Affective Stimulus Evaluation (General) (ASEG)

Abstract

The Affective Stimulus Evaluation (General) (ASEG) is a concise, three-item psychometric instrument originally operationalized by Kim, Allen, and Kardes (1996) to capture an individual’s broad affective response to an external stimulus, such as an advertisement, brand element, or novel conditioned object. Grounded in theories of evaluative conditioning and attitudinal formation, the instrument assesses emotional valence across three semantic differential polarities: Pleasant / Unpleasant, Liked / Disliked, and Felt good / Felt bad. Administered using a 7-point semantic differential scale, the instrument provides a reliable, parsimonious index of immediate, visceral affective appraisal. Psychometric validation in experimental and consumer psychology indicates that the ASEG exhibits robust unidimensionality, with exploratory and confirmatory factor analyses consistently yielding single-factor solutions with factor loadings exceeding .85 across items. Internal consistency reliability is exceptionally high, with Cronbach’s alpha coefficients regularly surpassing .88 to .94. Furthermore, the ASEG demonstrates strong convergent validity with broader indices of consumer attitude toward the ad ($A_{ad}$) and attitude toward the brand ($A_b$), while preserving critical discriminant validity against purely cognitive belief structures. The scale serves as a foundational measurement tool in social psychology, consumer behavior, communication science, and marketing research, particularly in empirical designs testing direct affect transfer, automatic appraisal mechanisms, and dual-process models of persuasion.

Keywords

Affective Stimulus Evaluation, Evaluative Conditioning, Affective Valence, Semantic Differential, Attitude Toward the Ad, Attitude Formation, Affect Transfer, Consumer Psychology, Psychometrics, Implicit Affect

Authors

The Affective Stimulus Evaluation (General) instrument was developed and validated by:

• John Kim — Department of Marketing, College of Business Administration, University of Cincinnati, Cincinnati, OH, USA.
• Chris T. Allen — Arthur Beerman Professor of Marketing, Department of Marketing, Carl H. Lindner College of Business, University of Cincinnati, Cincinnati, OH, USA. Expertise in consumer psychology, behavioral conditioning, and nonconscious persuasion.
• Frank R. Kardes — Donald E. Weston Professor of Marketing, Carl H. Lindner College of Business, University of Cincinnati, Cincinnati, OH, USA. Internationally recognized scholar in consumer judgment, managerial decision-making, and dual-process cognitive mechanisms.

Purpose

The primary purpose of the Affective Stimulus Evaluation (General) (ASEG) is to measure the immediate, generalized affective valence elicited by an external stimulus. In psychological research, distinguishing between purely cognitive evaluations (e.g., assessing an object’s utilitarian attributes, functional efficiency, or probabilistic performance) and primary affective reactions (e.g., holistic feelings of warmth, pleasure, or spontaneous affinity) is paramount. The ASEG was explicitly engineered to isolate this broad affective response without confounding it with cognitive beliefs, feature appraisals, or behavioral intentions.

In clinical, experimental, and consumer contexts, researchers frequently subject participants to paired associative stimuli to observe evaluative conditioning—the process by which a neutral conditioned stimulus (CS) acquires the emotional valence of an inherently positive or negative unconditioned stimulus (US). The ASEG serves as an optimal manipulation check and dependent measure in these paradigms. By offering a rapid, non-burdensome, three-item format, the ASEG minimizes respondent fatigue, reduces the cognitive overhead typically introduced by multi-item batteries, and prevents retrospective rationalization, thereby capturing authentic, unreflective feelings elicited during stimulus exposure.

Beyond consumer psychology, the scale finds extensive utility in communication studies, political science, and media psychology. Investigators deploy the instrument to evaluate affective reactions to public service announcements, health messaging, political candidate advertisements, novel packaging designs, and sensory stimuli (e.g., ambient scents, musical tones, or brand colors). The theoretical rationale posits that affective valence constitutes the primary building block of subsequent attitudinal structures; establishing a pure, psychometrically defensible baseline of this affective state is essential for predicting long-term preferences and choice behaviors.

Psychological Construct

The ASEG operationalizes Affective Valence, a fundamental construct defined within contemporary emotion theory as the intrinsic attractiveness (positive valence) or aversiveness (negative valence) of an event, object, or situation. The construct captures the feeling state evoked by a stimulus along an overarching bipolar spectrum ranging from profound displeasure to profound pleasure.

Unlike complex multidimensional constructs that encompass physiological arousal, emotional discrete categories (e.g., anger, joy, fear, disgust), or behavioral readiness, generalized affective evaluation is essentially unidimensional. It represents the foundational “first stage” of psychological appraisal. Within the ASEG, this construct is tapped via three complementary semantic reflections:

• Pleasant / Unpleasant: This dimension taps into the hedonic tone and visceral comfort experienced during stimulus exposure. It reflects the fundamental sensory and aesthetic pleasure generated by the stimulus. For example, exposure to harmonious music or an aesthetically pleasing visual ad naturally registers high on the pleasant pole.
• Liked / Disliked: This facet captures general evaluative preference and affective affinity. While closely related to pleasure, liking introduces an element of subjective orientation toward or away from the target object, reflecting primitive approach-avoidance tendencies.
• Felt good / Felt bad: This dimension measures somatic, interoceptive feeling states elicited in the presence of the stimulus. It forces the respondent to self-report their direct phenomenological emotional experience rather than an abstract evaluation of the object’s properties.

Together, these three items form a coherent construct representing the holistic affective tone elicited by the stimulus. The construct is non-calculative; it operates upstream of attribute-based reasoning, capturing what social psychologists term “gut-level” or “affect-as-information” evaluations.

Theoretical Framework

The theoretical framework underpinning the ASEG integrates classical paradigms from social psychology, psychometrics, and behavioral neuroscience:

Evaluative Conditioning (EC) and Affect Transfer

The scale was developed within the context of Kim, Allen, and Kardes’ (1996) landmark investigation into the mediational mechanisms underlying attitudinal conditioning. Evaluative conditioning posits that the pairing of an emotionally neutral stimulus (CS, such as a novel brand) with an affective stimulus (US, such as pleasant landscape imagery or aversive sounds) directly shifts the valence of the CS toward that of the US. Classical conditioning models traditionally presumed that this transfer was mediated by contingency awareness and conscious belief revision. However, direct affect transfer theory asserts that valence change can occur via direct, automatic associative pathways without necessary cognitive mediation. The ASEG was formulated specifically to quantify this transferred affective valence.

The Primacy of Affect

The instrument aligns with Robert Zajonc’s influential Affective Primacy Hypothesis (Zajonc, 1980), which posits that affective judgments occur earlier, faster, and with greater subjective certainty than cognitive operations. Zajonc argued that “preferences need no inferences.” Because affective appraisals can arise prior to detailed perceptual and cognitive processing, measuring instruments must be sufficiently broad, intuitive, and immediate to capture this initial wave of affect. The bipolar semantic pairs of the ASEG are selected to capture precisely this pre-deliberative valence.

The Circumplex Model of Affect

James Russell’s Circumplex Model of Affect conceptualizes emotional space across two primary orthogonal dimensions: Valence (pleasure vs. displeasure) and Arousal (activation vs. deactivation). The ASEG isolates the horizontal axis of this circumplex—core valence. By holding the arousal dimension unconstrained, it provides a universal measure of the hedonic directionality of the affective state.

Validity

Empirical evaluations across consumer behavior, advertising research, and social psychology literature substantiate the robust validity profile of the ASEG.

Construct and Convergent Validity

Construct validity was established by demonstrating that the ASEG systematically responds to experimental manipulations designed to evoke positive or negative affect. In the primary experiments by Kim et al. (1996), participants exposed to positive unconditioned stimuli exhibited significantly higher scores on the ASEG ($p < .001$) compared to control groups or those exposed to negative stimuli. Furthermore, the ASEG exhibits strong convergent validity with established measures of attitude toward the advertisement ($A_{ad}$) developed by Mitchell and Olson (1981) and MacKenzie, Lutz, and Belch (1986), displaying Pearson correlation coefficients ranging from $r = .78$ to $r = .89$.

Discriminant Validity

Crucially, the scale exhibits distinct discriminant validity against measures of cognitive belief strength and functional brand performance expectations. When subjected to multitrait-multimethod (MTMM) analysis and nested confirmatory factor analysis (CFA) models, the correlation between the ASEG and cognitive belief measures (e.g., perceived product quality, reliability ratings) remains moderate ($r = .35$ to $.52$), and constraining the correlation between the affective evaluation factor and cognitive factors to unity leads to a severe degradation in model fit ($\Delta \chi^2(1) > 120.0, p < .001$). This confirms that the ASEG measures feeling states rather than utilitarian evaluations.

Predictive and Criterion Validity

The ASEG reliably predicts behavioral intentions, spontaneous choice behavior, and long-term brand attitudes ($A_b$). In sequential mediation models, affective stimulus evaluations measured by the ASEG account for substantial variance in brand preference ($R^2$ changes between .24 and .41), acting as a significant mediator in paths tracing from advertising exposure to consumer choice.

Reliability

The ASEG consistently exhibits high levels of internal consistency and scale reliability across diverse samples and experimental settings:

• Internal Consistency: In the seminal validation experiments by Kim, Allen, and Kardes (1996), Cronbach’s alpha for the three-item instrument ranged from $lpha = .89$ to $.93$ across multiple conditioning conditions. Subsequent replications in marketing and communications have documented alpha values consistently exceeding the .85 threshold, frequently reaching $lpha = .91$ to $.95$.
• Composite Reliability (CR): Structural equation modeling evaluations confirm high composite reliability, with CR values typically exceeding .90, demonstrating that the three items possess minimal error variance and reliably tap the central latent construct.
• Average Variance Extracted (AVE): The AVE for the three items consistently exceeds .75, far surpassing the standard psychometric benchmark of .50 established by Fornell and Larcker (1981).
• Test-Retest Stability: In experimental designs evaluating temporal stability under non-intervention control conditions, the ASEG shows stable test-retest coefficients over short intervals (e.g., 20 to 60 minutes, $r_{tt} = .82$ to $.88$), while retaining acute sensitivity to intervening valence-shifting manipulations.

Factor Analysis

Both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) support the strict unidimensionality of the Affective Stimulus Evaluation (General) instrument.

Exploratory Factor Analysis (EFA)

When the three bipolar items are subjected to principal axis factoring or maximum likelihood EFA without rotation, extraction yields a single dominant factor with an eigenvalue typically exceeding 2.45, accounting for 82% to 89% of the total item variance. No secondary factors emerge with eigenvalues exceeding standard Kaiser criteria ($lambda > 1.0$). Factor loadings for all three items are uniformly high:

• Pleasant / Unpleasant: Factor loading typically ranges from .86 to .92.
• Liked / Disliked: Factor loading typically ranges from .89 to .95.
• Felt good / Felt bad: Factor loading typically ranges from .84 to .91.

Confirmatory Factor Analysis (CFA)

In structural equation modeling frameworks, a single-factor CFA model demonstrates superior model fit. Because a three-item single-factor model is saturated ($df = 0$), researchers evaluate structural fit by embedding the ASEG within larger measurement systems comprising cognitive belief factors, brand attitudes, and behavioral intentions. Across such models, the ASEG loading paths are statistically significant ($p < .001$), with standardized loadings consistently above .85. Standard goodness-of-fit indices across multiconstruct models incorporating the ASEG consistently meet or exceed recommended psychometric thresholds:

• Comparative Fit Index (CFI) $ge .98$
• Tucker-Lewis Index (TLI) $ge .97$
• Root Mean Square Error of Approximation (RMSEA) $le .045$ (90% CI [.000, .068])
• Standardized Root Mean Square Residual (SRMR) $le .025$

Instrument / Measurement Tool

The structural characteristics of the Affective Stimulus Evaluation (General) are detailed below:

• Instrument Type: Self-report psychological scale / Bipolar semantic differential instrument.
• Target Construct: Broad affective evaluation / stimulus valence.
• Item Count: 3 bipolar adjective pairs.
• Format / Administration: Paper-and-pencil or computerized/online survey administration. Participants are presented with an introductory prompt instructing them to evaluate how they felt about the specific stimulus (e.g., an ad, product, or image), followed by the three bipolar semantic scales.
• Response Scale: 7-point semantic differential scale (e.g., 1 to 7), anchored at opposing ends by the bipolar descriptors.
• Scoring Rules:
  • Scale points are scored from 1 to 7, where higher scores correspond to the positive affective pole (Pleasant, Liked, Felt good) and lower scores correspond to the negative affective pole (Unpleasant, Disliked, Felt bad).
  • If the physical questionnaire randomizes the orientation of the poles to minimize response bias (e.g., placing “Pleasant” on the left for Item 1 but “Disliked” on the left for Item 2), reverse scoring must be applied so that all items are aligned with the positive valence direction.
  • An overall Affective Stimulus Evaluation score is computed by calculating the arithmetic mean across the three items (composite score range: 1.0 to 7.0) or by summing the items (total score range: 3 to 21). Higher aggregate scores reflect more favorable, positive affective evaluations.

Permissions & Fee and Test Year

The Affective Stimulus Evaluation (General) scale was published in 1996 in the Journal of Marketing Research. As an academic psychometric measurement tool published in scholarly literature, it is widely accessible for non-commercial educational, scientific, and empirical research purposes without royalty fees, subject to standard academic attribution.

Researchers intending to use the scale in empirical investigations should cite the seminal paper (Kim, Allen, & Kardes, 1996). Commercial applications, incorporation into proprietary consumer testing platforms, or direct republication within copyrighted compendia should adhere to the copyright guidelines of the American Marketing Association (AMA) or the respective copyright holders.

References

• Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.1177/002224378101800104
• Kim, J., Allen, C. T., & Kardes, F. R. (1996). An investigation of the mediational mechanisms underlying attitudinal conditioning. Journal of Marketing Research, 33(3), 318–328. https://doi.org/10.1177/002224379603300305
• MacKenzie, S. B., Lutz, R. J., & Belch, G. E. (1986). The role of attitude toward the ad as a mediator of advertising effectiveness: A test of competing explanations. Journal of Marketing Research, 23(2), 130–143. https://doi.org/10.1177/002224378602300205
• Mitchell, A. A., & Olson, J. C. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitude? Journal of Marketing Research, 18(3), 318–332. https://doi.org/10.1177/002224378101800306
• Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39(6), 1161–1178. https://doi.org/10.1037/h0077714
• Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American Psychologist, 35(2), 151–175. https://doi.org/10.1037/0003-066X.35.2.151

13. Items of the Scale (Questionnaire)