Controllability (CON)

Abstract

The Controllability (CON) scale, developed and validated by Ronald L. Hess Jr., Shankar Ganesan, and Noreen M. Klein (2007), is a psychometric instrument designed to assess an individual’s perception of the degree to which an adverse event, specifically a service failure or organizational breakdown, was preventable and subject to the volitional power of a focal firm or service provider. Rooted in attribution theory, the scale operationalizes the distinct theoretical dimension of controllability as differentiated from locus of causality and temporal stability. The instrument comprises three self-report items evaluated on a 7-point Likert response format ranging from 1 (“Strongly Disagree”) to 7 (“Strongly Agree”). Extensive empirical validation within structural equation modeling (SEM) frameworks demonstrates robust psychometric properties, including high internal consistency reliability (Cronbach’s $\alpha ge .88$; Composite Reliability $ge .89$), high standardized factor loadings exceeding $.80$, and strong convergent and discriminant validity against related constructs such as stability, locus of causality, customer anger, and repurchase intentions. The scale serves as an indispensable tool in organizational behavior, consumer psychology, crisis management, and services marketing research to evaluate post-failure appraisals, blame attribution, and the efficacy of service recovery interventions.

Keywords

Controllability attribution, attribution theory, service failure, service recovery, perceived control, consumer behavior, organizational attributions, blame attribution, psychometrics, customer satisfaction

Authors

The Controllability (CON) scale was authored by researchers specializing in marketing strategy, relationship marketing, and consumer decision-making:

• Ronald L. Hess Jr. — Associate Professor of Marketing, Raymond A. Mason School of Business, College of William & Mary, Williamsburg, Virginia, USA. Primary research domains include customer relationship management, service failure and recovery, and sales force management.
• Shankar Ganesan — Professor of Marketing, Mendoza College of Business, University of Notre Dame, Notre Dame, Indiana, USA. Expertise spans interorganizational relationships, buyer-seller interactions, service marketing, and organizational attribution processes.
• Noreen M. Klein — Associate Professor of Marketing, Pamplin College of Business, Virginia Tech, Blacksburg, Virginia, USA. Specializes in consumer judgment, behavioral decision theory, and cognitive processing in marketing contexts.

Purpose

The primary purpose of the Controllability (CON) scale is to quantify an observer’s or customer’s cognitive attribution regarding whether an organization possessed the capacity, authority, resources, and foresight to avert a negative event. While adverse encounters—such as operational delays, interpersonal service breakdowns, product defects, or corporate crises—inevitably occur across organizational ecosystems, stakeholder psychological reactions are dictated not merely by the severity of the outcome, but by the underlying causal attributions assigned to the event.

In classical attribution paradigms, individuals function as intuitive psychologists who seek to diagnose the cause of unexpected or negative events. Although early research frequently conflated locus of causality (whether the cause is internal or external to the actor) with controllability (whether the cause is volitionally governable by the actor), modern psychometrics necessitates the empirical separation of these facets. An event may originate internally within a firm (e.g., a sudden mechanical failure caused by an undetected metallurgical flaw) yet remain uncontrollable; conversely, an event may involve external entities but remain within the firm’s preventative jurisdiction through adequate contingency planning or vendor oversight.

The scale was developed to isolate the volitional governability dimension. Specifically, it assesses:

• The perceived capacity and leverage of the firm to dictate the operational outcome.
• The preventability of the failure through reasonable organizational diligence or procedural safeguards.
• The overarching placement of the outcome within the firm’s managerial control.

In academic research, the instrument functions as a critical mediator or moderator within structural models of consumer forgiveness, affective reactions (e.g., righteous indignation, anger), retaliatory behaviors (e.g., negative word-of-mouth, boycotting), and compensatory expectations. In applied organizational settings, measuring controllability perceptions assists managers in calibrating service recovery strategies: high perceived controllability typically mandates substantive apologies, transparent root-cause disclosures, and financial compensation, whereas low perceived controllability may be effectively managed through relational reassurance and informational explanations.

Psychological Construct

The Controllability (CON) construct represents a specific cognitive appraisal within the multidimensional architecture of causal attribution. Drawing heavily from Bernard Weiner’s foundational attribution framework, causal explanations for outcomes can be systematically decomposed along three orthogonal dimensions:

1. Locus of Causality: Pertains to the ontological origin of the event—whether the primary catalyst resides internally within the target actor or externally in the environment or third parties.
2. Stability: Pertains to the temporal endurance and recurrence potential of the cause—whether the underlying factors are invariant over time or transient and fleeting.
3. Controllability: Pertains to the degree of volitional command, influence, or preventability that the actor can exert over the causal mechanisms producing the outcome.

The CON scale measures this third dimension with high theoretical purity. Within the context of service and organizational failures, controllability reflects a tripartite cognitive assessment:

1. Power to Influence (Capacity)

The first dimension addresses the structural and physical capability of the firm to alter the trajectory of events. When customers evaluate whether a firm had the “power to influence this outcome,” they evaluate organizational agency, technological capacity, resource availability, and domain expertise. A complete absence of power shifts attribution toward inevitable systemic constraints or force majeure.

2. Preventability (Foreseeability and Actionability)

The second facet examines whether proactive measures could have successfully intercepted the failure. Preventability captures elements of vigilance, managerial foresight, and adequate standard operating procedures. An event perceived as entirely preventable triggers heightened moral culpability, as the failure is attributed to managerial negligence, apathy, or structural incompetence rather than unavoidable chance.

3. Sphere of Control (Volitional Governance)

The final facet synthesizes general boundary conditions of control. It determines whether the final state of affairs was enclosed within the firm’s operational and normative sphere of responsibility. High perceived control signals that the outcome was not the product of external stochastic noise, but directly dependent on the focal actor’s choices, policies, and deliberate allocations of effort.

Theoretical Framework

The theoretical bedrock of the Controllability (CON) scale is Weiner’s Cognitive Attribution Theory of Achievement Motivation and Emotion (Weiner, 1985, 1986). Weiner posited that when individuals encounter unexpected or goal-inhibiting events, they engage in systematic causal searches. The causal inferences drawn along the dimensions of locus, stability, and controllability subsequently dictate emotional experiences, cognitive appraisals of fairness, and future behavioral intentions.

Valerie Folkes (1984, 1988) extended Weiner’s paradigm directly into consumer behavior, establishing that causal attributions fundamentally govern customer evaluations following service failures. Folkes demonstrated that controllability attributions act as the primary catalyst for moral outrage and attribution of blame. When a consumer infers that an organization caused a problem and possessed the unexercised power to prevent it, the psychological contract is perceived as intentionally violated.

Hess, Ganesan, and Klein (2007) integrated this attributional foundation with Social Exchange Theory and the concept of customer “pseudorelationships” (interactions where customers lack a dedicated personal relationship with an individual employee but maintain repeated encounters with a single firm). In such settings, customers rely heavily on firm-level organizational attributions. When an interactional service failure occurs, perceived controllability leads to:

• Cognitive Dissonance and Equity Violations: The customer perceives that their investment of capital and trust has been met with insufficient organizational diligence.
• Intensified Attribution of Blame: Controllability operates as the necessary condition for allocating culpability. Without perceived controllability, failure is perceived as misfortune; with high controllability, failure is coded as fault.
• Attenuated Relationship Commitment: High perceived controllability actively erodes the protective buffer of long-term customer relationships, precipitating sharp declines in repurchase intentions and sharp spikes in negative word-of-mouth.

Validity

The psychometric validity of the Controllability (CON) scale has been established through multiple laboratory experiments, cross-sectional field studies, and rigorous confirmatory modeling procedures.

Construct and Convergent Validity

Convergent validity evaluates whether the three scale items adequately converge upon a single underlying construct. In the validation study by Hess et al. (2007), all three standardized factor loadings were statistically significant ($p < .001$) and exceeded the recognized threshold of $.70$ (with loadings consistently between $.80$ and $.92$). The Average Variance Extracted (AVE) by the construct consistently surpasses $.70$, well above the classical Fornell-Larcker benchmark of $.50$, establishing that the scale captures substantial variance attributable to the true underlying construct rather than measurement error.

Discriminant Validity

Crucial to attributional instrumentation is demonstrating that controllability is empirically distinct from stability (whether the failure occurs repeatedly) and locus of causality (whether the firm or customer caused the issue). Hess et al. (2007) conducted pairwise confirmatory factor analysis chi-square difference tests ($\Delta\chi^2$) comparing constrained models (where the correlation between CON and related attributional dimensions was set to $1.0$) with unconstrained models. The unconstrained models exhibited statistically superior fit ($p < .001$). Furthermore, the square root of the AVE for the CON construct exceeded the inter-construct correlations, confirming robust discriminant validity according to Fornell-Larcker criteria and contemporary Heterotrait-Monotrait (HTMT) ratios of correlations.

Nomological and Predictive Validity

Nomological validity is demonstrated through the scale’s predictable relationships with upstream antecedents and downstream affective and behavioral variables. Empirical testing confirms that experimental manipulations of failure causes (e.g., internal system bugs vs. third-party network outages) successfully drive corresponding variations in CON scores. Downstream, elevated CON scores consistently and significantly predict:

• Increased consumer anger and felt betrayal ($r \approx .45$ to $.65, p < .001$).
• Decreased overall customer satisfaction and service quality perceptions.
• Heightened expectations for substantial organizational recovery efforts (e.g., compensatory refunds rather than mere social explanations).
• Depressed repurchase and retention intentions across varying relationship durations.

Reliability

The internal consistency of the Controllability (CON) scale has demonstrated exceptional stability across various empirical replications and sample populations.

In the original validation studies by Hess, Ganesan, and Klein (2007), the scale yielded internal consistency estimates that consistently exceeded recommended thresholds for psychological instrumentation:

• Cronbach’s Alpha ($\alpha$): Reported values across experimental scenarios ranged between $.88$ and $.93$, indicating superior internal consistency without redundant item bloat.
• Composite Reliability (CR): Structural equation modeling iterations demonstrated composite reliabilities ranging from $.89$ to $.94$, indicating that the latent variable is captured with high precision.
• Item-to-Total Correlations: Corrected item-to-total correlations for each of the three items systematically exceeded $.75$, confirming that each item contributes robust variance to the overarching composite.

Because the scale is brief (three items), these elevated alpha levels are particularly notable, as Cronbach’s alpha is inherently sensitive to scale length. This demonstrates that the high reliability is driven by strong inter-item correlations and minimal unique error variance rather than scale length.

Factor Analysis

Confirmatory Factor Analysis (CFA) conducted in structural equation modeling environments (e.g., LISREL, AMOS, Mplus) confirms that the Controllability (CON) scale is strictly unidimensional.

Factor Loadings and Parameter Estimates

When specified as a single latent factor with three manifest indicators, typical standardized parameter estimates are observed as follows:

• Item 1: “[The firm/service provider] had the power to influence this outcome.” — Standardized Loading ($lambda$) $\approx .82$ to $.87$.
• Item 2: “[The firm/service provider] could have prevented this problem from occurring.” — Standardized Loading ($lambda$) $\approx .88$ to $.93$.
• Item 3: “The outcome was within [the firm/service provider’s] control.” — Standardized Loading ($lambda$) $\approx .85$ to $.90$.

Model Fit Indices

Because a three-indicator single-factor measurement model possesses zero degrees of freedom ($df = 0$), it is statistically just-identified (saturated). However, when evaluated within comprehensive multi-factor measurement models alongside stability, failure severity, customer expectations, and relationship quality, the CON indicators exhibit negligible cross-loadings, zero substantial error covariances, and facilitate overall model fits well within conventional cutoff criteria:

• Comparative Fit Index (CFI): $ge .97$
• Tucker-Lewis Index (TLI): $ge .96$
• Root Mean Square Error of Approximation (RMSEA): $le .05$ (with $90%$ confidence intervals ranging from $.02$ to $.07$)
• Standardized Root Mean Square Residual (SRMR): $le .04$

These empirical findings verify that the construct space of perceived controllability is fully captured by the single-factor three-item model without multidimensional splitting.

Instrument / Measurement Tool

• Instrument Name: Controllability Scale (CON)
• Authors: Ronald L. Hess Jr., Shankar Ganesan, and Noreen M. Klein
• Publication Year: 2007
• Target Construct: Perceived Controllability Attribution (Firm/Provider Level)
• Instrument Type: Self-report psychometric scale (unidimensional)
• Number of Items: 3 items
• Administration Format: Paper-and-pencil or online digital survey; suitable for scenario-based experimental paradigms or retrospective recall of actual service breakdowns.
• Target Population: Consumers, organizational stakeholders, service recipients, and business partners.
• Response Scale: 7-point Likert scale (1 = Strongly Disagree to 7 = Strongly Agree)
• Scoring Protocol: All three items are positively keyed (no reverse scoring is necessary). Individual item responses are averaged (or summed) to yield a mean index of perceived controllability ranging from 1.00 to 7.00, where higher scores indicate stronger attribution of control, power, and preventability to the focal organization.

Permissions & Fee and Test Year

The Controllability (CON) scale was published in 2007 in the Journal of Retailing by Elsevier. Under standard academic fair use principles, the scale items may be utilized, adapted, and administered for non-commercial academic, psychological, and institutional research purposes without financial charge, provided proper scholarly citation is given to Hess, Ganesan, and Klein (2007).

For commercial applications, proprietary consulting interventions, or reproduction within commercial publications, permissions should be formally cleared through the copyright holder, Elsevier Inc., or via the Copyright Clearance Center (CCC). Inquiries regarding original experimental paradigms or conceptual expansions may be directed to the corresponding authors at their respective academic institutions.

References

• Bitner, M. J. (1990). Evaluating service encounters: The effects of physical surroundings and employee responses. Journal of Marketing, 54(2), 69–82. https://doi.org/10.1177/002224299005400206
• Folkes, V. S. (1984). Consumer reactions to product failure: An attributional approach. Journal of Consumer Research, 10(4), 398–409. https://doi.org/10.1086/208978
• Folkes, V. S. (1988). Recent attribution research in consumer behavior: A review and new directions. Journal of Consumer Research, 14(4), 548–565. https://doi.org/10.1086/209135
• 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
• Hess Jr., R. L., Ganesan, S., & Klein, N. M. (2007). Interactional service failures in a pseudorelationship: The role of organizational attributions. Journal of Retailing, 83(1), 79–95. https://doi.org/10.1016/j.jretai.2006.10.006
• Weiner, B. (1985). An attributional theory of achievement motivation and emotion. Psychological Review, 92(4), 548–573. https://doi.org/10.1037/0033-295X.92.4.548
• Weiner, B. (1986). An Attributional Theory of Motivation and Emotion. Springer-Verlag. https://doi.org/10.1007/978-1-4612-4948-1

Items of the Scale