Domain-Specific Innovativeness Scale (DSI)

Abstract

The Domain-Specific Innovativeness Scale (DSI), developed by Ronald E. Goldsmith and Charles F. Hofacker in 1991, represents a foundational psychometric breakthrough in consumer psychology, marketing science, and behavioral economics. Designed to address the empirical deficiencies of generalized, personality-level novelty-seeking scales, the DSI conceptualizes innovativeness as a middle-range, domain-bound behavioral tendency reflecting an individual’s inclination to learn about and adopt new products within a designated category. The scale comprises a parsimonious battery of six self-report items evaluated on a standard 5-point Likert scale ranging from 1 (Strongly disagree) to 5 (Strongly agree), incorporating systematically balanced positive and reverse-scored items to attenuate acquiescence bias.

Extensive psychometric evaluations across five decades of consumer research demonstrate that the DSI possesses exceptional construct, convergent, discriminant, and predictive validity. Unlike global measures of innovativeness—such as Michael Kirton‘s Adaption-Innovation Inventory (KAI) or generalized sensation-seeking batteries—the DSI consistently accounts for substantial proportions of variance in actual product ownership, early adoption behavior, opinion leadership, and ongoing product category involvement. Factor analytic studies universally confirm an essentially unidimensional latent structure, with high internal consistency estimates (Cronbach’s alpha values typically ranging between .81 and .92 across diverse consumer domains, including consumer electronics, fashion, financial technologies, and online services). This comprehensive profile reviews the historical development, theoretical foundations, psychometric properties, factor structure, scoring algorithms, and widespread cross-cultural applications of the DSI.

Keywords

Domain-Specific Innovativeness Scale, DSI, consumer innovativeness, adoption of innovations, Goldsmith and Hofacker, product category involvement, opinion leadership, psychometrics, scale validation, middle-range theory, early adopters, consumer psychology

Authors

The Domain-Specific Innovativeness Scale was conceptualized, operationalized, and psychometrically validated by two prominent scholars in marketing science and quantitative psychology at Florida State University:

• Ronald E. Goldsmith, Ph.D.: Professor Emeritus of Marketing at the College of Business, Florida State University (Tallahassee, Florida, United States). Dr. Goldsmith is recognized internationally for his seminal contributions to consumer behavior, personality assessment in market research, self-concept theory, fashion leadership, and the psychometric measurement of consumer traits. His research has appeared prolifically in leading journals, including the Journal of the Academy of Marketing Science, the Journal of Business Research, and the Journal of Consumer Psychology.
• Charles F. Hofacker, Ph.D.: Carl DeSantis Professor of Business Administration and Professor of Marketing at the College of Business, Florida State University. Dr. Hofacker is a leading authority on quantitative marketing methods, digital marketing, structural equation modeling, psychometric measurement theory, and e-commerce consumer interactions. He served extensively as the editor of international academic journals and has published foundational methodological works on scale construction and electronic consumer behavior.

Purpose

The fundamental purpose of the Domain-Specific Innovativeness Scale (DSI) is to provide a reliable, valid, and highly predictive operationalization of consumer innovativeness situated at the level of specific product or service categories. For decades prior to Goldsmith and Hofacker’s 1991 publication, innovation researchers were confronted with what was colloquially termed the “innovativeness paradox” or the “personality-behavior gap.” Foundational theorists such as Everett M. Rogers (1962, 1983) and Thomas S. Robertson (1971) established that the rate of diffusion for new commercial products depends critically on early adopters—individuals who embrace innovations before the general public. However, early efforts to identify and profile these consumers using generalized psychological instruments—such as Cattell’s 16PF, Jackson’s Personality Research Form, or generic sensation-seeking and novelty-seeking scales—consistently yielded weak, statistically non-significant, or practically trivial correlations with actual purchasing behavior.

Goldsmith and Hofacker recognized that innovativeness does not operate as an omnibus, monolithic personality trait that manifests identically across all behavioral arenas. An individual who is an avid, intrepid early adopter of cutting-edge computing hardware and software may be completely traditional, risk-averse, and conservative regarding fashion, culinary choices, or automotive selections. Consequently, measuring innovativeness as a global personality trait creates severe criterion-predictor level mismatch, violating fundamental psychometric principles of behavioral correspondence articulated by Icek Ajzen and Martin Fishbein. The primary objective of the DSI was to bridge this theoretical and empirical chasm by providing an adaptable, middle-range psychometric instrument that focuses directly on the respondent’s behavioral predispositions within an explicit, designated product domain.

In applied market research, academic consumer psychology, and product management, the DSI serves several vital purposes:

• Market Segmentation and Profiling: It enables enterprises and market analysts to segment consumer populations based on their latent propensity to adopt innovations within a targeted sector (e.g., green technologies, mobile applications, luxury apparel, smart-home automation), allowing for highly tailored promotional campaigns.
• New Product Forecasting: By identifying the proportion of domain-innovative consumers within a target market, organizations can evaluate trial rates, baseline adoption curves, and diffusion velocity prior to expensive mass-market commercialization.
• Lead-User Identification and Co-Creation: Highly innovative domain consumers often possess advanced product knowledge and emergent latent needs. The DSI serves as an efficient screening mechanism for identifying participants for user-centric design workshops, beta-testing initiatives, and open-innovation collaborations.
• Academic Theory Testing: The scale allows researchers to examine complex structural models exploring the antecedents and consequences of domain innovativeness, such as testing how perceived risk, price sensitivity, social contagion, and opinion leadership interact to govern consumer adoption trajectories.

Psychological Construct

The psychological construct captured by the Domain-Specific Innovativeness Scale occupies an intermediate level within the hierarchy of personality and behavioral dispositions. Drawing upon the conceptual distinctions established by Midgley and Dowling (1978), consumer innovativeness can be bifurcated into innate innovativeness (a generalized, domain-independent cognitive predisposition reflecting an openness to new experiences, variety seeking, and novelty tolerance) and actualized innovativeness (observable, early adoption behavior). Goldsmith and Hofacker conceptualized domain-specific innovativeness as a mediating construct located precisely between generalized personality traits and discrete purchase behaviors. They defined it as “the tendency to learn about and adopt innovations (new products) within a specific domain of interest.”

This construct is inherently behavioral-predispositional rather than purely cognitive or purely behavioral. It encompasses three core, inextricably linked psychological components:

1. Relative Time of Adoption and Self-Perceived Earliness

At the center of domain innovativeness is the consumer’s subjective perception of their adoption timing relative to their salient peer group. Rogers’ classic diffusion paradigm delineates consumer categories along an adoption timeline: Innovators (first 2.5%), Early Adopters (next 13.5%), Early Majority (34%), Late Majority (34%), and Laggards (16%). The DSI operationalizes innovativeness not by requiring respondents to recall precise historical dates of purchase—which are chronically plagued by retrospective recall errors and memory decay—but rather through comparative social evaluations (e.g., “In general, I am among the first in my circle of friends to buy a new [product category] when it appears”). This social comparison component taps into the respondent’s identity within their peer network.

2. Domain-Bound Information Seeking and Receptivity

Domain-innovative individuals exhibit an active cognitive alertness and low perceptual threshold toward novel market offerings within the focal category. While the non-innovative consumer ignores market announcements, advertisements, or discussions regarding unfamiliar brands or new product launches, the domain-innovative individual displays voluntary exposure to category-specific information. They actively monitor industry developments, retail assortments, and specialized media channels, resulting in superior awareness of brand names, product variants, and technological specifications (e.g., “In general, I am the last in my circle of friends to know the [names/brands] of the latest [product category]” [reverse-scored]).

3. Willingness to Incur Category-Specific Adoption Risk

Every commercial innovation embodies inherent performance, financial, social, and psychological risks. A new product may fail to operate as advertised, become obsolete rapidly, or invite social disapproval. The domain-innovative consumer displays heightened tolerance for category-specific ambiguity and risk. They are prepared to commit personal and financial resources to acquire a new product based on minimal prior social proof, external validation, or trial (e.g., “I will buy a new [product category] even if I haven’t heard it/tried it yet”). This trait distinguishes true innovators from cautious mainstream consumers who require extensive peer testimonials and risk-reduction guarantees before committing to an adoption decision.

Theoretical Framework

The DSI is grounded in a convergence of three prominent theoretical frameworks from sociology, social psychology, and consumer science: the Diffusion of Innovations Theory, the Sociological Theory of the Middle Range, and Social Comparison Theory.

1. Rogers’ Diffusion of Innovations Theory

Everett M. Rogers formulated the foundational sociometric paradigm governing how technological and social innovations diffuse through social systems over time. Rogers underscored that innovativeness is the degree to which an individual is relatively earlier in adopting an innovation than other members of their social system. While Rogers initially emphasized systemic, demographic, and sociological variables (such as cosmopolitanism, social status, and external media exposure), consumer psychologists later struggled to operationalize this construct reliably in cross-sectional research. Goldsmith and Hofacker leveraged Rogers’ theoretical framework by retaining the central tenet of relative adoption speed while shifting the psychometric lens from retrospective historical event reporting to an enduring individual-difference trait anchored in the consumer’s self-concept and immediate reference system.

2. Merton’s Middle-Range Theory and Trait Specificity

Sociologist Robert K. Merton advocated for “theories of the middle range”—theories that lie intermediate between minor working hypotheses in everyday research and all-inclusive grand social theories. In psychometric modeling, this principle dictates that predictive power is maximized when the level of specificity of the predictor trait matches the specificity of the criterion behavior. Midgley and Dowling (1978) formalized this within consumer research, arguing that global personality measures (e.g., generalized curiosity, openness to experience, sensation seeking) must filter through numerous mediating psychological layers before influencing specific shopping behaviors. By establishing the DSI as a middle-range construct—broader than a single situational brand purchase, yet far more focused than broad personality structure—Goldsmith and Hofacker aligned the measurement tool directly with the psychological locus where consumer adoption decisions occur.

3. Festinger’s Social Comparison Theory

According to Leon Festinger‘s (1954) Social Comparison Theory, individuals evaluate their own abilities, traits, and opinions by comparing themselves with similar peers rather than against absolute, abstract standards. In market environments, consumers rarely track the statistical percentiles of innovation adoption across an entire nation or global population; instead, they gauge their consumer identity and behavioral timing against their immediate interpersonal networks (“my circle of friends”). The theoretical architecture of the DSI incorporates this normative reference framework across multiple items, acknowledging that consumer innovativeness is fundamentally lived and validated within localized social ecologies.

Validity

The construct, criterion, convergent, and discriminant validity of the DSI have been corroborated through hundreds of independent empirical investigations across dozens of product categories and global cultures.

Construct and Factorial Validity

Goldsmith and Hofacker (1991) initially validated the scale across six independent sample cohorts comprising over 800 consumers across diverse product domains, including recorded rock music, fashion clothing, and personal computing hardware. Exploratory and confirmatory factor analyses verified that all six items load strongly onto a single, coherent latent construct representing domain innovativeness, with primary factor loadings routinely exceeding .65, and no significant secondary cross-loadings observed. Subsequent structural analyses (e.g., Goldsmith, Freiden, & Eastman, 1995; Flynn & Goldsmith, 1993) demonstrated that the single-factor specification exhibits robust goodness-of-fit across adult consumer cohorts, university student panels, and online purchasing populations.

Criterion-Related and Predictive Validity

The predictive validity of the DSI stands as its greatest psychometric advantage over generalized innovativeness inventories. In the original validation studies, Goldsmith and Hofacker demonstrated that DSI scores accounted for between 15% and 35% of the explained variance in self-reported and objectively logged product category ownership (e.g., number of newly released compact discs purchased, ownership of new electronic peripherals). In direct head-to-head empirical comparisons, the DSI significantly outperformed global novelty-seeking measures (such as the Hurt, Joseph, & Cook Innovativeness Scale or the Kirton Adaption-Innovation Inventory) in predicting specific innovation adoption. Numerous longitudinal studies have confirmed that consumers scoring in the upper quartile of the DSI display dramatically shorter adoption lead-times following product launch, spend higher proportions of category expenditures on newly released models, and exhibit higher frequencies of store visits during product launch phases.

Convergent Validity

The scale demonstrates robust convergent validity with theoretical constructs functionally linked to early adoption dynamics:

• Domain Involvement: Scores on the DSI correlate strongly and positively (typically r = .50 to .72) with Zaichkowsky’s Personal Involvement Inventory (PII) and domain-specific interest measures. Innovative consumers are consistently deeply involved with the product class.
• Opinion Leadership: Convergent correlations with scales measuring opinion leadership (e.g., the Childers Opinion Leadership Scale; King & Summers, 1970) typically range from r = .45 to .68. Highly innovative domain consumers frequently serve as informal information sources and advisors for subsequent adopters.
• Subjective and Objective Knowledge: DSI scores correlate significantly with both self-assessed domain knowledge and objective tests of category brand awareness and technical literacy.

Discriminant Validity

Crucially, the DSI demonstrates sharp empirical independence from potential confounding constructs:

• Generalized Traits: Correlations between the DSI and broad personality markers (e.g., the Big Five traits of Extraversion, Neuroticism, or Conscientiousness) are negligible to weak (rarely exceeding r = .20), proving that the DSI captures unique domain variance rather than baseline temperament.
• Cross-Domain Discriminant Power: In multitrait-multimethod analyses, Goldsmith and Hofacker demonstrated that a respondent’s DSI score for one domain (e.g., rock music) displayed near-zero correlation with their DSI score for an unrelated domain (e.g., fashion clothing), empirically proving that innovativeness within one sphere cannot be conflated with innovativeness in another.
• Demographic Variables: The DSI remains largely orthogonal to age, gender, household income, and educational level, confirming that early adoption within specific categories is governed by psychological engagement rather than pure demographic privilege.

Reliability

The internal consistency and temporal stability of the Domain-Specific Innovativeness Scale have been documented extensively across three decades of psychometric scrutiny.

Internal Consistency (Cronbach’s Alpha and Composite Reliability)

In the seminal 1991 validation work by Goldsmith and Hofacker, Cronbach’s alpha coefficients for the six-item scale were computed across six independent validation samples spanning multiple consumer categories:

• Sample 1 (College students, Rock Music): α = .84
• Sample 2 (College students, Rock Music replication): α = .83
• Sample 3 (Adult non-students, Rock Music): α = .82
• Sample 4 (College students, Fashion clothing): α = .81
• Sample 5 (College students, Personal computers): α = .86
• Sample 6 (Adults, Fashion clothing): α = .85

Subsequent psychometric evaluations across global contexts have consistently reaffirmed internal consistency values comfortably exceeding the standard .80 threshold established for psychological instrumentation. In structural equation modeling paradigms, composite reliability (CR) metrics routinely range between .83 and .91, with Average Variance Extracted (AVE) values typically surpassing the recommended .50 benchmark, confirming that the variance captured by the latent construct is substantially greater than the variance attributable to measurement error.

Test-Retest Stability

Goldsmith and Hofacker administered the DSI across multi-week intervals to evaluate temporal stability. Test-retest reliability coefficients yielded stability estimates ranging between r = .78 and r = .87 over a three-to-four-week testing window, demonstrating that while domain innovativeness is responsive to changes in product lifecycle stages, it operates as a stable, enduring behavioral predisposition rather than a volatile, state-dependent mood fluctuation.

Factor Analysis

The structural dimensionality of the DSI has undergone rigorous examination using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis Findings

In original EFA implementations utilizing principal components analysis and maximum likelihood factor extraction with varimax and oblimin rotations, the scree test and Kaiser-Guttman eigenvalue criterion (eigenvalues > 1.0) unequivocally revealed a dominant single-factor solution. The primary latent dimension consistently accounts for between 52% and 68% of the total variance across items. Factor loadings for individual items are uniformly high and statistically significant:

• Item 1 (Reverse-scored): Loadings typically range from .62 to .78
• Item 2 (Direct): Loadings typically range from .71 to .84
• Item 3 (Reverse-scored): Loadings typically range from .60 to .75
• Item 4 (Reverse-scored): Loadings typically range from .66 to .81
• Item 5 (Direct): Loadings typically range from .74 to .88
• Item 6 (Direct): Loadings typically range from .65 to .79

Confirmatory Factor Analysis and Model Fit

Confirmatory factor analytic investigations evaluating the hypothesized single-factor structure against alternative multidimensional configurations have consistently supported the unidimensional model. Across multiple structural equation modeling studies (utilizing LISREL, AMOS, and Mplus), the single-factor specification demonstrates exceptional global goodness-of-fit indices:

• Comparative Fit Index (CFI): Values routinely exceed .95 (frequently ranging from .96 to .99).
• Tucker-Lewis Index (TLI): Consistently reported above .94.
• Root Mean Square Error of Approximation (RMSEA): Typically falls between .038 and .065, well below the conservative .08 threshold indicating close approximate fit.
• Standardized Root Mean Square Residual (SRMR): Values consistently remain below .045.

Methodological Considerations: Reverse-Keyed Items and Method Effects

A recurrent methodological discussion in the psychometric literature surrounding the DSI concerns the presence of reverse-scored items (Items 1, 3, and 4). In some large-sample CFA evaluations, initial models with a strict one-factor specification produce slight elevations in modification indices between the error terms of negatively worded items. Methodologists (e.g., Marsh, 1996) have noted that this is a classic artifact of method variance associated with item wording polarity rather than true multi-dimensionality. When researchers specify a methodological factor representing negatively worded items (a Correlated Trait-Correlated Method [CTCM] or bifactor formulation), or when error covariances between items sharing the same phrasing direction are freed, model fit improves to near-perfect levels (χ²/df < 2.0; CFI > .98; RMSEA < .04) while leaving the primary substantive factor loadings robust and invariant.

Instrument / Measurement Tool

• Tool Name: Domain-Specific Innovativeness Scale (DSI)
• Alternative Designations: Goldsmith-Hofacker Innovativeness Scale
• Authors: Ronald E. Goldsmith, Ph.D., and Charles F. Hofacker, Ph.D. (1991)
• Construct Assessed: Consumer behavioral predisposition toward learning about and adopting new products within a specified category
• Administration Format: Self-administered pencil-and-paper questionnaire, electronic web survey, or integrated panel assessment
• Completion Time: Approximately 2 to 3 minutes
• Number of Items: 6 items
• Target Population: Adolescent, adult, and specialized consumer populations
• Response Format: 5-point Likert scale:
  • 1 = Strongly disagree
  • 2 = Disagree
  • 3 = Neither agree nor disagree (Neutral)
  • 4 = Agree
  • 5 = Strongly agree
• Scoring and Directionality:
  • Directly Scored Items: Item 2, Item 5, and Item 6 are positively keyed (1 = 1, 2 = 2, 3 = 3, 4 = 4, 5 = 5).
  • Reverse-Scored Items: Item 1, Item 3, and Item 4 reflect low innovativeness statements and must be reversed prior to aggregation (5 = 1, 4 = 2, 3 = 3, 2 = 4, 1 = 5). (Note: In the original 1991 published text key, items expressing non-innovative behaviors are reverse-scored so that higher summed values universally reflect greater innovativeness).
  • Total Score Calculation: Items are summed after reverse-scoring. Scores range from a minimum of 6 to a maximum of 30. Alternatively, researchers may calculate the mean score across the six items (ranging from 1.0 to 5.0) to facilitate interpretation on the original Likert metric.
• Score Interpretation:
  • 6 to 14 (Low Innovativeness / Laggards & Late Majority): Consumers exhibiting pronounced resistance to novelty, high reliance on social proof, and preference for mature, familiar products.
  • 15 to 21 (Moderate Innovativeness / Early Majority): Pragmatic consumers who adopt innovations once mainstream viability and interpersonal utility are established.
  • 22 to 30 (High Innovativeness / Innovators & Early Adopters): Proactive, novelty-seeking consumers who eagerly acquire category innovations early, tolerate adoption risks, and serve as market opinion leaders.
• Adaptation Protocol: Researchers must replace the bracketed placeholder “[product category]” with the specific category under investigation (e.g., “electric vehicles,” “smartphones,” “streaming services,” “craft beer,” “athletic footwear”) prior to administration.

Permissions & Fee and Test Year

The Domain-Specific Innovativeness Scale was formally published in 1991 in the Journal of the Academy of Marketing Science:

• Publication Year: 1991
• Copyright Holder: Academy of Marketing Science / Springer Nature
• Academic and Non-Commercial Use: The scale items and scoring protocol are in the public academic domain for non-commercial research, scholastic inquiry, thesis and dissertation projects, and scientific investigations, provided that appropriate scholarly attribution and standard academic citation are accorded to Goldsmith and Hofacker (1991).
• Commercial and Proprietary Use: Commercial market research firms, corporate consultancies, or commercial software developers integrating the DSI into proprietary diagnostics should consult standard copyright compliance guidelines governing academic publications or contact the authors directly through Florida State University.
• Fees: No licensing fees are required for scholarly, university-based academic research.

References

• Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84(5), 888–918. https://doi.org/10.1037/0033-2909.84.5.888
• Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7(2), 117–140. https://doi.org/10.1177/001872675400700202
• Flynn, L. R., & Goldsmith, R. E. (1993). A validation of the Goldsmith and Hofacker domain-specific innovativeness scale. Educational and Psychological Measurement, 53(4), 1105–1116. https://doi.org/10.1177/0013164493053004028
• Goldsmith, R. E., Freiden, J. B., & Eastman, J. K. (1995). The central role of interest in consumer innovativeness. Journal of Social Behavior and Personality, 10(4), 775–784.
• Goldsmith, R. E., & Hofacker, C. F. (1991). Measuring consumer innovativeness. Journal of the Academy of Marketing Science, 19(3), 209–221. https://doi.org/10.1007/BF02726497
• Hurt, H. T., Joseph, K., & Cook, C. D. (1977). Scales for the measurement of innovativeness. In P. Widener (Ed.), Communication Yearbook 1 (pp. 58–65). International Communication Association.
• Kirton, M. (1976). Adaptors and innovators: A description and measure. Journal of Applied Psychology, 61(5), 622–629. https://doi.org/10.1037/0021-9010.61.5.622
• Marsh, H. W. (1996). Positive and negative global self-esteem: A substantively meaningful distinction or artifactors? Journal of Personality and Social Psychology, 70(4), 810–819. https://doi.org/10.1037/0022-3514.70.4.810
• Merton, R. K. (1968). Social Theory and Social Structure (Enlarged ed.). Free Press.
• Midgley, D. F., & Dowling, G. R. (1978). Innovativeness: The concept and its measurement. Journal of Consumer Research, 4(4), 229–242. https://doi.org/10.1086/208701
• Robertson, T. S. (1971). Innovative Behavior and Communication. Holt, Rinehart and Winston.
• Rogers, E. M. (1983). Diffusion of Innovations (3rd ed.). Free Press.
• Zaichkowsky, J. L. (1985). Measuring the involvement construct. Journal of Consumer Research, 12(3), 341–352. https://doi.org/10.1086/208520

13. Items of the Scale (Questionnaire)