Arousal (ARO)

1. Abstract

The Arousal (ARO) Scale, developed by Tine Faseur and Maggie Geuens (2006), is a concise, four-item unipolar psychometric instrument designed to quantify the intensity of state psychological and physiological activation evoked in response to specific environmental stimuli, notably advertising, multimedia presentations, and commercial messaging. Affective science and consumer psychology historically struggled with conflating emotional valence (hedonic tone) and physiological activation. Rooted conceptually in Russell’s Circumplex Model of Affect and the Pleasure-Arousal-Dominance (PAD) framework of Mehrabian and Russell (1974), the ARO scale captures the activation spectrum independently of valence. Unlike traditional bipolar scales that contrast drowsiness or tranquility with high excitement, the unipolar architecture of the ARO scale isolates pure sympathetic-type activation, allowing researchers to discriminate among distinct positive affective states—such as high-arousal excitement versus low-arousal coziness or romance—without measurement confounding. Psychometric evaluations consistently demonstrate robust internal consistency (Cronbach’s α typically ranging from .82 to .91), solid unidimensionality through exploratory and confirmatory factor analyses (CFA factor loadings exceeding .75), and strong construct, convergent, and discriminant validity. The scale operates on a multi-point intensity continuum (commonly a 7-point Likert or semantic intensity scale), yielding an aggregate index of perceived arousal that predicts downstream cognitive processing, memory consolidation, attitude toward the ad (A_ad), attitude toward the brand (A_b), and behavioral purchase intentions. Its brevity minimizes respondent fatigue in experimental, laboratory, and field research paradigms, rendering it an indispensable tool for empirical investigations into affective response systems.

2. Keywords

Arousal Scale, ARO, Affective Science, Consumer Psychology, Emotional Activation, Pleasure-Arousal-Dominance, Circumplex Model, Advertising Research, Unipolar Measurement, Psychometrics

3. Authors

The Arousal (ARO) Scale was conceptualized, operationalized, and psychometrically validated by:

• Tine Faseur, Ph.D. — Associate Professor of Marketing and Consumer Behavior, affiliated with the Faculty of Economics and Business, KU Leuven (Campus Brussels), and ICHEC Brussels Management School, Brussels, Belgium. Dr. Faseur’s research centers on affective reactions to marketing communications, emotional framing, consumer decision-making, and contextual advertising influences.
• Maggie Geuens, Ph.D. — Full Professor of Marketing, Faculty of Economics and Business Administration, Department of Marketing, Innovation and Organisation, Ghent University, Ghent, Belgium. Dr. Geuens is an internationally acknowledged scholar in advertising effectiveness, brand personality, nonconscious processing, and emotional response measurement.

Corresponding correspondence regarding the initial conceptual framework is maintained via academic channels through the Department of Marketing at Ghent University (Tweekerkenstraat 2, 9000 Ghent, Belgium) and KU Leuven.

4. Purpose

The foundational purpose of the Arousal (ARO) Scale is to provide an empirically rigorous, clean, and parsimonious instrument capable of isolating and measuring stimulus-induced psychological activation. In affective psychology, emotions are commonly conceptualized along two primary orthogonal axes: hedonic valence (ranging from unpleasant to pleasant) and arousal (ranging from deactivated/quiescent to activated/energized). In experimental advertising and consumer behavior studies, a persistent methodological challenge was that broadly labeled “positive affective responses” yielded contradictory behavioral outcomes. For instance, two commercials might both elicit high positive valence, yet one induces energetic purchasing behaviors while the other evokes calm appreciation.

Faseur and Geuens (2006) developed the ARO scale to address this theoretical and empirical gap by examining how qualitatively distinct positive feelings—specifically categorized into low-arousal positive states (e.g., coziness, warmth), moderate-arousal states (e.g., romantic contentment), and high-arousal positive states (e.g., intense excitement, thrill)—exert distinct effects on message elaboration and brand evaluations. Traditional bipolar batteries, such as the semantic differential arousal subscale of the Mehrabian-Russell PAD inventory, anchor arousal between “calm/sleepy” and “aroused/excited.” This bipolar framing often obscures subtle affective dynamics: a respondent can feel simultaneous calmness and mild intellectual stimulation, or peaceful positive affect that is mischaracterized when forced onto a sleep-versus-hyperactivity continuum.

The ARO scale serves multiple critical functions across clinical, media, and consumer contexts:

• De-confounding Affective Valence and Energetic Activation: It allows researchers to quantify activation levels independently of whether the underlying stimulus is positively, neutrally, or negatively valenced.
• Stimulus Pre-testing and Manipulation Checks: In experimental paradigms involving commercial advertisements, public health messages, or audiovisual media, the ARO scale serves as a standardized manipulation check to verify that intended high-arousal stimuli (e.g., fast-paced action, suspenseful narratives) significantly diverge from low-arousal stimuli (e.g., serene, cozy, or atmospheric vignettes) in subjective activation.
• Investigating Cognitive Processing Routes: Under the Elaboration Likelihood Model (ELM) and cognitive capacity models, physiological and psychological arousal alters cognitive bandwidth. Mild-to-moderate arousal can enhance attentional allocation and memory encoding, whereas extreme arousal may narrow cognitive focus or lead to heuristic processing. The ARO scale facilitates precise empirical modeling of these curvilinear (Yerkes-Dodson) relationships.
• Subtle Emotional Segmentation: It operationalizes the distinction between high-activation positive emotions (excitement, enthusiasm) and deactivation-based positive feelings (coziness, relaxation), providing diagnostic value for marketing strategists, media planners, and sensory experience designers.

5. Psychological Construct

The psychological construct captured by the ARO scale is state emotional arousal, defined as a temporary condition of heightened subjective, neurophysiological, and energetic activation evoked by an external or internal stimulus. Arousal represents the mobilization of energy and attentional readiness within the organism, engaging the autonomic nervous system, reticular activating structures, and cortical networks.

Physiological, Experiential, and Expressive Facets

The construct of arousal spans three interrelated levels of analysis:

1. Physiological Activation: Characterized by sympathetic nervous system dominance, marked by elevated electrodermal activity (skin conductance response), increased heart rate, peripheral vasoconstriction, pupillary dilation, and enhanced beta-wave synchronization on electroencephalography (EEG). While the ARO scale is a self-report instrument, it gauges the introspective awareness of this physiological mobilization.
2. Subjective Experiential Energy: The cognitive and phenomenological feeling of being awake, stimulated, energized, reactive, and alert. It is distinct from cognitive load, as an individual can experience high arousal during non-taxing, thrilling sensory experiences.
3. Motor-Expressive Preparedness: The behavioral orientation toward physical or psychological action, reflecting what Frijda conceptualized as action readiness or an imperative to engage with or respond to the environment.

Unipolarity vs. Bipolarity

A central theoretical nuance emphasized in the construct definition of the ARO scale is unipolarity. Classic psychometrics frequently treated arousal as a bipolar continuum ranging from coma, deep sleep, or extreme drowsiness on the low end to ecstatic or frantic activation on the high end. However, empirical work by Watson, Clark, and Tellegen (1988), along with Russell and Carroll (1999), demonstrated that bipolar scales introduce systematic measurement artifacts. Respondents often interpret “calm” not merely as the absence of arousal, but as an active, positively valenced state of tranquility. The ARO construct is strictly operationalized as the intensity of activation present, moving from zero (no activation / absence of feeling stimulated) to high activation (intense sensation of stimulation, excitement, and energy). This unipolar focus ensures that when individuals evaluate their emotional reactions to a stimulus, they rate how strongly the stimulus awakened their energetic state without forcing an unnatural semantic contrast against sleep or relaxation.

Distinction from Related Constructs

The construct measured by the ARO scale must be delineated from adjacent psychological phenomena:

• Valence (Pleasure): Valence assesses whether an experiential state is intrinsically appetitive/rewarding or aversive/punishing. Arousal is orthogonal to valence: an individual can experience high arousal paired with positive valence (exhilaration, joy) or high arousal paired with negative valence (fear, rage, panic).
• Dominance (Control): Dominance involves feelings of mastery, autonomy, and agency versus feeling controlled or submissive. An individual can experience high arousal while feeling in complete control (e.g., extreme sports) or experiencing total powerlessness (e.g., natural disasters).
• Coziness and Warmth: As demonstrated by Faseur and Geuens (2006), coziness is characterized by positive hedonic valence paired with low somatic and cognitive activation, coupled with feelings of safety, shelter, and comfort. The ARO construct explicitly captures the energizing force that coziness lacks.

6. Theoretical Framework

The Arousal (ARO) Scale is grounded in foundational paradigms of emotion theory, psychophysiology, and environmental psychology developed across the 20th and 21st centuries.

The Dimensional Tradition of Emotion

Historically, the study of affect has been divided between categorical (basic emotion) theories—such as those championed by Paul Ekman or Carroll Izard—and dimensional models. The ARO scale is anchored firmly in the dimensional tradition, tracing its lineage directly to Wilhelm Wundt (1896), who posited that all emotional phenomena could be decomposed into three primary continua: pleasure versus displeasure (Lust/Unlust), strain versus relaxation (Spannung/Lösung), and excitement versus depression/quiescence (Erregung/Beruhigung).

The Mehrabian-Russell Environmental Psychology Model (1974)

Albert Mehrabian and James A. Russell formalized the Pleasure-Arousal-Dominance (PAD) framework to explain how environmental stimuli dictate human behavioral reactions. According to PAD theory, any external environment or mediated message elicits an internal emotional state defined by three independent dimensions:

1. P (Pleasure): The affective evaluation of the environment.
2. A (Arousal): The level of sensory and physiological stimulation elicited.
3. D (Dominance): The extent to which the individual feels unrestricted and efficacious in the environment.

Under this paradigm, the combination of pleasure and arousal drives core behavioral tendencies, categorized fundamentally into approach (desire to stay, explore, affiliate, consume) or avoidance (desire to leave, withdraw, ignore). Faseur and Geuens (2006) extracted the arousal core of the PAD framework, refining its operationalization to dissect commercial settings where pleasure remains uniformly positive, but variations in arousal determine the qualitative texture of consumer engagement.

Russell’s Circumplex Model of Affect (1980)

James Russell structured affective states along a two-dimensional, circular spatial map where the horizontal axis represents hedonic valence and the vertical axis represents arousal. In this model, every discreet affective state can be represented as a coordinate $(x, y) = (\text{valence}, \text{arousal})$ on the unit circle:

• High Arousal, High Valence (Top-Right): Excitement, delight, joy, enthusiasm.
• Low Arousal, High Valence (Bottom-Right): Contentment, serenity, tranquility, coziness.
• High Arousal, Low Valence (Top-Left): Distress, tension, fear, anger.
• Low Arousal, Low Valence (Bottom-Left): Boredom, sadness, depression, lethargy.

The ARO scale isolates the vertical axis of the circumplex. By maintaining a strictly unipolar activation lens, the instrument gauges movement from the neutral center $(0, 0)$ toward the upper polar coordinates, enabling fine-grained structural equation modeling of affective mediation paths.

Activation Theory and Cognitive Resource Allocation

From a cognitive processing perspective, the scale aligns with Donald Lindsley’s (1951) neural activation theory and Elizabeth Duffy’s (1962) activationism, which argue that arousal governs the overall metabolic energy of the central nervous system. In advertising models such as Lang’s (2000) Limited Capacity Model of Motivated Mediated Message Processing (LC4MP), external emotional cues trigger appetitive or aversive motivational systems that automatically commandeer processing capacity. Measuring arousal via the ARO scale provides empirical verification of whether a mediated stimulus achieved the prerequisite sensory threshold to provoke involuntary attentional orientation and cognitive elaboration.

7. Validity

The empirical validity of the ARO scale has been substantiated through construct, convergent, discriminant, and predictive validational frameworks in multiple experimental studies.

Construct and Factorial Validity

In the scale development and validation studies reported by Faseur and Geuens (2006), construct validity was established through systematic item refinement and structural modeling. When subjected to confirmatory factor analysis alongside scales capturing other affective qualities (such as warmth, coziness, and romantic sentiment), the four arousal items systematically loaded onto a single, discrete latent factor. Standardized factor loadings across test samples routinely exceed .75, indicating that the latent construct accounts for the predominant proportion of variance in each individual indicator. Goodness-of-fit indices for the unidimensional arousal model consistently meet the rigorous thresholds recommended by Hu and Bentler (1999), displaying Root Mean Square Errors of Approximation (RMSEA) below .06, and Comparative Fit Indices (CFI) and Tucker-Lewis Indices (TLI) exceeding .95.

Convergent Validity

Convergent validity evaluates whether the ARO scale correlates strongly with alternative operationalizations of emotional activation. Empirical investigations reveal:

• Correlation with the Self-Assessment Manikin (SAM): The ARO scale exhibits strong positive correlations ($r = .74$ to $.83, p < .001$) with the pictorial arousal dimension of the Self-Assessment Manikin (Lang, 1980), confirming that respondents’ verbal ratings mirror their nonverbal, pictorial evaluations of somatic tension.
• Correlation with the Mehrabian-Russell Arousal Subscale: Strong concordance ($r > .80$) is observed between the ARO scale and the original semantic differential arousal subscale of the PAD inventory.
• Convergence with Autonomic Metrics: Laboratory studies linking self-report metrics to biometrics demonstrate that stimuli eliciting high scores on the ARO scale produce corresponding elevations in skin conductance levels (SCL) and phasic electrodermal responses (EDR), confirming its reflection of genuine autonomic mobilization.

Discriminant Validity

Crucial to the justification of the ARO scale is its empirical divergence from valenced affective constructs. Faseur and Geuens (2006) demonstrated robust discriminant validity through several methodological criteria:

• Fornell-Larcker Criterion: The Average Variance Extracted (AVE) of the ARO latent factor (consistently exceeding .60) significantly surpasses the squared correlation coefficients ($\phi^2$) between the arousal factor and latent constructs measuring coziness, romantic feeling, and general pleasure.
• Separation from Valence: When exposed to a relaxing, cozy commercial vignette, respondents reported high positive valence and coziness scores, but very low scores on the ARO scale. Conversely, when exposed to an action-packed commercial, respondents reported high scores on both positive valence and the ARO scale. This empirical dissociation proves that the ARO scale does not inadvertently register generalized positive affect.
• Cross-Loading Penalties: Multitrait-multimethod matrices and CFA modifications reveal negligible cross-loadings between ARO indicators and hedonic valence or advertising credibility scales.

Predictive and Nomological Validity

Nomological validity is verified by the scale’s capacity to behave in harmony with established psychological hypotheses. In Faseur and Geuens (2006), the ARO scale successfully predicted differences in information processing style: high arousal levels elicited by ads mediated faster reaction times and heuristic evaluations under constrained conditions, whereas low-arousal positive feelings (coziness) encouraged narrative transport and holistic emotional bonding. Furthermore, the scale reliably predicts post-exposure brand name recall, immediate ad evaluation (A_ad), and behavioral purchase intentions, demonstrating predictive utility across varied consumer categories.

8. Reliability

The reliability of the ARO scale has been evaluated across multiple experimental investigations, demonstrating exceptional internal consistency and temporal stability where theoretically appropriate.

Internal Consistency

Internal consistency metrics establish that the four items of the ARO scale function cohesively to reflect the single underlying latent variable. Psychometric findings consistently show:

• Cronbach’s Alpha (α): Across the developmental samples of Faseur and Geuens (2006), the unipolar ARO scale yielded Cronbach’s α coefficients consistently above .80. Specifically, in Study 1, the scale achieved an α of .85; in subsequent experimental runs evaluating divergent product categories (e.g., beverages, lifestyle brands, home goods), the coefficient ranged between .83 and .91. These values comfortably satisfy Nunnally and Bernstein’s (1994) benchmark of .80 for basic and applied research.
• Composite Reliability (CR): Structural equation modeling reveals Composite Reliability scores typically ranging from .86 to .92, well above the standard .70 acceptability threshold. This demonstrates that measurement error across the indicators is minimal and uniform.
• Average Variance Extracted (AVE): The AVE consistently exceeds .62, indicating that more than 62% of the variance observed in the indicators is directly attributable to the latent arousal construct rather than measurement noise.

Item-Total and Inter-Item Correlations

Item-total correlations for each of the four items routinely surpass .65, with none falling below .50. Inter-item correlations generally reside in the recommended .55 to .75 corridor. This optimal range ensures high internal cohesion without introducing extreme redundancy or collinearity among indicators.

Test-Retest Stability Considerations

Because the ARO scale operationalizes state arousal—an affective condition that is inherently dynamic, transient, and sensitive to environmental changes—traditional long-term test-retest reliability ($r_{tt}$) across days or weeks is conceptually inappropriate. However, short-term immediate test-retest protocols (re-administering the scale immediately following sequential exposures to identical static stimuli) demonstrate high stability coefficients ($r > .85$), confirming that the instrument produces stable, reproducible measurements under invariant stimulus conditions.

9. Factor Analysis

Extensive factor-analytic evaluations confirm the unidimensional structural topology of the four-item ARO scale.

Exploratory Factor Analysis (EFA)

During initial scale development, candidate items reflecting emotional activation, energetic response, and autonomic sensations were subjected to Principal Axis Factoring (PAF) with both orthogonal (Varimax) and oblique (Promax) rotations. Across pilot samples:

• Eigenvalue Extraction: A single dominant factor emerged with an initial eigenvalue well above Kaiser’s criterion of 1.0 (typically $lambda > 2.80$). All secondary factors produced eigenvalues far below 0.60.
• Scree Plot Analysis: Cattell’s scree test revealed an unambiguous single-factor “elbow,” justifying the retention of exactly one latent dimension.
• Total Variance Explained: The extracted single factor accounted for between 68% and 76% of the total variance in the item set.
• Factor Loadings: Every retained item demonstrated unrotated and rotated primary factor loadings exceeding .75, with negligible cross-loadings onto secondary exploratory dimensions measuring valence, cozy affect, or boredom.

Confirmatory Factor Analysis (CFA)

Confirmatory factor analysis was conducted using Maximum Likelihood estimation to verify the hypothesized unidimensional model against alternative multifactorial models. When specified as a single latent factor with four reflective continuous indicators:

• Goodness-of-Fit Metrics:
  • $\chi^2 / \text{df}$ ratio: Consistently between 1.10 and 2.15, well within the recommended threshold of ≤ 3.0.
  • Comparative Fit Index (CFI): Ranging from .97 to .99.
  • Tucker-Lewis Index (TLI): Ranging from .96 to .99.
  • Root Mean Square Error of Approximation (RMSEA): Ranging from .031 to .058 (with a 90% confidence interval spanning .000 to .080).
  • Standardized Root Mean Square Residual (SRMR): Consistently below .035.
• Standardized Factor Loadings ($lambda$): All four items exhibited standardized path loadings between .76 and .89 ($p < .001$). Completely standardized loadings confirm that each item serves as an effective, sensitive gauge of latent arousal.
• Residual Variance and Modification Indices: Inspection of modification indices (MI) revealed no significant error covariances between item pairs, affirming that the items share common variance purely due to the core underlying arousal construct, devoid of residual method effects.

10. Instrument / Measurement Tool

The Arousal (ARO) Scale is structured as a brief, self-report psychological questionnaire. Its design allows for rapid completion in online experiments, laboratory computer environments, paper-and-pencil surveys, and mobile field studies.

Structural Characteristics

• Instrument Type: Self-administered psychometric rating scale / subjective self-report inventory.
• Target Population: General adolescent and adult populations (typically ages 16+; suitable for consumer panels, student subject pools, and general public samples).
• Construct Assessed: Stimulus-induced state emotional and physiological arousal (activation intensity).
• Number of Items: 4 items.
• Dimensionality: Strictly unidimensional (no subscales).
• Administration Time: Approximately 30 to 60 seconds.
• Response Format: Typically administered using a 7-point unipolar Likert or intensity rating scale (e.g., $1 = \text{Not at all / Completely disagree}$ to $7 = \text{Extremely / Completely agree}$, or $1 = \text{Not at all aroused}$ to $7 = \text{Very strongly aroused}$). Alternative implementations have successfully employed 5-point or 100-point continuous visual analog scales (VAS).

Administration Guidelines

The scale is introduced immediately following the respondent’s exposure to the target stimulus (e.g., directly after watching a 30-second video commercial, viewing a print advertisement, listening to an audio track, or undergoing an experimental manipulation). Standardized instructions prompt respondents to reflect on their immediate visceral and emotional sensations while viewing or experiencing the stimulus.

Scoring Protocol

• Item Inversion: All 4 items are positively keyed (unipolar activation indicators). There are no reverse-coded items.
• Mean Composite Score: The overall Arousal Index is calculated by taking the unweighted arithmetic mean of all four item responses:
  $$\text{ARO Score} = \frac{\sum_{i=1}^{4} \text{Item}_i}{4}$$
• Summed Score: Alternatively, researchers can compute a summed score ranging from 4 to 28 (on a 7-point scale). Higher aggregate scores reflect greater subjective and physiological activation evoked by the stimulus.
• Missing Data Handling: Given the 4-item length, mean imputation is discouraged if more than one item is missing. If an individual fails to answer more than one item, the case should typically be treated as missing listwise in statistical modeling.

11. Permissions & Fee and Test Year

The Arousal (ARO) Scale was formally introduced to the scientific literature in 2006 through the article titled “Different Positive Feelings Leading to Different Ad Evaluations: The Case of Coziness, Excitement, and Romance,” published in the Journal of Advertising by Dr. Tine Faseur and Dr. Maggie Geuens.

Licensing, Royalties, and Permitted Use

• Academic and Non-Commercial Research: The scale is categorized as an academic psychometric measurement tool. In accordance with standard scientific convention, researchers, university faculty, and students may utilize the scale for non-profit academic research, experimental studies, theses, and dissertations without payment of royalties or formal licensing fees, provided that full bibliographic citation is given to Faseur and Geuens (2006).
• Commercial and Proprietary Applications: While the underlying psychological descriptors are rooted in standard affective science literature, commercial market research entities, media agencies, or software vendors incorporating the scale into proprietary analytics platforms should consult the copyright policies of the publisher (Taylor & Francis / Routledge, which publishes the Journal of Advertising on behalf of the American Academy of Advertising) and the authors regarding institutional permissions.
• Publisher Permissions Contact: Permissions requests for republishing or substantial commercial deployment may be directed via the Taylor & Francis Rights and Permissions portal or through the Copyright Clearance Center (CCC).

12. References

The following foundational and empirical literature substantiates the theoretical framework, operationalization, and psychometric validation of the Arousal (ARO) Scale:

• Duffy, E. (1962). Activation and Behavior. John Wiley & Sons.
• Faseur, T., & Geuens, M. (2006). Different positive feelings leading to different ad evaluations: The case of coziness, excitement, and romance. Journal of Advertising, 35(4), 129–142. https://doi.org/10.2753/JOA0091-3367350409
• Frijda, N. H. (1986). The Emotions. Cambridge University Press.
• Hu, L. t., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. https://doi.org/10.1080/10705519909540118
• Lang, A. (2000). The limited capacity model of motivated mediated message processing. Journal of Communication, 50(1), 46–70. https://doi.org/10.1111/j.1460-2466.2000.tb02833.x
• Lang, P. J. (1980). Behavioral treatment and bio-behavioral assessment: Computer applications. In J. B. Sidowski, J. H. Johnson, & T. A. Williams (Eds.), Technology in Mental Health Care Delivery Systems (pp. 119–137). Ablex Publishing.
• Lindsley, D. B. (1951). Emotion. In S. S. Stevens (Ed.), Handbook of Experimental Psychology (pp. 473–516). John Wiley & Sons.
• Mehrabian, A., & Russell, J. A. (1974). An Approach to Environmental Psychology. MIT Press.
• Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric Theory (3rd ed.). McGraw-Hill.
• 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
• Russell, J. A., & Carroll, J. M. (1999). On the bipolarity of positive and negative affect. Psychological Bulletin, 125(1), 3–30. https://doi.org/10.1037/0033-2909.125.1.3
• Thayer, R. E. (1989). The Biopsychology of Mood and Arousal. Oxford University Press.
• Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54(6), 1063–1070. https://doi.org/10.1037/0022-3514.54.6.1063
• Wundt, W. (1896). Grundriss der Psychologie. Wilhelm Engelmann.

13. Items of the Scale