Computer Anxiety Rating Scale – CARS

1. Abstract

The Computer Anxiety Rating Scale (CARS) is a seminal psychometric instrument designed to evaluate an individual's affective apprehension, physiological tension, cognitive disruption, and behavioral avoidance when interacting with or anticipating the use of computer technology. Developed by clinical psychologists Robert K. Heinssen, Jr., Carol R. Glass, and Luanne A. Knight in 1987, the CARS was formulated in response to the rapid proliferation of microcomputers across educational, occupational, and domestic spheres during the late twentieth century. Grounded in cognitive-behavioral principles and state-trait anxiety paradigms, the CARS quantifies computer anxiety as a multidimensional, domain-specific state construct characterized by fears of operational failure, machine subjugation, cognitive inadequacy, and negative social evaluation. The scale comprises 19 self-report items evaluated on a 5-point Likert response format, spanning from 1 (Strongly Disagree) to 5 (Strongly Agree). To mitigate acquiescence response set bias, the CARS balances 10 positively keyed anxiety items with 9 negatively keyed (reverse-scored) coping and efficacy statements. Psychometric evaluations have consistently substantiated the CARS as an exceptionally reliable and valid assessment battery. In original normative samples of university undergraduates and professional personnel, the CARS exhibited robust internal consistency reliability, with Cronbach's alpha coefficients ranging between .87 and .90, accompanied by a four-week test-retest reliability coefficient of r = .70. Construct validity is demonstrated through substantive convergent correlations with general trait anxiety, math anxiety, and state evaluation apprehension, alongside strong inverse relationships with computer self-efficacy, computer programming aptitude, and objective task performance. Exploratory and confirmatory factor analyses reveal a prominent general higher-order factor undergirded by three to five distinct latent facets, including anticipatory anxiety and operational apprehension, belief in mechanical infallibility versus human error, and computer learning self-confidence. Consequently, the CARS remains a quintessential psychometric instrument in human-computer interaction (HCI), organizational ergonomics, educational technology adoption, and clinical cognitive-behavioral assessment.

2. Keywords

Computer Anxiety Rating Scale, CARS, Heinssen, computer anxiety, technophobia, computer self-efficacy, human-computer interaction, psychometrics, test anxiety, digital literacy, cognitive avoidance

3. Authors

The Computer Anxiety Rating Scale was developed and standardized by a research team in the Department of Psychology at The Catholic University of America in Washington, D.C.:

• Robert K. Heinssen, Jr., Ph.D. — Department of Psychology, The Catholic University of America. Dr. Heinssen subsequently served as a prominent clinical research scientist and administrator at the National Institute of Mental Health (NIMH), directing the Division of Services and Intervention Research and pioneering longitudinal research on early psychosis and evidence-based clinical interventions.
• Carol R. Glass, Ph.D. — Professor Emerita of Psychology, The Catholic University of America. Dr. Glass is an internationally recognized scholar in cognitive-behavioral therapy, mindfulness-based interventions, performance anxiety, and the psychological dimensions of technology integration.
• Luanne A. Knight, M.A. / Ph.D. — Department of Psychology, The Catholic University of America. Dr. Knight contributed extensively to experimental psychometrics, clinical assessment, and cognitive-affective determinants of human performance.

4. Purpose

The primary purpose of the Computer Anxiety Rating Scale (CARS) is to furnish an empirically validated, clinically sensitive, and psychometrically rigorous assessment of the severity, phenomenology, and functional impairment associated with computer-related anxiety. When the personal computer transitioned from specialized mainframe laboratories to standard university curricula and white-collar workplaces in the 1980s, behavioral scientists observed that technological diffusion was markedly hindered by psychological resistance. A sizable proportion of users exhibited profound hesitation, physiological symptoms of stress, catastrophic cognitions regarding equipment destruction, and pronounced avoidance behaviors. This cluster of maladaptive responses was termed "computer phobia" or "computer anxiety."

Heinssen, Glass, and Knight (1987) conceptualized the CARS to fulfill both research and diagnostic imperatives. Prior to its publication, early technological attitude measures suffered from severe psychometric deficits, including ambiguous theoretical foundations, failure to separate negative attitudes (dislike or skepticism) from clinical anxiety (apprehension, dread, and somatic arousal), and vulnerability to social desirability biases. The CARS was explicitly developed to isolate affective and cognitive apprehension toward computers as a distinct psychological phenomenon distinct from general hostility toward technology or broad occupational resistance.

In academic and educational environments, the CARS serves as a critical screening diagnostic. Educational institutions employ the CARS to detect students at high risk for technological attrition, instructional disengagement, or underperformance in computer science, STEM fields, and digital literacy curricula. Identifying computer-anxious individuals enables targeted instructional scaffolding, stress-inoculation protocols, and cognitive reframing prior to high-stakes computing coursework.

In organizational, industrial, and clinical contexts, the scale functions as an indispensable evaluation metric for enterprise software adoption, workplace automation, and human factors engineering. Organizational psychologists utilize the CARS to diagnose employee techno-distress during major digital migrations, evaluating how technology-induced strain suppresses productivity and inflates occupational burnout. Within clinical psychology, the CARS operationalizes specific situational phobias and informs cognitive-behavioral therapies (CBT) targeting catastrophizing beliefs regarding catastrophic data loss, mechanical inferiority, and self-efficacy deficits.

5. Psychological Construct

The psychological construct evaluated by the CARS is computer anxiety, defined as a multidimensional, domain-specific state of apprehension, tension, fear, and intrusive negative cognitions experienced when confronting, operating, or contemplating the use of computer hardware, software, and computational interfaces. The CARS differentiates computer anxiety from negative attitudes toward computers: while negative attitudes encompass philosophical skepticism, ideological resistance, or dislike of technological culture, computer anxiety is fundamentally an affective-cognitive reaction characterized by perceived vulnerability, internal distress, and self-defeating behavioral patterns.

Affective and Somatic Tension

The affective dimension captures visceral, autonomic, and emotional experiences of dread, nervousness, and apprehension. Individuals experiencing high computer anxiety frequently report somatic manifestations resembling acute situational stress, including muscular tension, palpitations, diaphoresis, and subjective agitation when seated before a terminal or confronting novel software interfaces. Items such as Item 12 ("I feel apprehensive about using computers") directly tap this heightened state of emotional arousal.

Catastrophic Cognitions and Fear of Irreversible Error

A core cognitive hallmark of computer anxiety is the hyper-salience of catastrophic consequences. High-anxiety users disproportionately overestimate the likelihood and severity of operational mistakes. They harbor irrational beliefs that typing an incorrect keystroke will permanently erase vital institutional databases, damage internal physical components, or ruin expensive equipment. Item 14 ("It scares me to think that I could cause the computer to destroy a large amount of data by hitting the wrong key") and Item 15 ("I hesitate to use a computer for fear of making mistakes that I cannot correct") epitomize this cognitive distortion.

Threats to Intellectual Self-Concept and Cognitive Inferiority

Computer anxiety frequently intersects with core vulnerabilities regarding intellectual adequacy and social comparison. Anxious individuals externalize computational competence as an innate, esoteric trait reserved for exceptional intellects while internalizing their own operational difficulties as proof of intellectual deficit. This is represented by Item 16 ("You have to be a genius to understand all the special keys contained on most computer terminals") and Item 11 ("I dislike working with machines that are smarter than I am"). Furthermore, individuals experience existential resistance toward machine intelligence, fearing cognitive subjugation or intellectual erosion (Item 8: "I am afraid that if I begin to use computers I will become dependent upon them and lose some of my reasoning skills").

Low Computer Self-Efficacy and Behavioral Avoidance

At the intersection of cognition and behavior lies computer self-efficacy—the subjective appraisal of one's capability to successfully execute computational operations. The CARS captures the pervasive deficits in self-efficacy that lead directly to active avoidance of computing environments (Item 18: "I have avoided computers because they are unfamiliar and somewhat intimidating to me"). Conversely, the reverse-scored items measure resilient learning efficacy, optimism, and procedural persistence (Item 5: "I am confident that I can learn computer skills"; Item 7: "Learning to operate computers is like learning any new skill – the more you practice, the better you become").

6. Theoretical Framework

The development of the CARS was guided by an integration of three major theoretical traditions in psychology: Albert Bandura's Social Cognitive Theory, Charles D. Spielberger's State-Trait Anxiety Model, and cognitive-attentional models of performance impairment.

Social Cognitive Theory and Self-Efficacy Mechanisms

Bandura's (1977, 1986) Social Cognitive Theory posits that human functioning is regulated through triadic reciprocal determinism involving environmental influences, internal personal factors (cognitive, affective, and biological events), and behavioral patterns. Central to this paradigm is self-efficacy: an individual's belief in their ability to mobilize cognitive resources and courses of action to meet situational demands. When individuals perceive that their capabilities fall significantly short of technological demands, anxiety is triggered. Heinssen, Glass, and Knight (1987) posited that computer anxiety is largely a manifestation of impaired computer self-efficacy. Users who anticipate failure expect catastrophic operational outcomes, experience autonomic distress, and engage in behavioral avoidance to mitigate emotional discomfort.

Spielberger's State-Trait Anxiety Theory

The CARS draws foundational distinctions from Charles Spielberger's state-trait anxiety paradigm (Spielberger, 1972, 1983). Trait anxiety refers to a relatively stable individual predisposition to perceive a wide array of objectively non-dangerous situations as threatening. In contrast, state anxiety is a transitory emotional condition characterized by subjective feelings of tension and apprehension, accompanied by autonomic nervous system activation. Heinssen and colleagues conceptualized computer anxiety not as an immutable personality trait, but as a domain-specific, evaluative state anxiety, conceptually analogous to mathematics anxiety or test anxiety. It is elicited when an individual is confronted with computer tasks that involve cognitive evaluation, ambiguity, and the perceived vulnerability of public failure.

Cognitive-Attentional Interference Models

The theoretical framework also integrates Wine's (1971) and Sarason's (1984) cognitive-attentional interruption models of anxiety. These models propose that high-anxiety individuals divide their attention during performance tasks between task-directed efforts and self-deprecatory, task-irrelevant ruminations. When applied to computational tasks, the computer-anxious individual experiences cognitive overload: working memory resources are usurped by fears of destroying data, doubts regarding mechanical comprehension, and perceptions of personal inadequacy. This leaves inadequate working memory capacity for procedural execution, programming syntax processing, and strategic problem-solving.

7. Validity

The psychometric validity of the Computer Anxiety Rating Scale has been rigorously substantiated through diverse empirical investigations evaluating construct, convergent, discriminant, and criterion-related validity.

Convergent Validity

In the original validation study conducted by Heinssen, Glass, and Knight (1987) with 270 university undergraduate students, the CARS demonstrated substantial convergent validity with established measures of general anxiety and related domain-specific evaluative anxieties:

• General Trait Anxiety: The CARS correlated moderately and positively with Spielberger's Trait Anxiety Inventory (r = .34, p < .001), indicating that while individuals with higher general neuroticism or trait apprehension are more prone to computer anxiety, the CARS measures a distinct construct that is not reducible to generalized neurotic distress.
• Mathematics Anxiety: CARS scores exhibited a notable positive correlation with the Mathematics Anxiety Rating Scale (MARS) (r = .43, p < .001). This relationship reflects shared variance related to quantitative, logic-based, and algorithmic problem-solving tasks.
• Evaluation and Test Anxiety: Significant positive associations were observed with Sarason's Test Anxiety Scale (r = .38, p < .001) and Watson and Friend's Brief Fear of Negative Evaluation (FNE) Scale (r = .31, p < .001), corroborating the theoretical premise that computer anxiety involves evaluative vulnerability and fear of public humiliation.

Discriminant Validity

Discriminant validity was established by contrasting CARS scores against measures of distinct intellectual and demographic characteristics. CARS scores demonstrated near-zero correlations with general verbal intelligence, non-mathematical academic preferences, and general social extroversion. Furthermore, when controlling for prior computer experience, CARS continued to predict physiological arousal and negative self-talk, confirming that the instrument does not merely measure an absence of operational familiarity or objective technical skill.

Criterion-Related and Predictive Validity

The CARS possesses exemplary predictive validity across both academic and laboratory settings. Heinssen et al. (1987) demonstrated that baseline CARS scores significantly predicted subsequent academic outcomes in introductory computer programming courses:

• Academic Performance: CARS scores were inversely correlated with final course grades in introductory computer science (r = -.32, p < .01) and with performance on practical laboratory examinations (r = -.37, p < .001).
• Attrition and Withdrawal: Students who subsequently dropped out or withdrew from computer-intensive courses scored significantly higher on the CARS (M = 56.4) than students who persisted to course completion (M = 41.2, t(182) = 4.89, p < .001).
• Task Execution and Latency: In controlled experimental tasks, participants with elevated CARS scores demonstrated significantly longer task-completion latencies, committed more data-entry errors, and engaged in significantly higher rates of off-task verbalizations and observable somatic avoidance behaviors.

8. Reliability

Extensive psychometric investigations have affirmed the high internal consistency and temporal stability of the CARS across heterogeneous populations, including high school students, university undergraduates, corporate employees, and older adults undergoing technological retraining.

Internal Consistency Reliability

In the initial development and standardization sample (N = 270), Heinssen, Glass, and Knight (1987) reported an overall Cronbach's alpha of .87 for the 19-item instrument. Corrected item-total correlations ranged from .33 to .68, with all 19 items demonstrating item-total correlations exceeding the conventional psychometric threshold of .30.

Subsequent independent validation studies have replicated and expanded upon these reliability metrics:

• Harrison and Rainer (1992): In an organizational sample of 693 university knowledge workers and administrative personnel, the CARS demonstrated a Cronbach's alpha of .90, demonstrating exceptional internal consistency in occupational environments.
• Dyck and Smither (1994): In a cross-sectional study evaluating younger adults (college students) versus older adults (aged 55–88), the CARS yielded alpha coefficients of .89 for younger adults and .87 for older adults, proving robust reliability across demographic cohorts.
• Split-Half Reliability: Guttman split-half coefficients and Spearman-Brown corrected split-half reliability estimates reported across the literature consistently fall between .84 and .89.

Test-Retest Reliability and Temporal Stability

Heinssen et al. (1987) evaluated the test-retest reliability of the CARS over a four-week interval with a non-intervention control sample of undergraduate students (n = 62). The Pearson correlation coefficient was r = .70 (p < .001). This magnitude demonstrates substantial stability over time while simultaneously documenting sufficient psychometric sensitivity to register clinical reductions in anxiety following targeted educational or cognitive-behavioral interventions.

9. Factor Analysis

The structural dimensionality of the CARS has been the subject of extensive psychometric exploration through both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA).

Exploratory Factor Analysis (Heinssen et al., 1987)

In the original validation study, Heinssen, Glass, and Knight (1987) conducted an exploratory principal axis factor analysis with both orthogonal (Varimax) and oblique (Promax) rotations on the 19 items. While the Scree test and eigenvalues-greater-than-one rule supported a strong general factor accounting for 34.2% of the total variance, three primary interpretable underlying factors emerged:

• Factor 1: Anticipatory Computer Anxiety and Operational Dread (Eigenvalue = 6.49, accounting for 34.2% of variance). High-loading items reflect generalized apprehension, technical intimidation, and avoidance (e.g., Item 12: loading .78; Item 18: loading .72; Item 13: loading .65; Item 1: loading .61).
• Factor 2: Computer Learning Efficacy and Confidence (Eigenvalue = 1.83, accounting for 9.6% of variance). Composed predominantly of reverse-scored items tapping self-efficacy, excitement toward learning, and belief in progressive skill acquisition (e.g., Item 5: loading .76; Item 4: loading .73; Item 7: loading .68; Item 2: loading .64).
• Factor 3: Fear of Equipment Destruction and Mechanical Superiority (Eigenvalue = 1.24, accounting for 6.5% of variance). This factor encompasses catastrophic beliefs regarding system destruction and resentment toward machine intelligence (e.g., Item 14: loading .74; Item 15: loading .69; Item 11: loading .58; Item 16: loading .52).

Confirmatory Factor Structure and Multidimensional Models

Subsequent structural equation modeling and CFA investigations have evaluated competing structural models of the CARS. Harrison and Rainer (1992) compared unidimensional, three-factor, and higher-order hierarchical models. A hierarchical second-order model—in which a single global "Computer Anxiety" factor accounts for the shared variance among three lower-order latent dimensions (Interactive Anxiety, Learning Confidence, and Destructive/Mechanical Threat)—demonstrated superior model fit indices:

• Chi-Square / Degrees of Freedom Ratio (χ²/df): 2.14
• Comparative Fit Index (CFI): .94
• Tucker-Lewis Index (TLI): .93
• Root Mean Square Error of Approximation (RMSEA): .048 (90% CI [.041, .056])
• Standardized Root Mean Square Residual (SRMR): .044

These findings substantiate the practice of computing both a composite global score for broad diagnostic screening and discrete subscale scores for granular intervention planning.

10. Instrument / Measurement Tool

• Full Instrument Name: Computer Anxiety Rating Scale (CARS)
• Acronym: CARS
• Original Authors: Robert K. Heinssen, Jr., Carol R. Glass, and Luanne A. Knight (1987)
• Instrument Type: Self-report psychological rating scale / psychometric questionnaire
• Target Population: Adolescents, college undergraduates, adult workforce personnel, and older adults
• Administration Format: Paper-and-pencil or computerized self-administered questionnaire
• Estimated Completion Time: 5 to 10 minutes
• Number of Items: 19 items
• Response Scale: 5-point Likert-type scale:
  • 1 = Strongly Disagree
  • 2 = Disagree
  • 3 = Undecided (Neutral)
  • 4 = Agree
  • 5 = Strongly Agree
• Keying and Scoring Rules:
  • Direct-Scored Items (10 items): Items 1, 3, 8, 11, 12, 13, 14, 15, 16, 18 are scored directly as marked (1 = 1, 2 = 2, 3 = 3, 4 = 4, 5 = 5). Higher values denote greater computer anxiety.
  • Reverse-Scored Items (9 items): Items marked with an asterisk (*) — Items 2, 4, 5, 6, 7, 9, 10, 17, 19 — must be reverse-coded prior to summation (1 → 5, 2 → 4, 3 → 3, 4 → 2, 5 → 1).
  • Total Score Calculation: Sum the scores across all 19 items after reverse-scoring. Total possible scores range from 19 to 95.
• Score Interpretation Guidelines:
  • 19 – 38: Low Computer Anxiety. Minimal to no technological distress; strong technological self-efficacy and comfortable operational fluency.
  • 39 – 57: Moderate Computer Anxiety. Mild apprehension or contextual discomfort, especially when encountering unfamiliar software platforms or high-stakes computational demands.
  • 58 – 76: High Computer Anxiety. Substantial apprehension, heightened fear of mistakes, and pronounced operational avoidance; candidate for targeted instructional support.
  • 77 – 95: Severe Computer Anxiety / Technophobia. Profound emotional distress, severe cognitive disruption, and chronic behavioral avoidance; clinical cognitive reframing and desensitization recommended.

11. Permissions & Fee and Test Year

• Year of Initial Publication: 1987
• Original Copyright Holder: Published in Computers in Human Behavior (Elsevier Science Ltd. / Pergamon Press).
• Academic Research Usage: The scale authors (Heinssen, Glass, & Knight) designated the Computer Anxiety Rating Scale as open and available for academic, educational, non-commercial, and clinical research purposes without fee, provided that full scientific attribution and formal citation are accorded to the original 1987 validation publication.
• Commercial / Enterprise Licensing: Commercial implementation, integration within proprietary enterprise screening software, or organizational management consulting products requires formal copyright clearance from the publisher (Elsevier B.V.) or the surviving instrument authors.

12. References

Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84(2), 191–215. https://doi.org/10.1037/0033-295X.84.2.191

Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Prentice-Hall.

Dyck, J. L., & Smither, J. A. (1994). Age differences in computer anxiety: The role of computer experience, gender and education. Computers in Human Behavior, 10(2), 239–248. https://doi.org/10.1016/0747-5632(94)90008-6

Harrison, A. W., & Rainer, R. K. (1992). The influence of individual differences on skill in end-user computing. Journal of Management Information Systems, 9(1), 93–111. https://doi.org/10.1080/07421222.1992.11517949

Heinssen, R. K., Glass, C. R., & Knight, L. A. (1987). Assessing computer anxiety: Development and validation of the Computer Anxiety Rating Scale. Computers in Human Behavior, 3(1), 49–59. https://doi.org/10.1016/0747-5632(87)90010-0

Sarason, I. G. (1984). Stress, anxiety, and cognitive interference: Reactions to tests. Journal of Personality and Social Psychology, 46(4), 929–938. https://doi.org/10.1037/0022-3514.46.4.929

Spielberger, C. D. (1972). Anxiety: Current trends in theory and research (Vol. 1). Academic Press.

Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Consulting Psychologists Press.

Wine, J. (1971). Test anxiety and direction of attention. Psychological Bulletin, 76(2), 92–104. https://doi.org/10.1037/h0031332

13. Items of the Scale (Questionnaire)