Ambivalent Sexism Inventory (ASI)

Abstract

The Ambivalent Sexism Inventory (ASI), developed by social psychologists Peter Glick and Susan T. Fiske in 1996, represents a transformative theoretical and psychometric framework designed to quantify gender bias. Moving beyond traditional unidimensional conceptions of prejudice that defined sexism solely as overt hostility or antipathy toward women, the ASI operationalizes sexism as a multidimensional, ambivalent construct comprised of two complementary and functionally interdependent subcomponents: Hostile Sexism (HS) and Benevolent Sexism (BS). Hostile Sexism reflects classic, adversarial beliefs and derogatory attitudes directed toward women who challenge male dominance, power structures, or traditional societal roles. In contrast, Benevolent Sexism encapsulates a subjectively positive, protective, yet paternalistic orientation toward women who conform to traditional gender expectations, casting women as pure, delicate, and in need of male protection and adoration.

The ASI consists of 21 self-report items administered on a 6-point Likert response scale ranging from 0 (Disagree strongly) to 5 (Agree strongly). Within the 11-item Benevolent Sexism dimension, three distinct yet intercorrelated sub-factors are captured: Protective Paternalism, Complementary Gender Differentiation, and Heterosexual Intimacy. Hostile Sexism is assessed through 10 items spanning dominative paternalism, competitive gender differentiation, and heterosexual hostility. Extensive psychometric evaluations across dozens of nations demonstrate exceptional internal consistency (Cronbach’s alphas routinely ranging between .80 and .92 for HS, and .73 to .92 for BS), robust factor structures verified via confirmatory factor analysis (CFA), and notable criterion-related, convergent, and discriminant validity. The instrument has fundamentally reshaped social psychological research into structural inequality, political psychology, intimate partner violence, system justification, and institutional gender dynamics worldwide.

Keywords

Ambivalent Sexism Inventory, Hostile Sexism, Benevolent Sexism, Gender Prejudice, Paternalism, Heterosexual Intimacy, Gender Roles, System Justification Theory, Psychometrics, Scale Validation, Social Psychology

Authors

The Ambivalent Sexism Inventory was conceived, validated, and published by:

• Peter Glick, Ph.D. — Henry Merritt Wriston Professor of the Social Sciences and Professor of Psychology at Lawrence University, Appleton, Wisconsin, United States. Dr. Glick is an internationally recognized authority on the psychology of prejudice, discrimination, stereotyping, and power dynamics across social hierarchies.
• Susan T. Fiske, Ph.D. — Eugene Higgins Professor of Psychology and Professor of Public Affairs at Princeton University, Princeton, New Jersey, United States. Dr. Fiske is a renowned social cognitive theorist celebrated for developing the Stereotype Content Model (SCM), power-as-control theory, and foundational treatises on interpersonal perception, prejudice, and social cognition.

Correspondence regarding original archival validation datasets, theoretical extensions, and permissions is typically directed to Dr. Peter Glick at the Department of Psychology, Lawrence University, or Dr. Susan T. Fiske at the Department of Psychology, Princeton University.

Purpose

Prior to the mid-1990s, the empirical measurement of sexism relied heavily on instruments such as the Attitudes Toward Women Scale (AWS) and the Modern Sexism Scale (MSS). While these instruments effectively captured overt resistance to women’s legal, occupational, and political equality, they suffered from significant psychometric limitations. Specifically, as societal norms shifted to condemn blatant expressions of misogyny, conventional scales exhibited pronounced ceiling effects, social desirability distortion, and an inability to account for widespread, subjectively affectionate attitudes that men expressed toward women. Researchers struggled to explain why traditional women were frequently idealized and venerated, even within deeply patriarchal societies where female autonomy was severely constrained.

The primary purpose of the Ambivalent Sexism Inventory is to reconcile this paradox by providing a nuanced, psychometrically rigorous instrument capable of simultaneously measuring both punitive, antipathetic orientations and chivalrous, patronizing attitudes. The tool was developed to achieve three primary objectives:

• Differentiating Hostile and Benevolent Facets: To systematically demonstrate that sexism is not a unitary affective dimension of pure antipathy, but rather a dual-component system wherein antipathetic attitudes (HS) and subjectively positive, paternalistic beliefs (BS) act as complementary mechanisms of social control.
• Exposing System-Justifying Ideologies: To provide researchers with a mechanism for evaluating how benevolent attitudes disarm women’s resistance to gender hierarchy by offering men’s protection, affection, and veneration in exchange for compliance with subordinate, traditional domestic roles.
• Facilitating Cross-Cultural and Clinical Investigation: To offer a cross-culturally invariant metric for examining interpersonal relationship functioning, workplace discrimination, evaluation of sexual harassment and rape victims, institutional hiring biases, and policy resistance in both experimental and observational contexts.

In applied and organizational contexts, the ASI serves as a diagnostic metric to identify hidden cultural hurdles, such as benevolent paternalism in corporate management (e.g., shielding female executives from challenging international assignments under the guise of work-family protection). In socio-legal research, the ASI assists in deconstructing victim-blaming attributions in sexual assault cases and assessing judicial biases regarding domestic violence.

Psychological Construct

The Ambivalent Sexism Inventory conceptualizes gender prejudice through two primary ideological dimensions and three sub-facets within the benevolent construct. These dimensions reflect how biological reproduction, sexual interdependence, and structural male dominance interact to shape intergroup attitudes between men and women.

1. Hostile Sexism (HS)

Hostile Sexism operationalizes an adversarial, antagonistic orientation toward women who are perceived as seeking to usurp male power, manipulate men through sexuality, or challenge traditional patriarchal authority. Rooted in social dominance orientation and competitive threat, HS characterizes women as inherently manipulative, overly sensitive, deceitful, and demanding unwarranted privileges under the banner of equality. The construct specifically targets non-traditional women, such as feminists, career-oriented women, female leaders, and sexually assertive women. Representative expressions include accusations that women exaggerate workplace discrimination or deploy seduction to exert control over male autonomy.

2. Benevolent Sexism (BS)

Benevolent Sexism is defined as a set of interrelated attitudes toward women that are subjectively positive in affective tone yet profoundly patronizing and functional in maintaining gender inequality. BS rewards women who conform to traditional maternal, domestic, and subordinate roles by framing them as delicate creatures who require male protection, cherishable companions who bring emotional completion to men, and moral compasses whose refined sensibilities elevate masculine character. BS is comprised of three distinct theoretical sub-constructs:

• Protective Paternalism: The belief that men have an obligation to provide for, defend, rescue, and shield women from life’s hardships. While appearing chivalrous, it implicitly frames women as weak, vulnerable, and incompetent in high-stress, dangerous, or autonomous domains.
• Complementary Gender Differentiation: The conviction that women naturally possess distinctive, refined virtues—such as superior moral purity, empathy, and artistic sensibility—that men lack. While laudatory, this construct pigeonholes women into traditional expressive roles while reserving instrumental, executive, and worldly authority for men.
• Heterosexual Intimacy: The psychological and physical reliance of men on women for dyadic fulfillment, romantic intimacy, and emotional completion. This facet posits that a man is inherently incomplete without the devotion and romantic love of a woman, which intensifies the masculine drive to possess, protect, and regulate female partners.

Theoretical Framework

The Ambivalent Sexism Inventory is grounded in Ambivalent Sexism Theory (AST), formulated by Glick and Fiske (1996, 2001). AST represents an integration of Social Identity Theory, System Justification Theory, and evolutionary social psychology. Unlike other forms of intergroup prejudice—such as racial, ethnic, or religious animus, where dominant outgroups frequently practice complete social segregation, expulsion, or physical subjugation—gender relations are uniquely characterized by the confluence of two contradictory sociological forces:

• Structural Male Dominance (Patriarchy): Across human societies, men have historically monopolized formal economic, political, and institutional power, resulting in status asymmetries and differential access to resources.
• Dyadic Interdependence: Due to human reproductive biology and sexual pair-bonding, men remain deeply dependent on women for reproduction, child rearing, and emotional and sexual gratification. Dominant men cannot simply banish or physically isolate the subordinate group upon whom their biological lineage and personal happiness depend.

This structural reality produces an inherently ambivalent psychological dynamic. As Glick and Fiske posited, patriarchy alone would foster pure hostile sexism: dominant group members disparage the subordinate group to rationalize their privileged status. However, dyadic interdependence necessitates benevolent sexism: dominant group members must cultivate warm, positive, rewarding affiliations with the subordinate group to maintain cooperative domestic partnerships.

Consequently, Hostile Sexism and Benevolent Sexism operate as an ideological “carrot-and-stick” system of behavioral reinforcement. Hostile Sexism serves as the punitive stick, punishing women who step outside traditional boundaries, demand institutional parity, or contest male prerogatives. Benevolent Sexism serves as the seductive carrot, rewarding women who embrace traditional domesticity, submissiveness, and sexual fidelity with male devotion, economic support, and physical protection. Together, they function as an exceptionally stable, self-reinforcing socio-cognitive mechanism that sustains gender hierarchy without requiring continuous coercion.

Validity

The psychometric validity of the ASI has been corroborated across hundreds of empirical studies spanning diverse cross-cultural, organizational, and experimental paradigms.

Construct and Structural Validity

Construct validity was initially established across six independent samples comprising over 2,200 participants by Glick and Fiske (1996). Confirmatory factor analyses demonstrated that models separating Hostile Sexism and Benevolent Sexism (with BS configured as a second-order factor subsuming Protective Paternalism, Complementary Gender Differentiation, and Heterosexual Intimacy) displayed superior fit indices compared to unidimensional or uncorrelated models. This bi-dimensional structure has demonstrated invariance across gender, age cohorts, and national cultures.

Convergent Validity

Hostile Sexism exhibits strong positive correlations with established measures of overt gender bias and authoritarianism, including:

• The Attitudes Toward Women Scale (AWS) (correlations typically ranging from r = -.50 to -.65 with conservative attitudes).
• The Modern Sexism Scale (MSS) and Old-Fashioned Sexism Scale (OFSS) (r = .55 to .70).
• Social Dominance Orientation (SDO) (r = .40 to .55), confirming that HS reflects a preference for group-based hierarchy and dominance.

Benevolent Sexism correlates moderately with Hostile Sexism (national averages typically range from r = .20 to .50 in men, and .10 to .40 in women), indicating that while they are conceptually distinct, they share an underlying endorsement of traditional gender division. BS demonstrates strong convergent associations with Right-Wing Authoritarianism (RWA), religiosity, and conventional morality, reflecting its focus on tradition, security, and social cohesion.

Predictive and Criterion Validity

In a landmark cross-national study across 19 nations (N = 15,000; Glick et al., 2000), national averages of HS and BS systematically predicted United Nations indices of gender inequality, specifically the Gender Development Index (GDI) and the Gender Empowerment Measure (GEM). In nations where men endorsed higher levels of hostile and benevolent sexism, women exhibited lower political empowerment, economic self-sufficiency, and literacy parity. Crucially, in countries characterized by intense male hostility toward women, women’s own endorsement of benevolent sexism increased, functioning as an adaptive, protective buffer against male aggression.

Experimental studies demonstrate that high HS scores predict victim-blaming in acquaintance rape scenarios, tolerance of domestic abuse, and hostility toward women seeking leadership roles. Conversely, high BS scores predict blaming female victims who violate traditional dress or decorum norms, endorsement of patronizing workplace feedback, and women’s self-reported reduction in cognitive performance and career aspirations following exposure to chivalrous comments.

Reliability

The Ambivalent Sexism Inventory consistently exhibits robust psychometric reliability across diverse demographic strata, linguistic translations, and temporal intervals.

Internal Consistency

Extensive psychometric investigations report high internal consistency across both primary scales:

• Hostile Sexism Subscale (10 items): Cronbach’s alpha (α) values routinely fall between .83 and .92 across male and female respondents in normative, collegiate, and general population samples. Composite reliability coefficients in structural modeling studies consistently exceed .85.
• Benevolent Sexism Subscale (11 items): Cronbach’s alpha values typically range between .73 and .92. When examining the three sub-facets individually, reliability estimates are acceptable given their short length: Protective Paternalism (α = .70 to .82), Complementary Gender Differentiation (α = .68 to .80), and Heterosexual Intimacy (α = .65 to .81).

Test-Retest Stability

Temporal stability assessments over intervals ranging from 2 weeks to 6 months demonstrate high test-retest reliability coefficients. Test-retest correlations for Hostile Sexism average r = .80 to .88, whereas Benevolent Sexism demonstrates stability coefficients ranging between r = .75 and .84. These findings substantiate that the ASI captures enduring socio-cognitive ideological orientations rather than transient affective states.

Factor Analysis

The latent structure of the ASI was established through exploratory factor analysis (EFA) and confirmed via confirmatory factor analysis (CFA) across numerous multi-ethnic and cross-national samples.

Exploratory Factor Analysis (EFA)

Principal axis factoring and maximum likelihood extractions with oblique (Promax or Oblimin) rotation consistently extract two dominant, correlated higher-order factors:

• Factor 1: Hostile Sexism: The 10 items assessing antipathy, resentment, and competitive gender suspicion cleanly load on this factor, with primary factor loadings typically ranging between .50 and .82, and cross-loadings onto the benevolent dimension remaining below .20.
• Factor 2: Benevolent Sexism: The 11 items assessing protective paternalism, gender complementarity, and heterosexual idealization load onto a separate dimension, with standardized loadings ranging between .42 and .78.

Confirmatory Factor Analysis (CFA) & Model Fit

Structural equation modeling confirms that a hierarchical, multidimensional model provides the most theoretically coherent and statistically superior fit to empirical data. Specifically, the optimal structural specification treats Hostile Sexism as a single, first-order latent factor (comprising 10 indicator items) and Benevolent Sexism as a second-order latent construct determined by three first-order latent sub-factors: Protective Paternalism (items 3, 8, 16, 19), Complementary Gender Differentiation (items 7, 18, 21), and Heterosexual Intimacy (items 1, 5, 11, 12).

Typical fit indices reported across the international literature demonstrate strong goodness-of-fit:

• Comparative Fit Index (CFI): ≥ .92 to .96
• Tucker-Lewis Index (TLI): ≥ .91 to .95
• Root Mean Square Error of Approximation (RMSEA): .042 to .058 (with 90% confidence intervals well within acceptable thresholds)
• Standardized Root Mean Square Residual (SRMR): ≤ .050

Multi-group CFA has further demonstrated metric and scalar measurement invariance across male and female cohorts, confirming that gender differences in mean scores reflect true latent distributional differences rather than measurement bias.

Instrument / Measurement Tool

• Test Type: Self-report psychometric rating scale; social attitudes inventory.
• Format: Paper-and-pencil or computer-assisted self-administered questionnaire.
• Total Item Count: 21 items.
  • Hostile Sexism Subscale: 10 items.
  • Benevolent Sexism Subscale: 11 items (subsuming Protective Paternalism: 4 items; Complementary Gender Differentiation: 3 items; Heterosexual Intimacy: 4 items).
• Response Scale: 6-point Likert-type scale scored from 0 to 5:
  • 0 = Disagree strongly
  • 1 = Disagree somewhat
  • 2 = Disagree slightly
  • 3 = Agree slightly
  • 4 = Agree somewhat
  • 5 = Agree strongly
• Reverse-Scored Items: Items 3, 5, 6, 12, 17, and 20 are negatively framed and must be reversed prior to computing subscale means (i.e., 0 becomes 5, 1 becomes 4, 2 becomes 3, 3 becomes 2, 4 becomes 1, and 5 becomes 0).

  Note on scoring conventions: In the original Glick & Fiske (1996) 22-item developmental instrument, an additional item was evaluated; in the finalized 21-item standardized ASI, the reverse-coded items are specifically 3, 5, 6, 12, 17, and 20.

• Subscale Item Composition:
  • Hostile Sexism (HS) (10 items): Calculate the arithmetic mean of items: 2, 4, 6*, 9, 10, 13, 14, 15, 17*, and 20* (where * indicates reverse scoring).
  • Benevolent Sexism (BS) (11 items): Calculate the arithmetic mean of items: 1, 3*, 5*, 7, 8, 11, 12*, 16, 18, 19, and 21 (where * indicates reverse scoring).
    • Protective Paternalism: Mean of items 3*, 8, 16, 19
    • Complementary Gender Differentiation: Mean of items 7, 18, 21
    • Heterosexual Intimacy: Mean of items 1, 5*, 11, 12*
• Overall Ambivalent Sexism Index: When an aggregate score is required, researchers compute the grand mean of all 21 items; however, literature strongly recommends analyzing HS and BS separately to capture their distinct behavioral and systemic correlates.

Permissions & Fee and Test Year

• Year of Initial Publication: 1996.
• Primary Citation: Glick, P., & Fiske, S. T. (1996). The Ambivalent Sexism Inventory: Differentiating hostile and benevolent sexism. Journal of Personality and Social Psychology, 70(3), 491–512.
• Intellectual Property & Permissions: The Ambivalent Sexism Inventory is copyrighted by the American Psychological Association (APA) and the original authors. However, the authors have placed the instrument in the public domain for academic, non-commercial research, educational instruction, and scientific replication purposes.
• Fee: There are no licensing fees, purchase costs, or royalty requirements for academic, non-profit empirical research. Commercial entities seeking to integrate the ASI into proprietary pre-employment screening, leadership consulting, or executive development programs should secure prior written permission from the authors or the APA.

References

• Fiske, S. T. (2000). Stereotyping, prejudice, and discrimination at the seam between the centuries: Evolution, culture, mind, and brain. European Journal of Social Psychology, 30(3), 299–322. https://doi.org/10.1016/B978-0-12-386915-9.00024-3
• Glick, P., & Fiske, S. T. (1996). The Ambivalent Sexism Inventory: Differentiating hostile and benevolent sexism. Journal of Personality and Social Psychology, 70(3), 491–512. https://doi.org/10.1037/0022-3514.70.3.491
• Glick, P., & Fiske, S. T. (2001). Ambivalent sexism. In M. P. Zanna (Ed.), Advances in Experimental Social Psychology (Vol. 33, pp. 115–188). Academic Press. https://doi.org/10.1016/S0065-2601(01)80005-8
• Glick, P., & Fiske, S. T. (2011). Ambivalent sexism revisited. Psychology of Women Quarterly, 35(3), 530–535. https://doi.org/10.1177/0361684311414832
• Glick, P., Fiske, S. T., Mladinic, A., Saiz, J. L., Wang, Q., Murphy, D. J., … & López, W. L. (2000). Beyond prejudice as simple antipathy: Hostile and benevolent sexism across 19 nations. Journal of Personality and Social Psychology, 79(5), 763–775. https://doi.org/10.1037/0022-3514.79.5.763

Items of the Scale