Attitudes Toward Self (ATS)

Abstract

The Attitudes Toward Self (ATS) scale is a brief, theoretically anchored psychometric instrument developed by Charles S. Carver and colleagues to assess cognitive-vulnerability dimensions within a self-regulatory framework. Spanning 10 self-report items, the ATS operationalizes three distinct cognitive predispositions: the establishment of excessively high personal standards (High Standards, 3 items), the tendency toward punitive self-reproach upon perceived shortfall (Self-Criticism, 3 items), and the catastrophic tendency to extrapolate a single localized failure to an overarching evaluation of global self-worth (Generalization, 4 items). Respondents rate items along a 5-point Likert-type response format ranging from 1 (I agree a lot) to 5 (I Disagree a lot), which are typically scored such that elevated values reflect higher cognitive vulnerability. Psychometric evaluations consistently confirm the three-factor orthogonal or modestly oblique structure through exploratory and confirmatory factor analyses. Internal consistency estimates across diverse community, collegiate, and clinical samples demonstrate robust reliability, with Cronbach’s alpha coefficients averaging approximately .76 for High Standards, .78 for Self-Criticism, and .78 for Generalization. Crucially, empirical investigations underscore the unique predictive validity of the Generalization subscale: whereas high standards and self-criticism are frequently shared across non-depressive perfectionistic and achievement-oriented profiles, only generalization from discrete failure interacts prospectively with adverse life events to predict the onset and escalation of depressive symptoms. Furthermore, comparative psychopathology research demonstrates that generalization discriminates unipolar depressive diatheses from bipolar manic vulnerabilities, establishing the ATS as an indispensable instrument in affective cognitive neuroscience, clinical psychology, and personality assessment.

Keywords

Attitudes Toward Self, ATS, Charles S. Carver, cognitive vulnerability, depression, self-regulation, generalization, self-criticism, high standards, diathesis-stress model, perfectionism, psychometrics

Authors

The Attitudes Toward Self (ATS) instrument was formulated by Charles S. Carver, Ph.D. (1947–2019), a preeminent American personality psychologist and Distinguished Professor of Psychology at the University of Miami, Coral Gables, Florida, United States. Carver’s pioneering scholarship fundamentally reshaped modern social, personality, and clinical psychology through his work on self-regulatory feedback systems, dispositional optimism, coping mechanisms, and behavioral activation and inhibition systems.

The initial conceptualization and formal empirical validation of the instrument were conducted in collaboration with colleagues across multiple foundational investigations, including L. La Voie, Julius Kuhl (University of Osnabrück), and Robert J. Ganellen (Northwestern University Feinberg School of Medicine). Later psychometric refinements and clinical extensions were realized in collaboration with Sheri L. Johnson (University of California, Berkeley) and Adele M. Hayes (University of Delaware).

Purpose

The primary clinical and research purpose of the Attitudes Toward Self (ATS) scale is to isolate, differentiate, and quantify specific maladaptive cognitive patterns that function as diatheses for affective disorders, most notably unipolar major depression. While classical cognitive models of depression—such as those articulated by Aaron T. Beck—posit that negative core schemas, dysfunctional attitudes, and cognitive distortions predispose individuals to affective collapse, broad measures of perfectionism or general dysfunctional attitudes frequently conflate benign or adaptive achievement striving with genuinely depressogenic vulnerabilities.

The theoretical rationale behind the development of the ATS rests on Carver and Michael F. Scheier’s cybernetic, control-process model of human self-regulation. Within this paradigm, human behavior is understood as organized around hierarchical feedback loops wherein individuals compare their perceived present state against salient reference values (standards or goals). When a discrepancy between the current state and the goal is detected, individuals normally experience negative affect and engage in corrective action. However, when an individual processes perceived failure through pathological cognitive operations, affective dysregulation and behavioral disengagement ensue.

The ATS was purpose-built to answer a fundamental psychopathological question: Which specific self-evaluative attitudes genuinely confer vulnerability to clinical depression? Is holding elevated standards inherently depressogenic? Is self-reproach upon failing to meet those standards sufficient to cause depressive collapse? Or does depression require a qualitative cognitive leap wherein a localized performance deficit is interpreted as diagnostic of total personal inadequacy? By decomposing self-evaluative cognition into three clean, factorially pure subscales—High Standards, Self-Criticism, and Generalization—the ATS provides researchers and clinicians with a nuanced diagnostic tool.

In clinical practice, the ATS serves as an incisive case conceptualization instrument. Cognitive-behavioral therapists utilize the scale to determine whether an individual’s distress is maintained by an excessively demanding aspirational standard, chronic self-punitive emotional reactivity, or overgeneralized catastrophizing. In longitudinal psychopathology research, the ATS serves as an established diathesis measure in prospective diathesis-stress protocols, permitting researchers to assess how stable cognitive vulnerability traits moderate the impact of acute negative life stressors over weeks, months, or years.

Psychological Construct

The Attitudes Toward Self instrument maps three distinct yet interrelated psychological constructs situated within the cognitive-personality interface: High Standards, Self-Criticism, and Generalization. Carver’s work demonstrates that these constructs occupy differential functional positions in the etiology of emotional distress.

1. High Standards (Items 1, 4, 7)

The High Standards construct represents the tendency of an individual to establish and endorse exceptionally elevated, ambitious performance benchmarks for personal conduct and achievement relative to others. This dimension reflects an aspirational reference value within the self-regulatory hierarchy. An individual endorsing high standards seeks exceptional outcomes, aspires higher than peer norms, and maintains ambitious expectations for their own behavior. Crucially, Carver and colleagues established that holding high standards is not inherently pathological. In non-clinical and achievement-oriented cohorts, high standards frequently correlate positively with conscientiousness, high self-efficacy, and goal pursuit. In psychopathological models, High Standards does not independently predict depressive episodes. It represents the evaluative threshold against which performance is compared, acting merely as the initial condition that sets the stage for possible discrepancy detection.

2. Self-Criticism (Items 3, 6, 9)

The Self-Criticism construct operationalizes an acute, punitive emotional and cognitive reaction triggered when an individual perceives that their efforts have failed to meet their self-imposed standards. Whereas High Standards reflects the height of the comparative goal, Self-Criticism reflects affective hostility directed inward—manifesting as anger, frustration, and distress targeting the self following an unsatisfactory outcome. Individuals high on this dimension experience marked subjective dissatisfaction and self-directed annoyance whenever a performance shortfall occurs. Although self-criticism correlates cross-sectionally with general distress and neuroticism, empirical research reveals that self-criticism alone is often insufficient to trigger clinical depression, provided the self-criticism remains circumscribed to the specific domain or task in which the shortfall occurred.

3. Generalization (Items 2, 5, 8, 10)

The Generalization construct (also termed “generalization of failure” or “catastrophic overgeneralization”) is the primary pathological mechanism assessed by the ATS. It reflects the cognitive propensity to extrapolate a single, localized, acute performance failure or perceived flaw to one’s total self-concept, global competence, and fundamental self-worth. When an individual characterized by high generalization encounters a setback in one specific task, they do not conclude, “I failed at this particular task”; rather, they infer, “I am an utter failure across all domains of life, and I am incapable of achieving anything meaningful.” Generalization represents an erroneous upward shift in cognitive abstraction—a single behavioral failure at a low level of the self-regulatory hierarchy is interpreted as conclusive proof of defectiveness at the highest system level (the idealized self). Across dozens of empirical investigations, this construct has emerged as the sole unique prospective predictor of depressive symptoms, demonstrating that cognitive vulnerability to depression is defined not by how high one’s goals are, nor by how upset one gets when failing them, but by the catastrophic cognitive leap from a discrete mistake to a ruined self.

Theoretical Framework

The ATS is firmly grounded in cybernetic control theory and self-regulation principles synthesized by Charles S. Carver and Michael F. Scheier (1981, 1990, 1998). This theoretical framework posits that human behavior, affect, and cognition operate through continuous, hierarchical, negative (discrepancy-reducing) feedback loops. Within this architecture, behavioral control is organized hierarchically, moving from concrete motor acts (muscle movements) to programs of action, to principles (values and traits), and ultimately up to system concepts, which embody the holistic sense of self and ideal identity.

Under normative psychological functioning, negative feedback loops continuously monitor input against reference values. When an individual encounters an environmental obstacle or fails an examination, discrepancy detection triggers a secondary meta-monitoring loop that assesses the rate of progress toward discrepancy reduction. If the rate of progress is judged acceptable, confidence remains high, positive affect is maintained, and effort continues. If the rate of progress is deemed unacceptably slow or stalled, negative affect ensues.

The critical theoretical premise underlying the ATS is that cognitive vulnerability arises when feedback loops across different hierarchical levels become pathologically coupled. In a psychologically resilient individual, feedback loops are compartmentalized: a failure at the program or principle level (e.g., getting a low grade on a single chemistry midterm) triggers local adjustments without collapsing higher-order loops. The person evaluates the programmatic failure, revises study habits, or concludes that chemistry is difficult, leaving their overarching system concept (“I am a worthwhile, capable human being”) intact.

In contrast, individuals with high ATS Generalization exhibit extreme vertical coupling across their self-regulatory hierarchy. A discrepancy at the concrete performance level instantly penetrates the boundary of the global system concept. Carver, La Voie, Kuhl, and Ganellen (1988) integrated this cybernetic framework with Aaron T. Beck’s cognitive theory of depression. Beck posited that depressed individuals activate latent, rigid cognitive schemas characterized by overgeneralization, all-or-nothing thinking, and arbitrary inference. The ATS operationalized these abstract schema concepts into concrete, testable self-regulatory parameters.

Moreover, the framework incorporates the diathesis-stress paradigm. High generalization does not continuously produce depression in the absence of failure; rather, it remains a latent cognitive diathesis. Only when an individual encounters an adverse event or perceived personal shortfall does this cognitive vulnerability activate, initiating rapid downward affective spirals, perceived helplessness, subjective hopelessness, and behavioral disengagement.

Validity

The psychometric validity of the Attitudes Toward Self scale has been extensively demonstrated across cross-sectional, prospective, longitudinal, and clinical paradigms over more than three decades.

Construct and Factorial Validity

Initial construct validation by Carver et al. (1988) confirmed that the three subscales—High Standards, Self-Criticism, and Generalization—form distinct conceptual factors that diverge markedly in their relations with affective outcomes. Factor analysis demonstrated that items cleanly segregated onto their hypothesized latent constructs without cross-loading. In confirmation of construct distinctiveness, High Standards exhibited near-zero correlations with depressive symptomatology, whereas Generalization demonstrated robust, significant positive correlations with validated measures of depression, including the Beck Depression Inventory (BDI) and the Center for Epidemiologic Studies Depression Scale (CES-D).

Prospective Predictive Validity

The definitive test of a cognitive diathesis is its ability to interact with life stress to predict symptom onset over time. In a landmark prospective investigation, Carver (1998) examined university students facing acute evaluative stress across a semester. While neither High Standards nor Self-Criticism interacted with adverse events to predict depressive escalation, the Generalization subscale demonstrated a potent, statistically significant diathesis-stress interaction: students who endorsed high generalization and subsequently experienced significant adverse academic or interpersonal events exhibited marked increases in depressive symptoms, whereas those with low generalization remained resilient even under severe stress.

Convergent and Discriminant Validity

Convergent validity has been established through positive correlations between ATS Generalization and related constructs, such as the Dysfunctional Attitude Scale (DAS; Weissman & Beck), perfectionistic concerns on the Frost Multidimensional Perfectionism Scale (FMPS), and high neuroticism. However, the ATS demonstrates superior discriminant efficiency: while the DAS blends perfectionism, approval dependency, and performance vulnerability across 40 items, the ATS isolates the specific mechanism of generalization in just 4 items. Furthermore, Hayes, Harris, and Carver (2004) demonstrated that ATS Generalization uniquely predicts self-esteem instability (moment-to-moment volatility in self-evaluations), whereas Self-Criticism and High Standards do not.

Discriminative Validity in Bipolar Disorder vs. Unipolar Depression

A seminal theoretical advance emerged from studies comparing cognitive profiles across unipolar depression and bipolar disorder. Eisner, Johnson, and Carver (2008) and Carver and Johnson (2009) examined whether cognitive responses to failure and success relate uniquely to depressive versus manic vulnerabilities. Utilizing the ATS alongside measures of manic tendencies (such as the Hypomanic Personality Scale), the researchers discovered that the Generalization subscale was selectively and uniquely associated with depressive symptoms, but completely decoupled from tendencies toward hypomania or mania. In contrast, manic vulnerability was linked to elevated reward-responsiveness and high goal pursuit. This divergence provided profound discriminant validity, proving that generalization is a specific marker of depressive vulnerability rather than a generalized proxy for broad psychopathology or emotional reactivity.

Reliability

The ATS demonstrates strong internal consistency and temporal stability across a diverse spectrum of non-clinical collegiate populations, community cohorts, and psychiatric outpatients.

Internal Consistency

Across multiple independent validation samples reported by Carver et al. (1988), Cronbach’s alpha (α) coefficients for the subscales exhibited satisfactory to strong internal reliability despite the brief length of each scale:

• High Standards Subscale (3 items): Cronbach’s α typically ranges from .74 to .78 (mean α ≈ .76).
• Self-Criticism Subscale (3 items): Cronbach’s α typically ranges from .76 to .81 (mean α ≈ .78).
• Generalization Subscale (4 items): Cronbach’s α typically ranges from .77 to .82 (mean α ≈ .78–.80).

Given that Cronbach’s alpha is mathematically sensitive to scale length, obtaining coefficients near or exceeding .80 for subscales comprising only 3 to 4 items indicates robust item intercorrelations and high conceptual homogeneity within each latent domain.

Test-Retest Reliability and Temporal Stability

Investigations evaluating the temporal stability of the ATS over multi-week and multi-month intervals indicate that the scale assesses enduring cognitive traits rather than transient mood-state artifacts. Test-retest reliability coefficients over intervals spanning 4 to 10 weeks have consistently yielded Pearson correlations (r) between .68 and .79 across all three subscales. Importantly, while mean levels of depressive symptoms fluctuate substantially in response to environmental shifts, baseline ATS Generalization scores exhibit marked rank-order stability over time, fulfilling the prerequisite criteria for a genuine cognitive vulnerability trait.

Factor Analysis

The structural dimensionality of the Attitudes Toward Self scale has been subjected to rigorous exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) across numerous empirical studies.

Exploratory Factor Analysis (EFA)

In the foundational scale development studies conducted by Carver et al. (1988), principal factor extractions followed by both orthogonal (Varimax) and oblique (Promax/Oblimin) rotations consistently yielded a clean, unforced three-factor solution based on the scree test and eigenvalues exceeding 1.0 (Kaiser-Guttman criterion). The three latent factors accounted for the vast majority of common variance:

• Factor 1: Generalization accounted for the largest proportion of common variance, with items 2, 5 (reverse-coded), 8, and 10 exhibiting primary factor loadings ranging from .62 to .81, with negligible cross-loadings (< .20) on alternative factors.
• Factor 2: High Standards comprised items 1, 4, and 7, with primary factor loadings spanning .68 to .84. These items loaded cleanly on an independent dimension reflecting aspirational benchmarks.
• Factor 3: Self-Criticism comprised items 3, 6, and 9, with primary factor loadings ranging from .65 to .83, capturing immediate self-directed punitive affect following failure.

Confirmatory Factor Analysis (CFA) and Structural Fit

Subsequent structural equation modeling and CFA evaluations (e.g., Carver, 1998; Hayes et al., 2004) have systematically tested the theoretical three-factor model against competing one-factor (unidimensional vulnerability) and two-factor (combining High Standards and Self-Criticism, or combining Self-Criticism and Generalization) models. The three-factor correlated model routinely demonstrates superior fit indices:

• Root Mean Square Error of Approximation (RMSEA): ≤ .048 (indicating excellent fit; 90% CI [.028, .064])
• Comparative Fit Index (CFI): ≥ .96
• Tucker-Lewis Index (TLI): ≥ .95
• Standardized Root Mean Square Residual (SRMR): ≤ .041

Alternative single-factor models exhibit catastrophic drops in fit (CFI < .75, RMSEA > .13), confirming that cognitive vulnerability cannot be conceptualized as an undifferentiated, unitary construct. Inter-factor correlations indicate that while High Standards and Self-Criticism correlate moderately (r ≈ .35 to .45), and Self-Criticism correlates moderately with Generalization (r ≈ .40 to .50), High Standards exhibits a near-orthogonal or weak association with Generalization (r ≈ .05 to .18). This factorial separation explains why individuals can maintain high personal standards without experiencing depressive overgeneralization.

Instrument / Measurement Tool

• Instrument Name: Attitudes Toward Self (ATS)
• Author: Charles S. Carver, Ph.D.
• Constructs Measured: Cognitive vulnerability to depression, High Standards, Self-Criticism, and Generalization from failure.
• Administration Format: Self-administered paper-and-pencil or computerized self-report questionnaire.
• Completion Time: Approximately 2 to 4 minutes.
• Total Item Count: 10 items.
• Response Scale: 5-point Likert-type scale:
  • 1 = I agree a lot
  • 2 = I agree a little
  • 3 = I'm in the middle–I neither agree nor disagree
  • 4 = I Disagree a little
  • 5 = I Disagree a lot
• Scoring and Directionality:
  • Items are scored such that higher scores reflect greater levels of the construct (higher cognitive vulnerability/higher standards).
  • Because the response anchors place agreement at 1 and disagreement at 5, reverse-coding is required for all items EXCEPT Item 5 (i.e., Items 1, 2, 3, 4, 6, 7, 8, 9, 10 must be recoded: 1 → 5, 2 → 4, 3 → 3, 4 → 2, 5 → 1).
  • Item 5 is positively phrased (“I hardly ever let unhappiness over one bad time influence my feelings about other parts of my life”); therefore, it retains its direct raw response value (or is scored directly: 1 = low generalization, 5 = high generalization), aligning with the recoded items where higher values indicate higher vulnerability.
• Subscale Composition:
  • High Standards Subscale: Items 1, 4, 7 (Total score range: 3–15).
  • Self-Criticism Subscale: Items 3, 6, 9 (Total score range: 3–15).
  • Generalization Subscale: Items 2, 5, 8, 10 (Total score range: 4–20).

Permissions & Fee and Test Year

The Attitudes Toward Self (ATS) scale was originally developed and published in 1988 by Charles S. Carver and colleagues. Dr. Carver adhered to an open-science philosophy, making his assessment instruments freely accessible to the global psychological and psychiatric research communities. The ATS is in the public domain for non-commercial academic research and clinical educational purposes. No licensing fees or royalty payments are required for legitimate educational or scientific investigations. Researchers and clinicians utilizing the ATS are requested to cite the original validation publications appropriately. Commercial deployment, automated diagnostic software integration, or inclusion in fee-for-service commercial testing packages requires explicit consultation with the Carver estate or copyright holders of the original publications.

References

Carver, C. S. (1998). Generalization, adverse events, and development of depressive symptoms. Journal of Personality, 66(4), 609–620. https://doi.org/10.1111/1467-6494.00027

Carver, C. S., & Johnson, S. L. (2009). Tendencies toward mania and tendencies toward depression have distinct motivational, affective, and cognitive correlates. Cognitive Therapy and Research, 33(6), 552–569. https://doi.org/10.1007/s10608-009-9255-0

Carver, C. S., La Voie, L., Kuhl, J., & Ganellen, R. J. (1988). Cognitive concomitants of depression: A further examination of the roles of generalization, high standards, and self-criticism. Journal of Social and Clinical Psychology, 7(4), 350–365. https://doi.org/10.1521/jscp.1988.7.4.350

Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-theory approach to human behavior. Springer-Verlag. https://doi.org/10.1007/978-1-4612-5887-2

Carver, C. S., & Scheier, M. F. (1990). Origins and functions of positive and negative affect: A control-process view. Psychological Review, 97(1), 19–35. https://doi.org/10.1037/0033-295X.97.1.19

Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. Cambridge University Press. https://doi.org/10.1017/CBO9781139174794

Eisner, L. R., Johnson, S. L., & Carver, C. S. (2008). Cognitive responses to failure and success relate uniquely to bipolar depression versus mania. Journal of Abnormal Psychology, 117(1), 154–163. https://doi.org/10.1037/0021-843X.117.1.154

Hayes, A. M., Harris, M. S., & Carver, C. S. (2004). Predictors of self-esteem variability. Cognitive Therapy and Research, 28(3), 369–385. https://doi.org/10.1023/B:COTR.0000031807.50244.60

Items of the Scale