Attitude Function Inventory (AFI)

Abstract

The Attitude Function Inventory (AFI) is an influential psychometric instrument developed by social psychologist Gregory M. Herek during the mid-1980s to operationalize and empirically evaluate functional theories of attitudes. Rooted in Herek’s neofunctional approach to social psychology, the inventory shifts analytical attention away from traditional unidimensional measures of attitude valence (e.g., positive vs. negative evaluations) toward the underlying psychological needs, motivational imperatives, and cognitive goals served by holding specific attitudes. While the broader AFI framework can theoretically be adapted to diverse target objects, its most validated and widely utilized incarnation evaluates heterosexual individuals’ attitudes toward gay men and lesbians. The standard operational scale consists of 10 self-report items mapped across four distinct motivational dimensions: experiential-schematic (evaluations grounded in direct personal interactions), social-expressive (attitudes functioning to secure social acceptance and signal group affiliation), defensive (ego-protective mechanisms mitigating internal conflict, personal anxiety, or psychosexual threat), and value-expressive (evaluations expressing core moral, ethical, ideological, or religious convictions). Administered on a 9-point Likert-type continuum ranging from 1 (“Very True of Me”) to 9 (“Not at all True of Me”), the AFI has exhibited internal consistency estimates ranging from moderate to high (α = .41 to .82 across subscales), robust construct and convergent validity when contrasted with prejudice indices, and strong predictive utility within functional matching persuasion paradigms. This article provides an exhaustive psychometric review of the AFI, addressing its theoretical foundations, structural factor architecture, cross-scale validity, operational methodology, scoring mechanics, and enduring relevance in modern social psychology, prejudice research, and intervention design.

Keywords

Attitude Function Inventory, functional approach to attitudes, neofunctional theory, Gregory Herek, sexual prejudice, value-expressive function, ego-defensive function, social-expressive function, experiential-schematic function, functional matching hypothesis

Authors

The Attitude Function Inventory was conceived, validated, and refined by Gregory M. Herek, Ph.D., an internationally recognized pioneer in the social psychological study of prejudice against sexual minorities, sexual orientation stigma, and hate crimes.

• Gregory M. Herek, Ph.D.: Professor Emeritus of Psychology at the University of California, Davis. Dr. Herek received his doctorate in social psychology from the University of California, Davis, followed by postdoctoral research at Yale University. Across four decades of scholarship, Herek has authored seminal treatises on the nature of heterosexism, homophobia, and sexual stigma, testifying before the United States Congress and serving as an expert witness in major civil rights litigation concerning LGBTQ+ equality. His empirical research on attitude functions reformulated mid-twentieth-century functional theories into testable, modern cognitive-motivational frameworks.

Purpose

For decades, attitude measurement in social psychology focused predominantly on valence (the direction of an evaluation, whether favorable or unfavorable) and extremity (the intensity of that evaluation). Traditional assessments of intergroup attitudes, such as early scales of anti-gay hostility or general prejudice, reliably documented that an individual held negative or positive views toward gay men and lesbians. However, these valence-centric metrics remained functionally blind to the psychological motives driving those evaluations. Two individuals may register identical negative scores on an explicit prejudice measure, yet one individual’s hostility might be rooted in a conscious effort to conform to family norms, whereas the other’s hostility might stem from deep psychosexual ego-defensiveness or rigid moral-theological dogma.

The primary purpose of the Attitude Function Inventory (AFI) is to diagnose, distinguish, and quantify the specific cognitive and psychological functions that an attitude fulfills for the individual holding it. Grounded in functional attitude theory, the AFI addresses the fundamental question: What psychological goal does this attitude accomplish for the self? In doing so, the inventory transforms attitude research from descriptive categorization into mechanistic, motivational inquiry.

In applied and basic research, the AFI serves multiple vital functions:

• Testing the Functional Matching Hypothesis: A central tenet of persuasion literature is that persuasive appeals are significantly more effective when the content of the message matches the primary psychological function sustaining the target’s attitude (the “matching hypothesis”). Researchers utilize the AFI to baseline participant attitude functions and subsequently test whether value-framed, social-norm-framed, or experiential-framed interventions yield differential attitude change across matched versus mismatched experimental conditions.
• Deconstructing Prejudice and Heterosexism: The AFI allows social scientists to move beyond pathologizing labels such as “homophobia” (which implies a generalized, undifferentiated irrational fear) by parsing sexual prejudice into empirical, identifiable components. It clarifies why some heterosexual individuals experience revulsion and ego-threat (defensive function), while others reference civil rights philosophies (value-expressive function) or peer group solidarity (social-expressive function).
• Intervention and Anti-Bias Curriculum Design: Educational campaigns designed to mitigate bias frequently fail because they deploy generic information (such as factual myth-busting) against attitudes that serve defensive or social-belonging functions. The AFI provides clinicians, educators, and institutional diversity strategists with an assessment framework to map the functional profile of a community, ensuring that educational interventions target the actual psychological engines of bias.

Psychological Construct

The Attitude Function Inventory conceptualizes an attitude not merely as a passive stored memory or simple associative evaluation, but as an active, goal-directed cognitive structure. According to Herek’s neofunctional taxonomy, attitudes fulfill distinct psychological needs across four primary structural dimensions:

1. Experiential-Schematic Function

The experiential-schematic function characterizes attitudes derived directly from an individual’s personal, interpersonal history with members of the target social group. Within this motivational framework, the attitude serves as an organizational schema that organizes, summarizes, and structures past behavioral interactions to guide future behavioral expectations. If an individual has developed close, positive relationships with gay individuals, their favorable attitude is sustained by concrete episodic memories and relational experiences. Conversely, if past encounters were perceived as negative or nonexistent, the cognitive schema reflects those boundaries. This dimension corresponds closely to Gordon Allport’s classical intergroup contact hypothesis, wherein direct relational familiarity bridges intergroup boundaries.

2. Social-Expressive Function

The social-expressive function posits that attitudes are adopted, maintained, or expressed to secure social acceptance, strengthen affiliations, avoid ostracism, and maintain membership within meaningful reference groups. In this modality, the content of the attitude is secondary to its social utility: the individual evaluates the target group in a manner that mirrors the perceived consensus of valued others, such as family members, close friends, or respected mentors. For example, a young adult who expresses disapproval of gay men not because of personal theological convictions, but because expressing disapproval wins affirmation from a conservative peer group, is utilizing the attitude in a social-expressive capacity. This function satisfies fundamental social-belonging and impression-management motives.

3. Defensive (Ego-Defensive) Function

Inherited from classical psychoanalytic formulations and refined within cognitive social psychology, the defensive (or ego-defensive) function operates to protect the self from internal psychodynamic conflicts, existential anxiety, identity ambiguity, or perceived threats to self-esteem. In the context of attitudes toward gay men and lesbians, defensive attitudes frequently manifest as visceral discomfort, revulsion, or active cognitive avoidance (“I would rather not think about homosexuality”). As demonstrated in Herek’s research on heterosexual masculinity, anti-gay hostility among certain heterosexual men serves a defensive role by reinforcing their own masculine self-concept and distancing themselves from socially stigmatized attributes associated with perceived gender-role nonconformity.

4. Value-Expressive Function

The value-expressive function represents evaluations that serve as direct conduits for an individual’s core moral, ethical, ideological, and philosophical values. Unlike defensive attitudes (which are oriented around threat mitigation) or social-expressive attitudes (which are oriented around peer conformity), value-expressive attitudes fulfill an intrinsic need for self-affirmation, integrity, and existential consistency. An individual who supports LGBTQ+ rights on the grounds of universal civil liberties and human dignity is enacting a value-expressive function. Likewise, an individual whose opposition to homosexuality stems from deeply internalized religious doctrine concerning the cosmic order is also exhibiting a value-expressive function. The common psychological denominator is that the attitude affirms who the individual is and what they fundamentally believe is right and just.

Theoretical Framework

The Attitude Function Inventory is situated within the broader intellectual evolution of functional attitude theory, a paradigm that originated in the late 1950s and experienced a substantial cognitive-motivational revival in the 1980s.

Classical Functional Foundations

Functional attitude theory was originally formulated independently by two seminal research groups: M. Brewster Smith, Jerome S. Bruner, and Robert W. White (1956) in their work Opinions and Personality, and Daniel Katz (1960) in his landmark paper “The Functional Approach to the Study of Attitudes.” Katz posited that attitudes perform four universal functions for personality:

1. Instrumental, adjustive, or utilitarian: Maximizing external rewards and minimizing punishments.
2. Ego-defensive: Protecting the self from intrapsychic tensions and harsh external realities.
3. Value-expressive: Giving positive expression to central values and the self-concept.
4. Knowledge: Providing consistency, structure, and clarity to an otherwise chaotic experiential world.

Despite the intuitive brilliance of Katz’s and Smith et al.’s models, classical functionalism stalled empirically throughout the 1960s and 1970s. The early theories lacked operational measurement instruments; researchers had no standardized methodology to assess which function an individual’s attitude served, rendering the theoretical claims largely unfalsifiable.

Herek’s Neofunctional Approach

In his landmark 1986 and 1987 publications, Gregory Herek revived functionalism by introducing a “neofunctional” perspective, specifically designed to translate theoretical functional classifications into measurable psychometric dimensions. Herek synthesized cognitive psychology’s insights regarding knowledge representation with motivational models of the self.

Herek posited that attitude functions can be categorized along two fundamental axes: evaluative versus expressive, and object-centered versus self-centered. The neofunctional model explicitly reconceptualized how attitudes toward social groups operate:

• Evaluative Functions: The attitude is oriented around the direct qualities of the attitude object itself (represented in the AFI by the experiential-schematic domain, wherein evaluations reflect actual behavioral interactions with gay people).
• Expressive Functions: The attitude is oriented around personal and social identity preservation, utilizing the attitude object as a symbolic vehicle to communicate social alignment (social-expressive), moral commitments (value-expressive), or intrapsychic safety (defensive).

By producing the AFI, Herek provided the first standardized self-report metric capable of directly classifying participants based on the self-reported functional underpinnings of their social attitudes.

Validity

The psychometric validity of the Attitude Function Inventory has been empirically documented through extensive investigations evaluating construct, convergent, discriminant, and predictive validities across diverse university and community samples.

Construct and Structural Validity

Construct validity for the AFI was established by demonstrating that the four operational subscales systematically map onto theoretically differentiated psychological traits. In foundational validation studies conducted by Herek (1987, 1988), participants who scored highest on the defensive subscale consistently registered significantly higher levels of generalized psychological defense mechanisms, anxiety regarding personal heterosexuality, and elevated scores on measures of traditional gender-role rigidity. In contrast, participants scoring high on the value-expressive subscale exhibited strong correlations with overarching ideological orientations, including Right-Wing Authoritarianism (RWA), personal religiosity, and principled commitments to civil liberties.

Convergent and Discriminant Validity

The AFI demonstrates robust convergent validity when examined alongside established metrics of sexual prejudice, such as Herek’s Attitudes Toward Lesbians and Gay Men (ATLG) Scale:

• High scores on the defensive subscale converge strongly and positively with negative attitude valence on the ATLG (particularly among heterosexual male respondents evaluating gay men), supporting the theoretical proposition that anti-gay hostility often fulfills an ego-defensive role in policing masculine boundaries.
• Scores on the experiential-schematic subscale correlate strongly with objective indices of interpersonal contact (e.g., number of openly gay friends, frequency of contact, and depth of personal disclosure), establishing discriminant validity from purely ideological or defensive motives.
• The social-expressive subscale demonstrates moderate-to-strong associations with measures of social conformity, need for social approval, and perceptions of peer group norms, distinguishing it from internal personal moral values.

Predictive and Persuasion Validity: The Matching Effect

The ultimate test of validity for any functional attitude inventory lies in its ability to predict differential susceptibility to persuasive messaging. In functional matching experiments (e.g., Herek, 1987; Tureau, 2003), investigators exposed participants to persuasive arguments specifically crafted to resonate with either value-expressive concerns (e.g., framing gay rights through the lens of constitutional fairness or religious charity) or social-expressive concerns (e.g., presenting peer-norm data demonstrating growing social acceptance). Participants categorized via the AFI as possessing primarily value-expressive attitudes exhibited significantly greater attitude change when exposed to value-matched messages compared to mismatched messages. This predictive capability firmly validates the AFI as a measure of underlying functional orientation rather than an artifact of semantic response bias.

Reliability

The internal consistency of the Attitude Function Inventory has been extensively evaluated across multiple validation cohorts, including college student samples, general adult populations, and targeted demographic groups.

Internal Consistency Reliability

Cronbach’s alpha coefficients (α) for the four subscales across Herek’s primary validation literature exhibit the following ranges:

• Experiential-Schematic Subscale (4 items): α = .67 to .82. This subscale displays robust internal consistency, reflecting strong item-total correlations among items measuring personal contact, relational history, and future interaction likelihood.
• Social-Expressive Subscale (2 items): α = .41 to .62. The lower alpha range for this subscale is largely a mathematical function of its short length (consisting of only two items). Psychometric evaluation via the Spearman-Brown prophecy formula indicates adequate internal coherence given the brevity of the index.
• Defensive Subscale (2 items): α = .72 to .82. Despite consisting of only two items, this subscale demonstrates exceptionally strong internal reliability, reflecting high inter-item correlation between avoidance (“rather not think about”) and visceral negative affect (“feelings of discomfort or revulsion”).
• Value-Expressive Subscale (2 items): α = .53 to .61. The moderate internal reliability observed in this subscale stems from the intentional inclusion of distinct ideological facets (civil liberties vs. overarching moral beliefs), which represent related but non-identical value frameworks.

Test-Retest Stability

Longitudinal assessments of the AFI indicate acceptable test-retest reliability across 4- to 8-week intervals, with stability coefficients typically ranging between r = .65 and r = .79 across the subscales. The defensive and value-expressive dimensions demonstrate the highest temporal stability, consistent with the theoretical postulate that core intrapsychic defense structures and internalized moral architectures remain relatively enduring over time in the absence of targeted cognitive interventions.

Factor Analysis

The structural dimensionality of the Attitude Function Inventory has been rigorously examined through both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Structure

In Herek’s original factor-analytic investigations (1984, 1987), principal axis factoring with both orthogonal (Varimax) and oblique (Promax / Direct Oblimin) rotations was conducted on the pool of functional items. Because functional orientations are not assumed to be strictly mutually exclusive, oblique rotations provide the most conceptually sound representation. The exploratory analyses repeatedly resolved into a clean four-factor solution accounting for the majority of the total common variance:

• Factor 1 (Experiential-Schematic): Items 1, 2, 3, and 4 cleanly load onto this factor, with primary factor loadings typically ranging between .58 and .84. Cross-loadings on defensive or value-expressive factors remain consistently below .25.
• Factor 2 (Defensive): Items 7 and 8 load heavily onto this dimension, with primary loadings exceeding .75, capturing visceral avoidance and somatic revulsion.
• Factor 3 (Social-Expressive): Items 5 and 6 display primary loadings exceeding .60, reflecting the influence of respected peers and close social circles.
• Factor 4 (Value-Expressive): Items 9 and 10 load onto this separate ideological factor, with primary loadings between .55 and .72.

Confirmatory Factor Analysis and Model Fit

Subsequent psychometric evaluations utilizing Confirmatory Factor Analysis (e.g., Tureau, 2003) have tested competing structural models to verify the superiority of the four-factor neofunctional model against alternative configurations (e.g., a unidimensional model or a two-factor “instrumental versus symbolic” model). Structural equation modeling results consistently demonstrate that the four-factor correlated model achieves acceptable goodness-of-fit indices:

As illustrated above, the four-factor model yields a Comparative Fit Index (CFI) of .95, a Root Mean Square Error of Approximation (RMSEA) below .05, and a Standardized Root Mean Square Residual (SRMR) of .042, confirming that the four theoretical dimensions represent empirically verifiable, distinct psychological constructs.

Instrument / Measurement Tool

• Instrument Name: Attitude Function Inventory (AFI)
• Target Domain: Attitudes toward gay men and lesbians (functional motivational foundations)
• Administration Format: Self-administered paper-and-pencil or computerized questionnaire
• Total Item Count: 10 items
• Subscale Item Composition:
  • Experiential-Schematic Subscale: Items 1, 2, 3, and 4 (4 items)
  • Social-Expressive Subscale: Items 5 and 6 (2 items)
  • Defensive Subscale: Items 7 and 8 (2 items)
  • Value-Expressive Subscale: Items 9 and 10 (2 items)
• Response Scale: 9-point Likert-type continuum anchored from 1 (“Very True of Me”) to 9 (“Not at all True of Me”).
• Scoring and Transformation Rules:
  • Depending on the research protocol, items are frequently reverse-scored (e.g., transformed via 10 - Score) so that higher values denote greater endorsement of that specific attitude function (i.e., 9 = Highly True of Me, 1 = Not at all True of Me).
  • Subscale scores are calculated by averaging or summing the corresponding item scores within each dimension.
  • Because the subscales represent distinct functional dimensions rather than facets of a single unidimensional trait, an aggregate “total score” is conceptually meaningless and should not be calculated. Researchers instead generate a four-dimensional functional profile for each respondent.
• Administration Duration: Approximately 3 to 5 minutes.

Permissions & Fee and Test Year

The foundational development and initial validation of the Attitude Function Inventory were published by Dr. Gregory M. Herek across a series of journal articles between 1986 and 1987 (Herek, 1986a, 1987). As an academic psychometric tool developed under institutional and federal research support, the AFI resides in the public domain for non-commercial scholarly and educational research purposes.

• Commercial Usage & Royalties: There are no licensing fees, purchase costs, or commercial royalties required for utilizing the AFI in academic investigations, thesis dissertations, or non-profit educational research.
• Usage Permissions: Formal written permission from the author is generally not required for non-commercial scholarly research, provided that appropriate academic attribution and citations are given to Dr. Gregory M. Herek’s published works in any resulting publications, presentations, or data repositories.
• Institutional Reference: Research inquiries regarding the historic development of the scale may be directed to the Department of Psychology at the University of California, Davis.

References

The following peer-reviewed publications and scholarly sources document the development, validation, and theoretical framework of the Attitude Function Inventory:

• Herek, G. M. (1984). Attitudes toward lesbians and gay men: A factor-analytic study. Journal of Homosexuality, 10(1–2), 39–51. https://doi.org/10.1300/J082v10n01_03
• Herek, G. M. (1984). Beyond “homophobia”: A social psychological perspective on attitudes toward lesbians and gay men. Journal of Homosexuality, 10(1–2), 1–21. https://doi.org/10.1300/J082v10n01_01
• Herek, G. M. (1986a). The instrumentality of attitudes: Toward a neofunctional theory. Journal of Social Issues, 42(2), 99–114. https://doi.org/10.1111/j.1540-4560.1986.tb00228.x
• Herek, G. M. (1986b). On heterosexual masculinity: Some psychical consequences of the social construction of gender and sexuality. American Behavioral Scientist, 29(5), 563–577. https://doi.org/10.1177/000276486029005005
• Herek, G. M. (1987). Can functions be measured? A new perspective on the functional approach to attitudes. Social Psychology Quarterly, 50(4), 285–303. https://doi.org/10.2307/2786814
• Herek, G. M. (1988). Heterosexuals’ attitudes toward lesbians and gay men: Correlates and gender differences. The Journal of Sex Research, 25(4), 451–477. https://doi.org/10.1080/00224498809551476
• Herek, G. M. (1990). The context of anti-gay violence: Notes on cultural and psychological heterosexism. Journal of Interpersonal Violence, 5(3), 316–333. https://doi.org/10.1177/088626090005003006
• Herek, G. M. (2002). Gender gaps in public opinion about lesbians and gay men. Public Opinion Quarterly, 66(1), 40–66. https://doi.org/10.1086/338631
• Katz, D. (1960). The functional approach to the study of attitudes. Public Opinion Quarterly, 24(2), 163–204. https://doi.org/10.1086/266945
• Smith, M. B., Bruner, J. S., & White, R. W. (1956). Opinions and Personality. John Wiley & Sons.
• Tureau, Z. L. (2003). College student identity and attitudes toward gays and lesbians (Doctoral dissertation, University of North Texas). UNT Digital Library. https://digital.library.unt.edu/ark:/67531/metadc4286/

Items of the Scale