional detail into how participants emotionally responded to racial violence.

2.3.5. RoBERTa

Both sentiment and emotion models were implemented using the Hugging Face Transformers library, executed on GPU when available. For tokenization and classification, the RobertaTokenizer and model-specific pipelines were applied directly to raw text without additional preprocessing (e.g., stopword removal or lemmatization). Sentiment labels were converted into numeric classes for evaluation, and the model’s performance was assessed using scikit-learn metrics. The use of RoBERTa enabled high-quality classification of emotionally complex responses due to its ability to capture contextual embeddings from unstructured text. GenerativeAI (OpenAI’s ChatGPT—Version 4o) was utilized to assist with code troubleshooting.

3. Results

Sentiment and emotion analysis was conducted using the RoBERTa model, with a particular focus on identifying positive sentiments within participants’ responses. Of the responses analyzed, the majority were classified as negative (n = 334), followed by positive (n = 48), and neutral (n = 2). Figure 1 presents the distribution of sentiment categories identified in the sentiment analysis of participants’ responses. The majority of responses were classified as negative (87%), followed by positive (12.5%), and neutral (0.5%) sentiments.

Figure 2 shows the distribution of primary emotions expressed by participants’ responses to racial violence. The most frequently reported emotions were fear (22.1%), disgust (21.1%), and sadness (16.4%), highlighting a predominance of negatively valenced emotional experiences. In contrast, positively valenced emotions such as approval, surprise, and caring were expressed far less frequently, each accounting for fewer than 1.5% of responses.

Further analysis of the emotion categories reveals the distribution of emotions within each sentiment classification (negative, neutral, and positive) based on participants’ responses. As shown in Figure 3, the negative sentiment category was primarily composed of fear, sadness, and anger. The neutral sentiment category was entirely represented by the “neutral” emotion label. In contrast, the positive sentiment category reflected a broader range of positively valenced emotions, including joy, gratitude, admiration, and love. This breakdown underscores the complexity of emotional expression within each sentiment type.

Figure 4 displays the confusion matrix averaged over five-fold cross-validation showing the performance of the sentiment classification model. The model demonstrated strong accuracy in identifying negative (52 correct) and positive (45 correct) sentiments, while neutral sentiments were frequently misclassified as either negative or positive. Specifically, neutral responses were never correctly predicted and were most often confused with negative (10 instances) and positive (10 instances) sentiments. The misclassification of neutral responses suggests that the model had insufficient exposure to neutral sentiment during training, likely due to its underrepresentation in the dataset.

In addition to the confusion matrix, evaluation metrics were averaged across five folds of cross-validation to assess overall model performance. The sentiment classification model achieved an average accuracy of 74.1%, with a precision of 64.98%, recall of 74.1%, and F1 score of 68.66%. Fold-by-fold performance is provided in

Table 2. Evaluation metrics across five-fold cross-validation.

Fold	Accuracy	Precision	Recall	F1 Score
1	0.666667	0.641667	0.666667	0.633222
2	0.692308	0.609391	0.692308	0.648070
3	0.807692	0.686391	0.807692	0.740602
4	0.807692	0.686391	0.807692	0.740602
5	0.730769	0.625000	0.730769	0.670330
Average	0.741026	0.649768	0.741026	0.686565

, showing consistent results across iterations. These metrics suggest moderate-to-strong performance in identifying sentiment, although precision remains somewhat lower. This indicates some misclassifications, particularly for neutral responses.

4. Discussion

This study examines psychosocial factors that may contribute to positive emotional responses to racial violence among Black women in Missouri, with a focus on coping mechanisms, racial identity beliefs, spirituality and religiosity, and resilience and strength. Our findings align with prior research that utilizes sentiment and emotion analysis to uncover nuanced social dynamics. For example, one study demonstrated the value of analyzing differential public opinions through advanced sentiment analysis to effectively detect racist content on social media platforms [56] (pp. 9717–9728).

Similarly, an emotion analysis of #BlackLivesMatter tweets emphasized the prominence of positivity in the context of protest-related discourse [19]. Using domain-adapted neural models, they analyzed a large dataset to trace the prevalence and patterns of emotions such as anger, disgust, and positivity. Their work demonstrated the utility of few-shot emotion classification in capturing the emotional dynamics of social movements. Building on these methodological insights, our study seeks to deepen understanding of the complex emotional responses of Black women to racial violence.

However, the presence of positive sentiments and emotions in response to racial violence suggests the influence of underlying psychosocial mechanisms that may not be immediately visible. These mechanisms may include coping strategies that alleviate psychological trauma, racial identity beliefs that reinforce a sense of self and belonging, spirituality or religiosity that offers emotional support and meaning, and resilience that fosters strength and empowerment in the face of adversity. Together, these factors may help explain the emergence of positive emotional responses in such contexts. This leads us to one of the central questions of our study: What contextual or psychosocial factors contribute to the expression of positive sentiments or emotions in response to racial violence?

Although the prevalence of negative and neutral sentiments and emotions is expected given the context of racial violence, the emergence of positive sentiments and emotions presents a compelling dimension that merits deeper exploration. This phenomenon highlights the importance of investigating the underlying contexts and psychosocial processes that may give rise to such responses. In light of this complexity, our study explores several contributing factors, including trauma-related coping mechanisms, racial identity beliefs, spirituality and religiosity, and resilience and strength.

4.1. Negative Emotions and Racial Violence Among Black Women in Missouri

The widespread presence of negative emotions among Black women in response to racial violence is a critical finding in this study. These emotions, captured through sentiment and emotion analysis, are deeply rooted in the historical and ongoing realities of systemic racism and its disproportionate impact on Black communities [18]. Traumatic incidents involving police brutality and racialized violence, such as the deaths of Sonya Massey, Breonna Taylor, Atatiana Jefferson, Sandra Bland, and Michelle Cusseaux, have reinforced collective feelings of fear, anger, distress, and despair among Black women. This aligns with research underscoring the detrimental effects of police brutality on both mental and physical health [57] (pp. 1113–1122).

Importantly, these emotional responses are not isolated or momentary; instead, they reflect the compounded trauma of persistently navigating a racially oppressive society [58] (pp. 466–485). As existing literature shows, chronic exposure to racial stress is linked to serious health consequences, including psychological distress, hypertension, and cardiovascular disease [59,60,61] (see [61], pp. 800–816). For Black women, this burden is uniquely intensified by intersecting roles, as individuals enduring racism themselves and, often, as caretakers or advocates within their families and communities. The emotional toll of confronting racial violence is compounded by the responsibility of shielding loved ones and preserving a sense of stability among ongoing sociopolitical threats [62,63] (pp. 457–467).

4.2. Neutral Emotions and Racial Violence Among Black Women in Missouri

Although less frequent than negative emotions, neutral emotional responses among Black women in the context of racial violence reveal a complex and often overlooked dimension of the emotional experience. These expressions of neutrality may reflect avoidant coping strategies and psychological mechanisms used to shield oneself from the emotional toll of persistent racial trauma [64] (pp. 609–617). Such coping can manifest as emotional detachment or a muted affective response to racially charged events [64] (pp. 609–617). For Black women, adopting a neutral stance may serve as a protective buffer, helping to maintain emotional stability while navigating ongoing exposure to racial violence. This form of neutrality may also be indicative of emotional numbing, a response commonly observed in communities facing chronic trauma [65] (pp. 207–224).

In this context, neutrality may not signify indifference but rather a psychological adaptation that enables Black women to preserve their mental health and fulfill daily responsibilities in the face of unrelenting stress. However, these neutral responses carry broader social implications. When misinterpreted as emotional disengagement or lack of concern, they can obscure the actual psychological burden of racial violence and potentially hinder collective action against systemic injustice. Therefore, recognizing and understanding the role of neutral emotional responses is essential in capturing the full spectrum of Black women’s lived experiences and crafting effective, culturally informed strategies for support and healing.

4.3. Positive Emotions and Racial Violence Among Black Women in Missouri

The presence of positive emotions in response to racial violence, while unexpected, offers valuable insight into the resilience and adaptive coping strategies employed by Black women. These expressions of positivity may arise from several sources, including a profound sense of solidarity and community support that often surfaces in the wake of collective trauma [66,67] (see [66], pp. 240–256). Positive emotional responses may also reflect hope and a commitment to social change, as many Black women engage in activism, advocacy, and mutual care efforts aimed at advancing racial justice [68] (pp. 519–533). Furthermore, the concept of post-traumatic growth provides another lens through which to interpret these responses, suggesting that some Black women may experience increased inner strength, meaning-making, and empowerment in the aftermath of racial trauma [69] (pp. 390–393). For these women, shared experiences of injustice may deepen community bonds and reinforce a collective determination to challenge systemic oppression.

However, it is essential to situate these positive responses within a broader emotional experience characterized mainly by negative and neutral sentiments. The dominant presence of distress and emotional numbing underscores the deep psychological impact of racial violence on Black women’s well-being. While the emergence of positive emotions is significant and contributes to a richer, more complex understanding of emotional responses, it should not be interpreted as diminishing the severity of the harm experienced. A comprehensive approach to supporting Black women must recognize the full emotional spectrum, from trauma and pain to hope and resilience and address both the psychological consequences and structural roots of racial violence.

4.4. Examining Factors Contributing to Positive Emotions Among Black Women in Missouri

4.4.1. Coping Mechanisms

Coping mechanisms can play a crucial role in shaping how Black women emotionally respond to racial violence. Strategies such as cognitive reappraisal, meaning-making, emotional regulation, and community-based coping allow individuals to manage the psychological toll of trauma while maintaining a sense of control [70,71] (see [70], pp. 310–331; [71], pp. 58–74). For some Black women, these coping strategies may foster emotional detachment from distressing events or facilitate a re-framing of trauma into narratives of strength, survival, and resistance. Through these mechanisms, positive emotions such as hope, pride, or gratitude may emerge, not because the trauma is diminished but because the coping process enables emotional transformation in the face of adversity.

4.4.2. Racial Identity Beliefs

Racial identity beliefs serve as a powerful foundation for fostering positive emotional responses to racial violence. A strong sense of racial identity can buffer against the internalization of negative societal messages and cultivate pride, cultural affirmation, and belonging [72] (pp. 384–400). This identity may serve as a source of empowerment for Black women, connecting individual experiences to a broader historical and communal struggle for justice. By embracing a collective identity rooted in strength and resilience, Black women may respond to racial violence not only with grief or anger but also with pride in their heritage, solidarity with others, and a determination to uplift and protect their community.

However, it is also important to consider the emotional implications of internalized racism and colorblind ideology, which function in contrast to a positive racial identity. Internalized racism, or the acceptance of negative societal stereotypes about one’s racial group, can diminish self-worth and distort emotional processing, potentially leading to disconnection or emotional suppression [72] (pp.384–400). Similarly, colorblind ideology, which involves the denial or minimization of race and racism, often undermines the salience of racial group membership and discourages acknowledgment of systemic harm [73] (pp. 258–275). Research shows that individuals who endorse colorblind or power-evasive beliefs are less likely to express empathy for racially marginalized groups and more likely to uphold anti-Black prejudice [73] (pp. 258–275).

4.4.3. Spirituality and Religiosity

Spirituality and religiosity are often central to the lives of many Black women and can provide a major source of emotional strength in the face of racial violence. Faith-based frameworks offer tools for making sense of suffering, envisioning justice beyond the present moment, and sustaining hope [74] (pp. 65–84). Spiritual beliefs may allow Black women to view their experiences within a larger divine narrative, reinforcing the idea that justice will ultimately prevail and that their pain is not in vain. Religious practices such as prayer, worship, and communal gatherings also offer emotional release, comfort, and connection, contributing to feelings of peace, hope, and collective purpose during times of crisis [74] (pp. 65–84).

4.4.4. Resilience and Strength

Resilience and strength are frequently cited as defining characteristics of Black women’s responses to racial violence, often enabling them to endure and navigate systemic violence with perseverance and purpose [75] (pp. 395–404). These traits are shaped by historical and intergenerational survival experiences, collective resistance, and caregiving within structurally oppressive environments. However, the cultural narrative surrounding the Strong Black Women (SBW) schema adds complexity to how these traits are experienced and expressed. The SBW schema characterizes Black women as inherently resilient, self-sacrificing, emotionally restrained, and unwaveringly strong, with expectations that can both empower and constrain [76] (pp. 89–98). On the one hand, this internalized ideal may fuel a sense of pride, agency, and communal responsibility, contributing to expressions of strength, dignity, and optimism even in the face of racial violence [77] (pp. 668–683).

On the other hand, adherence to the SBW ideal can discourage vulnerability, emotional openness, and help-seeking behaviors, potentially leading to emotional suppression and psychological strain [78] (pp. 1–16). In this context, positive emotional responses may reflect a strategic affirmation of strength rooted in cultural survival, but they may also obscure unacknowledged pain. Understanding how resilience and strength function within and beyond the SBW schema is critical for interpreting Black women’s emotional expressions and developing support frameworks that affirm their complexity while challenging the pressures of emotional invulnerability.

4.5. Public Health and Clinical Implications

The findings of this study have important implications for public health practice, particularly in mental health service delivery for Black women. Given the prominence of emotional suppression and limited emotional expression in participants’ responses, clinicians and community health practitioners must be trained to recognize culturally specific expressions of distress that may not align with traditional diagnostic criteria. Emotionally neutral or seemingly positive affect may mask deep psychological strain rooted in racial trauma. As such, culturally competent care should not only include trauma-informed practices but also integrate racial trauma frameworks that acknowledge the historical and systemic nature of oppression.

Community-based mental health interventions, such as healing circles, culturally grounded group therapy, and partnerships with Black-led community organizations, may offer more accessible and affirming spaces for Black women to process and express their emotional experiences. Increasing representation among mental health providers, incorporating spiritual and communal strengths into care, and developing targeted screening tools that account for emotion masking are critical for reducing disparities in mental health outcomes.

5. Limitations

Several limitations should be acknowledged when interpreting the findings of this study. First, the emotional responses analyzed were derived from a checklist-based self-report measure (MAACL-R). While this approach allowed for the integration of psychometric and computational techniques, the use of predefined adjectives may have limited participants’ emotional expression, potentially excluding more complex or culturally specific feelings not captured by the checklist.

Second, although transformer-based models like RoBERTa offer advanced contextual understanding, they are still limited by the characteristics of the datasets on which they were pre-trained and fine-tuned. While our dataset consisted exclusively of Black women’s responses, the emotion classification models may have been developed using corpora where Black women’s linguistic patterns, emotional expressions, or cultural references were underrepresented. As a result, certain subtleties in emotional expression, such as emotion masking or culturally embedded meanings, may have been misinterpreted or overlooked. Although code-switching was not present in this dataset due to its structured format, future work involving open-ended or narrative responses should consider that code-switching and culturally specific language can present additional challenges for computation models.

Third, model performance was affected by a class imbalance in the dataset. Specifically, the number of negative responses (n = 334) far exceeded the number of positive (n = 48) and neutral (n = 2) responses. This imbalance likely impaired the model’s ability to accurately detect underrepresented sentiment categories, particularly neutral expressions. As evidenced by the confusion matrix, all neutral responses were misclassified. While the imbalance reflects the real emotional perspective of Black women in Missouri responses to racial violence, it poses challenges for model generalizability and highlights the need for future research to explore advanced techniques such as class weighing, synthetic oversampling, or ensemble methods to enhance sensitivity to underrepresented classes. Lastly, this class imbalance suggests a limitation in the model’s ability to identify emotional uncertainty or suppression, key components in understanding responses to racial trauma.

Fourth, while the discussion explores psychosocial factors such as coping mechanisms, racial identity beliefs, and spirituality to contextualize emotional responses, these variables were not directly measured in the present analysis. Their inclusion is grounded in prior literature, but future studies should incorporate direct assessments of these constructs to validate their influence on emotional expression empirically.

Lastly, the focus on Black women in Missouri provides important region-specific insight but may limit generalizability to Black women in other states or sociopolitical environments. Missouri’s specific history and contemporary racial dynamics may uniquely show how racial violence is experienced and expressed emotionally. Broader geographic replication is needed to understand how context influences these emotional patterns.

6. Conclusions

This study examined the emotional responses of Black women in Missouri to incidents of racial violence using an integrated approach that combined self-reported data from the MAACL-R with sentiment and emotion analysis via transformer-based models. The findings highlight the predominance of negative emotional responses, such as fear, disgust, and sadness, while also highlighting less frequent but meaningful expressions of positive sentiment, including joy, admiration, and love. These emotional patterns reflect not only the psychological toll of racial trauma but also the complex coping strategies and cultural strengths that shape how Black women navigate racial violence.

By incorporating racial trauma theory and centering the lived experiences of Black women, this study offers a complex understanding of how systemic oppression manifests in emotional expression. The presence of positive emotions among racial violence points to the influence of psychosocial factors, such as racial identity, spirituality, and resilience, that may foster emotional resistance and post-traumatic growth. At the same time, the emergence of neutral or emotionally muted responses draws attention to the coping mechanisms used to shield against chronic distress, such as emotional suppression or detachment.

Methodologically, this research demonstrates the value of integrating psychometric tools with computational techniques to analyze emotional responses in a culturally relevant and empirically grounded manner. While challenges remain, particularly in capturing emotional complexity through pre-trained models and addressing class imbalances, this study contributes to a growing body of scholarship that uses Natural Language Processing to explore racial violence and emotional health in marginalized communities.

Ultimately, this work highlights the importance of centering Black women’s emotional realities in both research and policy. Future studies should extend this analysis across geographic regions, incorporate qualitative narratives to complement computational findings, and directly measure psychosocial factors that shape emotional outcomes. In doing so, researchers and practitioners alike can better understand, support, and advocate for the emotional well-being of Black women confronting systemic racial violence.