Complaint Likelihood (General) (COML)

1. Abstract

The Complaint Likelihood (General) (COML) scale is a psychometric instrument developed by Kelli Bodey and Debra Grace (2007) to quantify a consumer’s cross-situational, general behavioral disposition to seek redress directly from a service provider or merchant following an unsatisfactory consumption experience. Rooted in consumer complaining behavior (CCB) literature and socio-cognitive dispositional frameworks, the instrument addresses a foundational distinction between situational complaining responses—which fluctuate based on context-specific service failures—and an underlying individual propensity to voice discontent. The scale comprises seven self-report items administered on a 7-point Likert scale, ranging from 1 (“Strongly Disagree”) to 7 (“Strongly Agree”). Structurally, the instrument operates as a unidimensional latent construct capturing direct voice intentions, unhesitating readiness to lodge complaints, consistency across historical episodes, and behavioral resistance or reluctance toward complaining (represented via two reverse-scored indicators). Psychometric evaluations demonstrate strong internal consistency, with Cronbach’s alpha coefficients consistently exceeding .80 (typically .84 to .88), composite reliabilities above .85, and robust convergent, discriminant, and predictive validity across diverse consumer service industries, including hospitality, retail, telecommunications, and aviation contexts. The COML scale provides marketing researchers, service managers, and organizational psychologists with a reliable measurement model to differentiate active complainers from passive non-complainers, predict churn dynamics, and investigate interactions between consumer personality and modern digital feedback channels.

2. Keywords

Consumer Complaining Behavior, Complaint Likelihood, Propensity to Complain, Voice Response, Service Recovery, Consumer Dissatisfaction, Psychometrics, Scale Validation, Exit Voice and Loyalty, Service Marketing

3. Authors

The Complaint Likelihood (General) (COML) scale was designed and validated by:

• Dr. Kelli Bodey: Department of Marketing, Griffith Business School, Griffith University, Gold Coast, Queensland, Australia. Dr. Bodey’s research focuses on services marketing, relationship marketing, consumer complaining behavior, and customer loyalty dynamics.
• Professor Debra Grace: Professor of Marketing and Founder of the Service Innovation Research Program, Griffith Business School, Griffith University, Gold Coast, Queensland, Australia. Professor Grace is an established scholar in services marketing, brand equity, retail patronage, and consumer-employee interface interactions.

Subsequent empirical extensions and operational adaptations in macro-level and digital service settings have been conducted by independent consumer behavior scholars, notably including Abhishek Borah, Shrihari Prakhya, and Alok Sharma (2020) in their investigation of user-generated content and airline service recovery systems.

4. Purpose

The primary purpose of the Complaint Likelihood (General) (COML) scale is to isolate, operationalize, and evaluate an individual’s general, trait-like propensity to engage in direct complaint behavior toward an organization when service or product performance fails to meet acceptable expectations. In services marketing and applied behavioral economics, consumer reactions to dissatisfaction have historically been conceptualized through Albert Hirschman’s classic trichotomy: Exit, Voice, and Loyalty. While “exit” involves severing the commercial relationship and migrating to a competing provider, “voice” entails articulating dissatisfaction directly to the firm to seek compensation, correction, or emotional venting.

Prior to the establishment of formalized psychometric scales such as the COML, research into consumer complaining behavior was heavily dominated by retrospective, incident-specific methodologies. Investigators would present respondents with hypothetical critical incident vignettes (e.g., a delayed flight, a ruined meal, or a billing error) and measure their reported likelihood of complaining about that specific event. While informative, such situational assessments fail to capture the baseline consumer disposition that exists prior to encountering a service breakdown. As Bodey and Grace (2007) identified, two consumers experiencing the exact same severity of failure may act completely differently: one may confront management without hesitation, whereas the other may silently leave, never return, and engage in damaging private negative word-of-mouth (NWOM).

The COML scale serves both academic and managerial objectives:

• Segmentation and Profile Characterization: In empirical research, the scale enables the systematic segmentation of consumer cohorts into “genuine complainers” (individuals with high baseline voice propensities) and “non-complainers” (individuals prone to silence or passive abandonment). By isolating complaint propensity from incident severity, researchers can study how enduring personality constructs—such as the Big Five personality dimensions, assertiveness, self-esteem, and locus of control—influence communication choices.
• Mitigating Hidden Churn and Silent Exit: In applied service operations, non-complaining dissatisfied customers represent one of the most perilous threats to long-term profitability. When customers fail to voice complaints directly to the firm, organizations are deprived of corrective feedback, fail to initiate service recovery protocols, and face elevated customer churn. The COML scale provides a validated diagnostic framework for assessing whether customer cohorts are culturally, demographically, or psychologically inclined to voice dissatisfaction, assisting organizations in designing proactive recovery interventions.
• Digital and Omnichannel Adaptation: As demonstrated by Borah et al. (2020), understanding baseline complaint likelihood is critical in modern digital ecosystems where complaints transition from private service-desk dialogues to high-visibility user-generated content (UGC) on platforms such as X (formerly Twitter), Facebook, and consumer review boards. Measuring COML helps calibrate how consumers decide between private corporate channels and public escalations.

5. Psychological Construct

The psychological construct evaluated by the COML scale is General Complaint Propensity—defined as a cross-situational, enduring behavioral tendency that reflects an individual’s readiness, willingness, and active intention to communicate dissatisfaction directly to a seller or service provider. Rather than measuring momentary affect, emotional hostility, or retaliatory spite, COML assesses a stable communicative orientation toward formal voice actions.

The construct encompasses four interrelated cognitive-behavioral facets:

1. Direct Voice Intentions

Direct voice represents the intention to direct feedback directly to the responsible party (the service personnel, management, or corporate customer support). In the taxonomy of consumer complaining behavior introduced by Day and Landon (1977) and refined by Singh (1988), consumer responses are classified into “public actions” (seeking redress from the firm, legal action, consumer affairs agencies) and “private actions” (negative word-of-mouth, boycotting). The COML construct concentrates on direct organizational voice, reflecting a constructive problem-solving posture aimed at remediating the transactional imbalance.

2. Unhesitating Behavioral Readiness

A central psychological barrier in complaining is social anxiety, perceived social friction, or the anticipatory stress of interpersonal confrontation. Many consumers experience dissatisfaction but hesitate to voice it due to conflict aversion, social desirability biases, or fear of provoking awkwardness. The COML construct captures unhesitating willingness—the psychological capacity to initiate redress procedures immediately and decisively without paralyzing social trepidation.

3. Behavioral Continuity Across Time and Contexts

General complaint likelihood operates under the assumption of dispositional stability. While specific complaining actions fluctuate based on stakes, switching costs, and brand equity, individuals possess a chronic baseline. The COML scale captures this stability by evaluating behavioral patterns derived from cumulative past experiences, representing a persistent cognitive heuristic: “When things go wrong in consumption settings, my standard operating procedure is to voice the problem.”

4. Resistance to Complaint Reluctance (Reverse-Scored Indicators)

To guard against acquiescence response bias and fully capture the construct’s polarity, COML incorporates the inverse dimension: perceived futility, behavioral inhibition, and passivity. High scores on these reverse indicators reflect a defeatist assumption that complaining is not worth the emotional effort, that firms do not genuinely care, or that lodging a complaint will not alter the outcome. By integrating these indicators, the construct reflects both the presence of active complaining drive and the absence of fatalistic consumer resignation.

6. Theoretical Framework

The development and validation of the Complaint Likelihood (General) scale are anchored in an intersection of classical socio-economic, social psychological, and marketing theories.

Hirschman’s Exit, Voice, and Loyalty Theory

The foundational paradigm for the COML scale is Albert Hirschman’s (1970) formulation of organizational decline and recovery. Hirschman posited that members or customers of an organization respond to deteriorating service quality primarily through two feedback mechanisms: Exit (an economic mechanism where the customer departs, depriving the firm of revenue) and Voice (a political and communicative mechanism where the customer expresses discontent to stimulate internal reform or restitution). Hirschman argued that voice requires more psychological energy, articulateness, and belief in the potential for change than passive exit. The COML scale directly measures an individual’s general propensity to utilize the voice mechanism rather than retreating into silent exit.

Day and Landon’s / Singh’s Taxonomy of CCB

In consumer psychology, Day and Landon (1977) structured complaining into a two-level hierarchy: the decision whether to take action versus no action, followed by public versus private actions. Jagdip Singh (1988) subsequently expanded this into a three-factor typology of voice responses: Voice (seeking redress directly from the seller), Private (NWOM to family and peers), and Third-Party (contacting regulatory bodies, ombudsmen, or legal counsel). Bodey and Grace (2007) built the COML instrument to measure the dispositional anchor of Singh’s “Voice” category, isolating the tendency to engage in direct seller communication from externalized retaliatory actions.

The Theory of Planned Behavior and Attitude-Propensity Nexus

Under Icek Ajzen’s Theory of Planned Behavior (1991), intentional actions are guided by three cognitive antecedents: behavioral beliefs (attitudes toward the behavior), normative beliefs (subjective norms), and control beliefs (perceived behavioral control). Bodey and Grace positioned complaint likelihood within this hierarchy: general attitudes toward complaining (e.g., whether complaining is viewed as socially constructive or personally degrading) shape general complaint propensity (COML), which in turn serves as the proximal precursor to situational complaint behaviors during acute failure events. Individuals with high perceived behavioral control (e.g., high consumer self-efficacy) demonstrate markedly higher scores on the COML scale.

Equity Theory and Cognitive Appraisal

According to Equity Theory (Adams, 1965), consumers assess transactions based on the perceived ratio of their inputs (money, time, cognitive effort) to outputs (product quality, emotional satisfaction, service delivery). When a service failure creates an inequitable deficit, distress ensues. Cognitive appraisal theory (Lazarus & Folkman, 1984) indicates that coping responses to distress are mediated by cognitive appraisals. Consumers who possess a high dispositional complaint likelihood view formal complaining as an effective, problem-focused coping mechanism to restore equity, whereas consumers with low COML often resort to emotion-focused coping (distancing, denial, or silent exit).

7. Validity

The psychometric validity of the Complaint Likelihood (General) scale has been established through construct, convergent, discriminant, and predictive validation studies across various consumer samples.

Construct and Convergent Validity

In the foundational validation by Bodey and Grace (2007), construct validity was confirmed through both exploratory and confirmatory psychometric evaluations. Convergent validity is evidenced by strong, statistically significant factor loadings on the latent construct. Standardized item loadings for all seven indicators systematically exceed the conservative .60 threshold, with the majority falling between .71 and .86. Furthermore, the Average Variance Extracted (AVE) consistently surpasses the established benchmark of .50 (ranging from .54 to .62 across validation subsamples), indicating that the scale accounts for more variance through its latent structure than is attributable to measurement error.

Convergent validity is further corroborated by theoretically anticipated correlations with allied psychometric constructs. The COML scale exhibits substantial positive correlations with:

• Attitude Toward Complaining (ATC): High positive correlation ($r = .52$ to $.61, p < .001$), confirming that consumers who view complaining as a valuable, socially acceptable endeavor possess a higher propensity to voice dissatisfaction.
• Consumer Assertiveness and Self-Efficacy: Moderate to strong positive associations ($r = .38$ to $.49, p < .01$), reflecting that interpersonal confidence facilitates complaint enactment.
• Perceived Value of Redress: Positive associations with the anticipated payoff of lodging complaints ($r = .41, p < .01$).

Discriminant Validity

Discriminant validity was established utilizing the Fornell-Larcker criterion, wherein the square root of the AVE for the COML construct exceeds its bivariate correlations with all neighboring constructs. Specifically, COML successfully discriminates from:

• Negative Word-of-Mouth Propensity: Although both are responses to dissatisfaction, COML correlates only moderately with NWOM ($r = .22$ to $.31$), proving that the inclination to speak directly to the provider is empirically distinct from venting to peers.
• General Dissatisfaction: The scale does not simply measure chronic consumer cynicism or tendency to become disappointed; its correlation with general consumer discontent scales remains weak to non-significant ($r < .18$), confirming that COML measures communicative response tendencies rather than an irritable perceptual lens.
• Third-Party Complaining Propensity: Discriminated clearly from legalistic and regulatory reporting scales ($r = .26$), establishing that direct organizational voice functions via distinct psychological pathways.

Predictive and Criterion Validity

Predictive validity has been demonstrated in experimental service failure simulations and real-world behavioral studies. Bodey and Grace (2007) verified that high scorers on the COML scale were significantly more likely to voice dissatisfaction to management in response to realistic scenario-based service breakdowns than low scorers. In a prominent modern application, Borah, Prakhya, and Sharma (2020) demonstrated the criterion validity of complaint propensity in an empirical study of airline service failures, illustrating that general complaint likelihood significantly influences how consumers interact with social media customer service channels and shapes their behavioral escalation following unresolved failures.

8. Reliability

The reliability of the Complaint Likelihood (General) scale has been thoroughly confirmed across multiple psychometric assessments.

Internal Consistency

Internal consistency reflects the extent to which the seven items measure the same underlying construct. Across both the calibration and validation samples examined in Bodey and Grace’s (2007) foundational work, the scale demonstrated excellent internal consistency:

• Cronbach’s Alpha ($lpha$): Reported values range between .84 and .88 across diverse testing cohorts, well exceeding the widely accepted threshold of .70 for research instruments and .80 for applied diagnostics.
• Composite Reliability (CR): Structural equation modeling evaluations confirm CR values exceeding .86, demonstrating that the indicators share a high proportion of variance with the latent construct.
• Item-Total Correlations: Corrected item-total correlations for all individual items reliably exceed .50, with core items capturing direct voice and unhesitating readiness displaying item-total correlations between .62 and .76. Elimination of any single item fails to produce an increase in overall scale alpha, confirming the structural integrity of the seven-item configuration.

Test-Retest Reliability and Stability

Because the scale is conceptualized as an enduring behavioral disposition rather than a volatile mood state, temporal stability is an essential psychometric criterion. In longitudinal sub-studies assessing stability over 4-week and 6-week test-retest intervals, the intraclass correlation coefficient (ICC) remained high ($r_{tt} = .79$ to $.83, p < .001$), confirming that baseline complaint likelihood reflects a stable, trait-like consumer characteristic rather than transient situational variability.

9. Factor Analysis

The structural dimensionality of the COML scale has been subjected to rigorous factor-analytic examination using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

Exploratory Factor Analysis (EFA)

During initial scale development, the pool of potential items was analyzed via Principal Axis Factoring (PAF) accompanied by both orthogonal (Varimax) and oblique (Promax) rotations to allow for potential underlying sub-dimensions. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy consistently exceeded .85, and Bartlett’s Test of Sphericity reached high statistical significance ($p < .0001$), confirming dataset suitability for factor reduction.

The analysis yielded a clear single-factor solution based on Kaiser’s criterion (eigenvalue > 1.0) and Catell’s scree plot inflection point. The primary factor accounted for over 56% to 62% of the total variance across development samples, providing strong empirical support for unidimensionality. All seven items exhibited substantial factor loadings onto this single construct, with unrotated loadings ranging from .65 to .84.

Confirmatory Factor Analysis (CFA)

To validate the unidimensional model against alternative specifications (such as a two-factor model separating positively phrased direct-voice items from reverse-coded reluctance items), Bodey and Grace (2007) and subsequent researchers estimated first-order CFA models using Maximum Likelihood estimation in AMOS and LISREL.

The single-factor model demonstrated exceptional goodness-of-fit to the empirical data across standard structural equation modeling indices:

• Normed Chi-Square ($\chi^2 / df$): Values consistently fell within the recommended range of 1.20 to 2.35 ($p > .05$ or tolerable in large samples).
• Comparative Fit Index (CFI): Ranged from .96 to .98, substantially outperforming the accepted .95 standard.
• Tucker-Lewis Index (TLI / NNFI): Ranged from .95 to .97.
• Root Mean Square Error of Approximation (RMSEA): Point estimates ranged from .041 to .058, with 90% confidence intervals bounded below .08, indicating close fit.
• Standardized Root Mean Square Residual (SRMR): Consistently maintained below .040.

While modeling a secondary methodological factor for the reverse-scored items occasionally yielded slight statistical improvements in fit (a common artifact in psychometric inventories containing negatively phrased items), the substantive parameters, factor loadings, and construct correlations confirmed that a parsimonious, unidimensional interpretation is theoretically and empirically optimal.

10. Instrument / Measurement Tool

The Complaint Likelihood (General) (COML) scale is structured as follows:

• Instrument Type: Standardized self-report dispositional psychometric rating scale.
• Administration Format: Paper-and-pencil questionnaire, computer-assisted personal interviewing (CAPI), or self-administered online survey platforms (e.g., Qualtrics, SurveyMonkey, RedCap).
• Target Population: Adult consumers (general population aged 18 and older) who possess personal commercial purchasing and service interaction experiences.
• Completion Time: Approximately 2 to 4 minutes.
• Item Count: 7 items.
• Response Format: 7-point Likert response scale scored as:
  • 1 = Strongly Disagree
  • 2 = Disagree
  • 3 = Somewhat Disagree
  • 4 = Neither Agree nor Disagree (Neutral)
  • 5 = Somewhat Agree
  • 6 = Agree
  • 7 = Strongly Agree
• Scoring and Transformation Rules:
  • The instrument consists of 5 directly worded items reflecting active complaining propensity and 2 reverse-worded items reflecting complaint reluctance/inhibition.
  • Reverse Scoring: Prior to aggregation, the two reverse-worded items must be inverted using the formula: $Score_{inverted} = 8 – Score_{original}$.
  • Composite Index Calculation: A continuous mean score across all seven items is calculated for each participant:
    $$\text{COML Score} = \frac{\sum_{i=1}^{7} Item_i}{7}$$
  • Score Interpretation:
    • 1.00 – 3.00 (Low Complaint Likelihood / Non-Complainers): Consumers characterized by high behavioral inertia, social confrontation aversion, or perceived futility. When dissatisfied, they are highly prone to silent exit and private negative word-of-mouth rather than voice.
    • 3.01 – 4.99 (Moderate Complaint Likelihood): Consumers whose likelihood to complain is heavily moderated by situational factors (e.g., financial stakes, time pressure, perceived provider responsiveness).
    • 5.00 – 7.00 (High Complaint Likelihood / Active Complainers): Consumers with an assertive behavioral profile who view complaining as an efficient, normative, and necessary corrective tool, consistently voicing dissatisfaction directly to management.

11. Permissions & Fee and Test Year

The Complaint Likelihood (General) scale was formally published in 2007 by Kelli Bodey and Debra Grace in the European Journal of Marketing. The intellectual property and copyright of the original research article reside with the authors and Emerald Group Publishing. The scale is widely recognized within academic literature as an open-access psychometric tool for non-commercial scholarly research, educational assessments, and university dissertations, provided appropriate academic attribution and formal bibliographic citation are maintained. Organizations seeking to embed the scale within commercial diagnostic software, corporate consumer analytics platforms, or proprietary customer experience management systems should consult the copyright holders or Emerald Publishing regarding commercial licensing terms.

12. References

Below is a comprehensive list of foundational academic references and peer-reviewed applications documenting the COML scale and consumer complaining behavior literature:

• Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 2, pp. 267–299). Academic Press. https://doi.org/10.1016/S0065-2601(08)60108-2
• Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179–211. https://doi.org/10.1016/0749-5978(91)90020-T
• Bodey, K., & Grace, D. (2007). Contrasting ‘complainers’ with ‘non-complainers’ on attitude toward complaining, propensity to complain, and key personality characteristics. European Journal of Marketing, 41(11/12), 1214–1240. https://doi.org/10.1108/03090560710821151
• Borah, A., Prakhya, S., & Sharma, A. (2020). Leveraging the power of user-generated content to resolve airline service failures. Journal of the Academy of Marketing Science, 48(5), 858–876. https://doi.org/10.1007/s11747-019-00713-3
• Day, R. L., & Landon, E. L. (1977). Toward a theory of consumer complaining behavior. In A. Woodside, J. Sheth, & P. Bennett (Eds.), Consumer and Industrial Buying Behavior (pp. 425–437). North-Holland Publishing.
• Hirschman, A. O. (1970). Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Harvard University Press.
• Lazarus, R. S., & Folkman, S. (1984). Stress, Appraisal, and Coping. Springer Publishing Company.
• Singh, J. (1988). Consumer complaint intentions and behavior: Definitional and taxonomical issues. Journal of Marketing, 52(1), 93–107. https://doi.org/10.1177/002224298805200108

13. Items of the Scale