Self-perceived Accent

Abstract

The Self-perceived Accent Scale (SPAS), developed by social psychologists John F. Dovidio and Agata Gluszek (Dovidio et al., 2010; Gluszek & Dovidio, 2010), is a psychometric instrument designed to capture an individual's subjective appraisal of their own spoken accent relative to a dominant linguistic standard. While objective sociophonetic evaluations measure acoustic features such as voice onset time, vowel formant trajectories, and prosodic variations, the SPAS evaluates the phenomenological and psychological reality of speaking with a non-native or non-standard accent. The scale comprises an initial categorical screening item with a qualitative open-ended specification followed by five core dimensional items presented along a 9-point bipolar semantic differential continuum. These items assess perceived phonetic distance from the standard American accent, self-appraised nativeness, ease of understanding (intelligibility), communicative interference, and self-perceived spoken English competence. Across non-native English-speaking immigrant populations, international students, and domestic ethnolinguistic minorities, the instrument demonstrates robust psychometric properties, featuring high internal consistency (typically ranging from Cronbach's $\alpha = .86$ to $.91$), a robust single-factor or tightly integrated hierarchical factor structure, and strong convergent validity with measures of perceived discrimination, linguistic stigma consciousness, intergroup communication apprehension, and diminished psychological sense of belonging. The instrument has become a foundational tool in sociolinguistics, cross-cultural psychology, and intergroup relations, shedding light on the mediating role of linguistic self-concept between external societal bias and individual psychological well-being.

Keywords

Self-perceived Accent, Linguistic Stigma, Foreign Accent, Sociolinguistics, Intergroup Relations, Communicative Competence, Social Belonging, Linguistic Profiling, Gluszek and Dovidio, Psychometrics, Acculturation, Language Discrimination

Authors

The Self-perceived Accent Scale was developed through collaborative empirical investigations conducted by researchers in the Department of Psychology at Yale University:

• John F. Dovidio, Ph.D.: Carl I. Hovland Professor Emeritus of Psychology and Professor of Public Health at Yale University. Internationally recognized for his seminal contributions to social psychology, implicit bias, aversive racism, and intergroup communication.
• Agata Gluszek, Ph.D.: Social psychologist whose doctoral and postdoctoral research at Yale University focused extensively on the social psychology of accents, linguistic diversity, social identity, and the psychological costs of communicative marginalization.
• Additional Collaborators (2010 Cohort): Michael John, Ruth Ditlmann, and Paul Lagunes, who collaborated with Dovidio and Gluszek on investigations regarding intergroup bias, dimensions of difference, and systemic social exclusion among Latino populations in the United States.

Purpose

The primary purpose of the Self-perceived Accent Scale is to operationalize and quantify the subjective experience of one's spoken accent within a specific sociocultural and linguistic context. Sociolinguistic and applied linguistic research has long demonstrated that accents are not merely acoustic phenomena; they function as potent social signifiers that cue categorical social identity, nationality, race, ethnicity, socioeconomic status, and geographic provenance (Baugh, 2003). In predominantly English-speaking societies such as the United States, divergence from what is colloquially termed “General American” or “Standard American English” frequently elicits conscious or subconscious linguistic profiling, negative stereotyping, and overt interpersonal or institutional discrimination.

Historically, phonetic research prioritized laboratory-based, objective assessments of accent strength, relying on acoustic phonetics (e.g., spectrograms, fundamental frequency analysis) or blinded evaluations by trained native-speaker raters. While these methodologies provide precision regarding acoustic divergence, they fail to capture the psychological salience of an individual's own perception of their voice. The SPAS addresses this methodological gap by focusing squarely on the speaker's metacognitive and affective evaluation of their spoken output. The rationale underlying the instrument is that an individual's cognitive appraisal of their accent—specifically whether they believe their speech diverges from the societal standard and whether they perceive it as an impediment to mutual comprehension—exerts an immediate influence on their social agency, communicative confidence, interpersonal anxiety, and global self-esteem.

In research contexts, the SPAS serves as a vital explanatory, mediating, or moderating variable in studies examining:

• Immigrant acculturation and adaptation trajectories.
• The psychological mechanisms linking linguistic diversity to perceived social exclusion and reduced institutional belonging.
• Stereotype threat and public speaking anxiety among foreign-born and bilingual professionals.
• Health disparities, particularly regarding how communicative self-efficacy affects interactions with healthcare providers among ethnolinguistic minorities.

In applied and clinical settings, such as university counseling centers, diversity, equity, and inclusion (DEI) diagnostics, and accent modification or communication coaching environments, the SPAS provides practitioners with a validated index of linguistic insecurity. It enables clinicians and coaches to distinguish between objective speech intelligibility challenges and internalized linguistic stigma, facilitating psychological interventions that target communicative self-efficacy rather than merely enforcing linguistic assimilation.

Psychological Construct

The central psychological construct measured by the SPAS is subjective accent appraisal, which encompasses the cognitive representations, self-evaluations, and behavioral efficacy beliefs an individual holds regarding their spoken English in relation to the societal standard. Although the five core dimensional items of the SPAS are frequently aggregated into a single composite score representing overall self-perceived accent strength and communicative impediment, an in-depth psychometric analysis reveals that the construct spans several interrelated dimensions:

1. Perceived Acoustic Divergence (Foreignness)

This facet captures the respondent's self-reported distance from the standard linguistic prototype. It is assessed via items examining the degree of “foreign accent” relative to a standard American accent and the extent to which the respondent believes they sound like a native English speaker. This dimension reflects an awareness of acoustic non-conformity. Rather than indicating communicative failure, it reflects the degree to which an individual views their vocal signature as carrying the phonetic markers of an outgroup.

2. Perceived Intelligibility and Decodability

Intelligibility involves the degree to which a speaker believes their speech sounds can be decoded and comprehended by interlocutors without cognitive strain. Assessed by items contrasting “extremely easy to understand” with “impossible to understand,” this dimension reflects the speaker's expectation of perceptual processing fluency in others. Cognitive and social psychological research demonstrates that when listeners experience low processing fluency due to an accent, they often misattribute this cognitive effort to the speaker's lack of competence or credibility. The SPAS assesses whether the speaker has internalized this dynamic.

3. Communicative Interference and Disruption

This dimension measures the extent to which the speaker perceives their accent as an active barrier in transactional and interpersonal dialogue. Assessed on a continuum from “does not interfere with communication” to “interferes extremely with communication,” this facet taps into functional impairment. A speaker may acknowledge possessing a prominent regional or foreign accent yet perceive zero communicative interference; conversely, an individual with subtle phonetic variations may perceive high interference due to heightened linguistic anxiety or experiences of interpersonal rejection.

4. Spoken Linguistic Self-Efficacy (Communicative Competence)

The final dimension evaluates the overarching self-concept of spoken language mastery, contrasting “competent English speaker” with “not a competent English speaker.” Grounded in Albert Bandura's self-efficacy framework, this dimension assesses the degree to which linguistic differences erode an individual's belief in their ability to articulate complex thoughts, negotiate meaning, and command authority in spoken discourse.

By capturing these multifaceted appraisals, the SPAS measures the intersection between phonological self-awareness and sociocognitive vulnerability. It operationalizes accent not as an isolated speech characteristic, but as a dynamic psychological construct that shapes how individuals navigate their social worlds.

Theoretical Framework

The conceptual foundation of the Self-perceived Accent Scale is rooted in several converging theoretical frameworks across social psychology, sociolinguistics, and intergroup communication:

Communication Accommodation Theory (CAT)

Formulated by Howard Giles and colleagues, Communication Accommodation Theory posits that individuals adjust their speech patterns—including accent, dialect, speech rate, and lexical choices—to signal social affiliation, psychological proximity, or social distance. When an individual perceives a substantial gap between their own accent and that of the dominant culture, it reflects perceived linguistic divergence. According to CAT, sustained non-accommodation or perceived inability to converge toward the prestigious linguistic norm can lead to psychological strain, communicative apprehension, and expectations of social censure.

Social Identity Theory and Ethnolinguistic Identity Theory

Under Social Identity Theory (Tajfel & Turner, 1986) and Ethnolinguistic Identity Theory (Giles, Bourhis, & Taylor, 1977), language and accent are core markers of ingroup identity and outgroup differentiation. Standard varieties of a language often represent institutional power, status, and societal legitimacy, whereas non-standard or foreign varieties are frequently stigmatized. The SPAS captures the extent to which an individual views their spoken identity as aligning with the societal ingroup or being relegated to outgroup status. Dovidio and Gluszek's framework suggests that individuals who perceive their accent as markedly divergent and disruptive experience lower levels of social belonging within the host society.

Stigma Consciousness and Social Identity Threat

The SPAS is deeply informed by research on social stigma (Crocker, Major, & Steele, 1998). Stigmatized individuals often develop a chronic awareness that their social identity may serve as a basis for devaluation. When non-native or dialectal speakers hold high self-perceptions of accent strength and communicative interference, their voice functions as a visible (or audible) stigma. This auditory stigma can induce stereotype threat during evaluative communicative tasks, increasing cognitive load, undermining executive control, and impairing performance.

Linguistic Profiling and Aversive Bias

Dovidio's broader theoretical model of aversive racism and subtle prejudice indicates that in modern egalitarian societies, bias often manifests through subtle, rationalizable mechanisms rather than blatant discrimination. Accent bias represents a socially tolerated form of discrimination; individuals often justify prejudice against accented speakers under the guise of “communication difficulties” or “clarity concerns.” Consequently, the SPAS evaluates an individual's internal representation of this societal dynamic, operationalizing how individuals navigate environments where their speech may trigger subtle, systemic biases.

Validity

The validity of the Self-perceived Accent Scale has been established across multiple studies involving diverse linguistic, cultural, and racial populations, including Latino Americans, Asian immigrants, and multinational student bodies.

Construct Validity

Construct validity is evidenced by the scale's capacity to measure subjective linguistic positioning. Studies show that scores on the continuous 5-item scale correlate strongly with objective sociodemographic markers of language acquisition. For instance, in Gluszek and Dovidio (2010), self-perceived accent scores were negatively correlated with the length of time an individual had resided in the United States and positively correlated with the age at which they began acquiring English. Individuals who learned English after puberty exhibited higher scores on foreignness and perceived communicative interference, aligning with critical period hypotheses in second language acquisition.

Convergent Validity

The SPAS demonstrates robust convergent validity with constructs related to intergroup functioning, psychological distress, and perceived marginalization:

• Perceptions of Discrimination: Across samples of Latino adults (Dovidio et al., 2010; Rivera, 2012), elevated SPAS scores correlated positively with self-reported experiences of institutional and interpersonal discrimination ($r$ values typically between $.30$ and $.45, p < .001$).
• Belonging and Social Inclusion: Gluszek and Dovidio (2010) demonstrated that higher self-perceptions of foreign accent were significantly associated with lower feelings of belonging in the United States, even after controlling for objective English language proficiency and general self-esteem.
• Psychological Distress and Microaggressions: In doctoral research conducted at Columbia University, Rivera (2012) utilized the SPAS to demonstrate that higher self-perceived accent scores significantly amplified the psychological impact of racial and ethnic microaggressions on self-reported depressive symptoms and somatic health outcomes among Latina/o Americans.

Discriminant Validity

Empirical evidence supports the discriminant validity of the SPAS by showing that it does not simply reflect general negative affectivity or low global self-worth. When correlated with standardized measures of global self-esteem (e.g., the Rosenberg Self-Esteem Scale), the correlations are generally small to non-significant ($r$ values ranging from $-.08$ to $-.18$), indicating that an individual can possess high global self-esteem while maintaining an accurate, localized perception of their non-native accent. Furthermore, the SPAS demonstrates divergence from general cognitive ability and formal written English literacy assessments, establishing that it specifically indexes spoken communicative appraisal.

Predictive and Criterion Validity

The scale effectively predicts behavioral outcomes, including avoidance of intergroup interactions, unwillingness to volunteer for leadership roles requiring public speaking, and communicative anxiety during clinical or legal encounters. In experimental and field paradigms, individuals with higher SPAS scores reported greater hesitation to engage in spontaneous conversations with native English speakers and expressed lower expectations of interpersonal warmth from conversational partners.

Reliability

The reliability of the Self-perceived Accent Scale has been documented across empirical studies using various indicators of internal consistency and temporal stability.

Internal Consistency

The 5 continuous dimensional items demonstrate high internal consistency across diverse empirical samples:

• In the foundational validation study by Gluszek and Dovidio (2010), which examined non-native English speakers residing in the United States from diverse linguistic backgrounds (including Spanish, Russian, Chinese, Korean, and French), the 5-item scale yielded a Cronbach's alpha of $\alpha = .86$.
• In Dovidio et al. (2010), analyzing perceptions of bias and exclusion among Latino adults, the internal consistency of the continuous accent scale reached $\alpha = .88$.
• In Rivera's (2012) study examining Latina/o American community members and university students, the scale achieved an internal consistency reliability coefficient of $\alpha = .91$.
• Corrected item-total correlations across these investigations consistently exceed $.60$, with items measuring perceived foreignness and ease of understanding frequently showing item-total correlations above $.75$.

Split-Half Reliability and Composite Reliability

Analyses evaluating split-half coefficients have shown Spearman-Brown corrected reliability indices exceeding $.85$. In structural equation modeling (SEM) applications, composite reliability (Raykov's rho or McDonald's omega) values regularly surpass $.88$, demonstrating that the scale exhibits minimal measurement error when modeling latent linguistic self-concept.

Test-Retest Stability

Although an individual's accent may undergo gradual changes over years of immersion or targeted phonetic training, accent self-perception remains stable over short-to-medium intervals. In longitudinal observational cohorts evaluated over 4-to-8-week intervals without speech interventions, the test-retest reliability coefficient was established at $r_{tt} = .82$ to $.87$, indicating robust temporal stability.

Factor Analysis

Both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) have been employed to evaluate the latent dimensionality of the 5 continuous items comprising the core SPAS assessment.

Exploratory Factor Analysis (EFA)

Initial principal components and principal axis factoring analyses conducted by Gluszek and Dovidio revealed a dominant single-factor solution. Scree plot analyses and Kaiser-Guttman criteria (eigenvalues > 1.0) consistently show a single primary factor accounting for between 62% and 74% of the total variance across items. Factor loadings on this primary unrotated factor are substantial:

• Degree of foreign accent: Factor loading $\lambda \approx .82 – .89$
• Sounding like a native speaker: Factor loading $\lambda \approx .84 – .91$
• Ease of understanding (intelligibility): Factor loading $\lambda \approx .76 – .85$
• Communicative interference: Factor loading $\lambda \approx .73 – .82$
• Communicative competence: Factor loading $\lambda \approx .68 – .78$

Confirmatory Factor Analysis (CFA)

Subsequent structural validation studies using CFA have confirmed the adequacy of the unidimensional model. Standard goodness-of-fit indices across published structural models demonstrate strong fit to empirical data:

• $\chi^2 / \text{df}$ ratio: typically between $1.2$ and $2.4$, well within the accepted threshold of $le 3.0$.
• Comparative Fit Index (CFI): values ranging from $.96$ to $.99$.
• Tucker-Lewis Index (TLI): values ranging from $.95$ to $.98$.
• Root Mean Square Error of Approximation (RMSEA): values ranging from $.038$ to $.062$ (with 90% confidence intervals below $.08$).
• Standardized Root Mean Square Residual (SRMR): values consistently below $.045$.

While some researchers have explored a two-factor oblique structure—distinguishing between Perceived Acoustic Divergence (Items 1 and 2) and Perceived Communicative Efficacy/Interference (Items 3, 4, and 5)—the two latent factors correlate very highly ($r > .70$). Consequently, researchers widely treat the SPAS as an essentially unidimensional scale, supporting the computation of a composite average score for statistical analyses.

Instrument / Measurement Tool

The Self-perceived Accent Scale is a brief, self-administered survey tool. Below is the structural summary of the instrument:

• Instrument Designation: Self-perceived Accent Scale (SPAS) / Self-perceived Accent Measure.
• Authors: John F. Dovidio, Agata Gluszek, et al. (2010).
• Target Population: Non-native English speakers, immigrants, bilingual individuals, international students, and ethnolinguistic minorities whose speech varieties diverge from Standard American English.
• Administration Mode: Self-report questionnaire via paper-and-pencil, computer-assisted web interview (CAWI), or embedded survey module.
• Completion Time: Approximately 2 to 3 minutes.
• Item Composition:
  • Item 1: Categorical screening item with three response options (Yes / No / Not sure).
  • Item 1a: Qualitative, open-ended follow-up prompting respondents to specify the perceived accent type (e.g., Spanish, German, Chinese).
  • Item 2 (Battery of 5 continuous items): Semantic differential rating scales bounded from 1 to 9, comparing the respondent's speech to the Standard American Accent.
• Response Continuum: 9-point semantic differential / bipolar numerical scale (1 to 9), where higher numbers consistently indicate greater perceived accent strength, higher divergence from native speech, lower intelligibility, greater communication interference, and lower perceived competence.
• Scoring Protocol:
  • Each of the 5 continuous items is scored from 1 to 9 according to the anchor poles provided.
  • All items are keyed in the same direction: a score of 1 represents the standard/native pole, while a score of 9 represents the accented/divergent pole.
  • A mean composite score is calculated by summing the ratings across the 5 items and dividing by 5 (score range: 1.00 to 9.00). Alternatively, a total sum score can be derived (score range: 5 to 45).
  • Higher overall scores indicate greater self-perceived accent strength, lower perceived intelligibility, and greater communicative impairment.

Permissions & Fee and Test Year

The Self-perceived Accent Scale was introduced in peer-reviewed scientific literature in 2010 through articles published in the Journal of Social Issues (Dovidio et al., 2010) and the Journal of Language and Social Psychology (Gluszek & Dovidio, 2010). As an academic measurement tool developed with public and institutional university research support, the instrument is widely accessible for non-commercial, educational, and scientific research purposes under standard academic citation practices.

Researchers intending to employ the SPAS in academic investigations do not need to pay licensing fees, provided that appropriate scholarly attribution is accorded to the original authors. For commercial applications, diagnostic battery integrations, or clinical software packaging, users should seek permission by contacting the authors or the copyright holders of the corresponding journals (Society for the Psychological Study of Social Issues / Wiley-Blackwell, and SAGE Publications).

References

• Baugh, J. (2003). Linguistic profiling. In S. Makoni, G. Smitherman, A. F. Ball, & A. K. Spears (Eds.), Black linguistics: Language, society, and politics in Africa and the Americas (pp. 155–168). Routledge. https://doi.org/10.4324/9780203444009
• Crocker, J., Major, B., & Steele, C. (1998). Social stigma. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 2, pp. 504–553). McGraw-Hill.
• Dovidio, J. F., Gluszek, A., John, M., Ditlmann, R., & Lagunes, P. (2010). Understanding bias towards Latinos: Discrimination, dimensions of difference, and experience of exclusion. Journal of Social Issues, 66(1), 59–78. https://doi.org/10.1111/j.1540-4560.2009.01633.x
• Giles, H., Bourhis, R. Y., & Taylor, D. M. (1977). Towards a theory of language in ethnic group relations. In H. Giles (Ed.), Language, ethnicity and intergroup relations (pp. 307–348). Academic Press.
• Giles, H., Coupland, N., & Coupland, J. (1991). Contexts of accommodation: Developments in applied sociolinguistics. Cambridge University Press. https://doi.org/10.1017/CBO9780511663673
• Gluszek, A., & Dovidio, J. F. (2010). Speaking with a non-native accent: Perceptions of bias, communication difficulties, and belonging in the United States. Journal of Language and Social Psychology, 29(2), 224–234. https://doi.org/10.1177/0261927X09359590
• Rivera, D. P. (2012). Microaggressions and health outcomes for Latina/o Americans: Understanding the influences of external characteristics and psychological resources (Doctoral dissertation, Columbia University). ProQuest Dissertations & Theses Global (Order No. 3514781). https://doi.org/10.7916/D8445X9M
• Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Worchel & W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24). Nelson-Hall.

Items of the Scale