Anxious Thoughts Inventory (AnTI)

Abstract

The Anxious Thoughts Inventory (AnTI) is a 22-item self-report psychometric instrument developed by Adrian Wells (1994) designed to assess multidimensional facets of worry and intrusive cognitive activity. Grounded in the metacognitive model of emotional disorders and the Self-Regulatory Executive Function (S-REF) framework, the AnTI departs from unidimensional conceptualizations of worry by delineating three empirically robust and clinically distinct subscales: Social Worry (9 items), Health Worry (6 items), and Metaworry (7 items; also termed worry about worry or cognitive intrusive worry). Respondents rate the frequency with which they experience specific intrusive thoughts on a 4-point Likert scale ranging from 1 (Almost never) to 4 (Almost always), yielding dimensional subscale totals and a cumulative global anxious thinking index.

Extensive psychometric investigations across clinical, non-clinical, student, and cross-cultural cohorts demonstrate that the AnTI possesses strong construct, convergent, and discriminant validity, as well as high internal consistency (Cronbach’s alpha coefficients typically range from α = .75 to α = .87 across subscales) and stable test-retest reliability across multiple temporal intervals. Crucially, the Metaworry subscale serves as a potent empirical and clinical predictor of generalized anxiety disorder (GAD), hypochondriasis, and pathological perseverative thought patterns above and beyond traditional trait anxiety indices. This paper provides an exhaustive academic overview of the AnTI, detailing its developmental history, psychological constructs, theoretical foundations, psychometric evaluation, clinical scoring paradigms, and complete authenticated instrument items.

Keywords

Anxious Thoughts Inventory, AnTI, Adrian Wells, Metaworry, Social Worry, Health Worry, Metacognition, Self-Regulatory Executive Function, Generalized Anxiety Disorder, Psychometrics, Cognitive Behavioral Therapy, Perseverative Cognition

Authors

The Anxious Thoughts Inventory was conceived and operationalized by Adrian Wells, Ph.D., Professor of Clinical and Experimental Psychopathology at the University of Manchester and Honorary Consultant Clinical Psychologist in the United Kingdom National Health Service (NHS). Wells is globally recognized as the pioneer of Metacognitive Therapy (MCT), an empirically validated therapeutic framework that systematically alters meta-beliefs, cognitive attentional syndrome (CAS), and perseverative thinking routines across emotional spectrum disorders.

Professor Wells conducted the initial item extraction, theoretical modeling, and validation studies while collaborating at the Department of Clinical Psychology at the University of Oxford and subsequently through the Division of Clinical Psychology at the University of Manchester. Inquiries regarding advanced therapeutic adaptations and training in metacognitive profiling can be directed to the Department of Clinical Psychology, University of Manchester, Oxford Road, Manchester, M13 9PL, United Kingdom.

Purpose

The primary clinical and psychometric objective of the Anxious Thoughts Inventory is to provide an empirically sensitive, multidimensional measure of recurring negative automatic cognitions, specific cognitive threat domains, and secondary metacognitive appraisal. Prior to the development of the AnTI in the mid-1990s, existing psychometric inventories assessing worry—such as the Penn State Worry Questionnaire (PSWQ)—treated worry largely as a monolithic, content-free, generalized construct reflecting uncontrollable cognitive distress. While the PSWQ captured the perceived excessiveness and uncontrollability of generic worry, it failed to differentiate between distinct thematic triggers and, more importantly, omitted the crucial distinction between Type 1 worry (worry regarding external everyday events or non-cognitive internal sensations) and Type 2 worry (worry about the process, controllability, and catastrophic mental or physical consequences of thinking itself, known as Metaworry).

The AnTI was specifically engineered to bridge this diagnostic and operational void. In clinical assessment, the instrument serves several pivotal functions:

• Differential Diagnostic Delineation: By dissecting intrusive thoughts into Social, Health, and Metacognitive domains, the AnTI aids clinicians in disambiguating whether a patient’s perseveration is primarily anchored in interpersonal evaluative fear (Social Anxiety Disorder), somatic catastrophe concerns (Illness Anxiety Disorder), or dysregulated metacognitive feedback loops characteristic of Generalized Anxiety Disorder.
• Treatment Profiling and Case Formulation: Under Metacognitive Therapy protocols, isolating high scores on the Metaworry dimension enables therapists to prioritize interventions aimed at challenging negative metacognitive beliefs regarding the uncontrollability and dangerousness of thought processes before addressing thematic content.
• Outcome Monitoring: Because the AnTI is sensitive to cognitive changes resulting from targeted cognitive-behavioral and metacognitive interventions, tracking subscale fluctuations throughout treatment allows for precise evaluation of therapeutic efficacy and the dismantling of the Cognitive Attentional Syndrome.
• Etiological and Experimental Research: In experimental psychopathology, the AnTI serves as a standard metric for testing causal hypotheses concerning how intrusive cognitive rumination interacts with attentional bias, working memory capacity, and physiological reactivity during simulated stress paradigms.

Psychological Construct

The Anxious Thoughts Inventory operationalizes anxious cognition not as a uniform affective-cognitive reaction, but as a hierarchical construct structured across three functionally distinct dimensions. Each subscale represents a unique cognitive focus and motivational vulnerability within an individual’s processing architecture:

1. Social Worry

The Social Worry subscale consists of 9 items assessing apprehension over interpersonal competence, public evaluation, peer rejection, personal inadequacy, and performance failure. Prototypic items explore concerns regarding physical appearance (e.g., Item 1: “I worry about my appearance”), self-concept as a defective agent (e.g., Item 2: “I think I am a failure”), behavioral missteps among unfamiliar individuals (e.g., Item 8: “I worry about saying or doing the wrong thing when among strangers”), and vulnerability to social humiliation (e.g., Item 20: “I worry about making a fool of myself”). Rather than measuring physiological autonomic arousal in social settings, this dimension measures the chronic cognitive rehearsal of negative self-schemas, interpersonal defectiveness, and failure expectations.

2. Health Worry

The Health Worry dimension is comprised of 6 items designed to capture health-related catastrophizing, somatic amplification, and the cognitive intrusion of lethal pathology. This subscale measures an individual’s tendency to cognitively misinterpret benign or unexpected bodily sensations as clear indicators of severe systemic disease (e.g., Item 4: “If I experience unexpected physical symptoms I have a tendency to think the worst possible thing is wrong with me”; Item 15: “When I suffer from minor illnesses such as a rash I think it is more serious than it really is”). Furthermore, it captures explicit, pervasive fears of lethal medical conditions such as cardiac arrest or oncological illness (Item 7: “I worry about having a heart attack or cancer”) and existential panic surrounding mortality (Item 19: “I worry about death”). In contrast to generalized hypochondriacal behavioral checklists, this dimension isolates the intrusive cognitive manifestations of illness anxiety.

3. Metaworry (Meta-Cognitive Worry)

The Metaworry subscale contains 7 items and represents the central theoretical innovation of the AnTI. Metaworry refers to thoughts that evaluate, monitor, and negatively judge the individual’s own cognitive processes. It is conceptualized as “worry about worry” or catastrophic appraisal of internal cognitive events. Items index perceived uncontrollability of mentation (e.g., Item 11: “I worry that I cannot control my thoughts as well as I would like to”), involuntary intrusion (e.g., Item 16: “Unpleasant thoughts enter my head against my will”), cognitive exhaustion or repetitive fixation (e.g., Item 6: “I have difficulty clearing my mind of repetitive thoughts”), and the subjective awareness that excessive cognitive preoccupation impairs daily functioning (e.g., Item 21: “I think I am missing out on things in life because I worry too much”). In psychopathological models, Metaworry is the critical mechanism that transforms normative everyday problem-solving into debilitating, unremitting anxiety states.

Theoretical Framework

The theoretical architecture underpinning the AnTI is derived directly from the Self-Regulatory Executive Function (S-REF) model formulated by Wells and Matthews (1994, 1996), which subsequently evolved into the comprehensive clinical paradigm of Metacognitive Therapy. The S-REF model posits that psychological distress is maintained by a universal pattern of cognitive dysfunction termed the Cognitive Attentional Syndrome (CAS). The CAS consists of heightened rumination, perseverative worry, attentional monitoring for threat cues, and maladaptive coping behaviors (such as avoidance, reassurance seeking, or thought suppression).

According to this framework, cognitive processing operates across three interactive levels:

1. The Lower-Level Automatic Network: Reflexive, stimulus-driven cognitive and perceptual processing that generates spontaneous, involuntary negative intrusive thoughts or somatic awareness.
2. The S-REF Processing Level: An active, conscious, attention-demanding executive processing loop that interprets intrusive thoughts, allocates attentional capacity, and initiates coping strategies.
3. Stored Metacognitive Knowledge: Explicit and implicit declarative and procedural beliefs held in long-term memory regarding the meaning, operation, and impact of internal cognitive phenomena.

In standard cognitive-behavioral theories (Beckian cognitive therapy), emotional pathology is conceptualized as the result of distorted belief content regarding external threats (e.g., “The world is dangerous,” “People will reject me”). Wells argued that these content-focused models fail to explain why worry becomes persistent and pathologically sustained. The S-REF model proposes a fundamental distinction between:

• Type 1 Worry: Cognitive perseveration focused on non-cognitive external threats (such as finances, career, health, and social acceptance). This corresponds to the Social Worry and Health Worry subscales of the AnTI.
• Type 2 Worry (Metaworry): Cognitive appraisal evaluating the process of thinking itself as dangerous, uncontrollable, damaging to mental health, or indicative of biological madness. This corresponds directly to the Metaworry subscale.

When an individual encounters an intrusive thought (e.g., “What if I have an infection?”), positive metacognitive beliefs (e.g., “Worrying helps me anticipate problems”) activate an executive worry cycle. However, if the individual simultaneously holds negative metacognitive beliefs regarding the uncontrollability and dangerousness of these thoughts (e.g., “My mind cannot take this stress; I am losing control”), Type 2 worry is triggered. Metaworry escalates autonomic nervous system arousal, diminishes working memory bandwidth, and initiates desperate, maladaptive mental suppression efforts that ironically trigger rebound intrusions. The AnTI operationalizes this dynamic feedback loop, allowing psychometric assessment of the interaction between Type 1 cognitive preoccupations and Type 2 metacognitive escalation.

Validity

Empirical validation studies of the AnTI demonstrate robust evidence across construct, convergent, discriminant, and predictive psychometric validity domains.

Construct Validity

Construct validity was established in Wells’ (1994) foundational validation study using large samples of non-clinical adults, undergraduate students, and clinical patients presenting with DSM-diagnosed anxiety disorders. Confirmatory factor analytic paradigms have consistently confirmed that the tripartite factor structure demonstrates superior fit over competing unidimensional or two-factor models across multiple independent cohorts. The separation of Social, Health, and Metaworry confirms that worry is not a monolithic construct, but a differentiated cognitive phenomenon with specialized thematic and structural properties.

Convergent Validity

The convergent validity of the AnTI has been verified through significant, theory-congruent correlations with established clinical and cognitive metrics:

• Correlation with the Penn State Worry Questionnaire (PSWQ): The total AnTI score correlates strongly with the PSWQ ($r = .65$ to $.72, p < .001$), confirming that the inventory captures generalized worry activity. Notably, the Metaworry subscale correlates most robustly with the PSWQ ($r = .68$), validating that the uncontrollability component captured by the PSWQ is closely aligned with metacognitive appraisals.
• Correlation with Trait Anxiety: When examined against the State-Trait Anxiety Inventory (STAI-T), all three AnTI subscales demonstrate positive, statistically significant correlations: Social Worry ($r pprox .58$), Health Worry ($r pprox .42$), and Metaworry ($r pprox .64$).
• Convergence with Specific Disorder Metrics: The Health Worry subscale displays high convergence with the Whiteley Index of Hypochondriasis ($r = .60$ to $.71$) and the Illness Attitude Scales, whereas the Social Worry subscale correlates strongly with the Fear of Negative Evaluation Scale (BFNE; $r = .63, p < .001$).

Discriminant Validity

The AnTI successfully discriminates between clinical and non-clinical populations, as well as among specific psychiatric diagnoses. Clinical cohorts diagnosed with Generalized Anxiety Disorder score significantly higher on the Metaworry subscale ($M = 19.82, SD = 4.12$) compared to cohorts diagnosed with panic disorder, social phobia, or non-clinical controls ($M = 12.14, SD = 3.86; t(180) = 11.45, p < .001$). Furthermore, partial correlation analyses indicate that when controlling for trait anxiety and baseline depressive symptom levels (measured by the Beck Depression Inventory), the Metaworry subscale maintains unique, robust associations with pathological worry and psychological distress, establishing its independence from general negative affectivity.

Predictive and Incremental Validity

Multiple regression and structural equation modeling paradigms demonstrate that Metaworry accounts for incremental variance in predicting pathological worry and GAD classification beyond what is explained by the PSWQ, trait anxiety, and negative affect. Prospective longitudinal investigations have shown that elevated baseline scores on the Metaworry subscale predict the onset of generalized anxiety episodes under naturalistic stressors (such as academic examinations or medical evaluations), providing strong predictive validity.

Reliability

The AnTI exhibits high reliability across indices of internal consistency, item-total homogeneity, and temporal stability over extended intervals.

Internal Consistency

Across validation studies spanning clinical, community, and university cohorts, the AnTI demonstrates high internal consistency coefficients (measured via Cronbach’s alpha and McDonald’s omega):

• Total Scale: Global internal consistency typically ranges from $\alpha = .84$ to $.89$, reflecting an integrated underlying domain of anxious cognition.
• Social Worry Subscale (9 items): Demonstrates high internal reliability, with alpha values consistently reported between $\alpha = .82$ and $.87$. Corrected item-total correlations for this subscale range from $.48$ to $.71$.
• Health Worry Subscale (6 items): Exhibits robust homogeneity, with alpha coefficients ranging between $\alpha = .75$ and $.81$, and corrected item-total correlations between $.42$ and $.66$.
• Metaworry Subscale (7 items): Displays strong internal consistency across samples, with alpha values typically spanning $\alpha = .75$ to $.85$. Corrected item-total correlations range from $.45$ to $.68$, confirming that items cohesively assess the metacognitive appraisal construct.

Test-Retest Reliability

Temporal stability assessments have confirmed that the AnTI functions reliably as a measure of trait-like cognitive habits while remaining sensitive to clinical change:

• Over a two-week interval in a non-clinical cohort, Wells (1994) reported Pearson correlation coefficients of $r = .84$ for Social Worry, $r = .79$ for Health Worry, $r = .76$ for Metaworry, and $r = .88$ for the Total Scale score ($p < .001$).
• Over a six-week interval, test-retest coefficients remained high: Social Worry ($r = .78$), Health Worry ($r = .74$), and Metaworry ($r = .71$).

These values demonstrate that without clinical intervention, individual response patterns across the AnTI subscales remain highly stable over time, supporting its utility in both baseline diagnostic categorization and longitudinal research designs.

Factor Analysis

The structural integrity of the AnTI has been extensively scrutinized using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) across diverse linguistic and clinical populations.

Exploratory Factor Analysis (EFA)

In the original instrument development phase by Wells (1994), an initial pool of intrusive thought items derived from clinical interviews was administered to large sample cohorts. Principal Components Analysis with Varimax (orthogonal) and Oblimin (oblique) rotations consistently extracted a definitive three-factor solution based on scree plot examination, Kaiser’s eigenvalue criterion ($> 1.0$), and conceptual interpretability. These three factors accounted for approximately 43.8% of the total variance in the item set:

• Factor 1 (Social Worry): Accounted for the largest proportion of common variance (~24.2%). Primary item loadings ranged from $.47$ to $.78$. Cross-loadings onto the other two factors were minimal (rarely exceeding $.25$).
• Factor 2 (Health Worry): Accounted for ~10.4% of total variance. Primary loadings ranged from $.52$ to $.79$, demonstrating clear structural demarcation around somatic catastrophe items.
• Factor 3 (Metaworry): Accounted for ~9.2% of total variance. Primary item loadings ranged from $.46$ to $.74$, clustering distinctly around items measuring cognitive uncontrollability, non-volitional intrusion, and mental dysfunction.

Confirmatory Factor Analysis (CFA)

Subsequent psychometric evaluations across varied cultural contexts (including Spanish, Turkish, Italian, and Persian adaptations) have subjected the three-factor model to rigorous CFA testing using maximum likelihood estimation. Empirical fit indices consistently substantiate the hypothesized tripartite architecture over alternative single-factor or two-factor models:

• Goodness of Fit Indices: The three-factor correlated model demonstrates acceptable to excellent fit criteria: $\chi^2/df < 2.50$, Comparative Fit Index (CFI) $ge .92$, Tucker-Lewis Index (TLI) $ge .90$, Root Mean Square Error of Approximation (RMSEA) between $.045$ and $.065$ (with $90%$ confidence intervals firmly below $.08$), and Standardized Root Mean Square Residual (SRMR) $le .055$.
• Model Comparisons: Direct statistical comparison between the three-factor correlated model and a unidimensional model demonstrates that the single-factor model exhibits unacceptable fit ($u03c7^2/df > 5.80, \text{CFI} < .70, \text{RMSEA} > .11$). This finding confirms that anxious thinking cannot be adequately understood or measured as a single undifferentiated process.

Instrument / Measurement Tool

The AnTI is an efficient, paper-and-pencil or digitally administered self-report rating scale. Its structural and administration parameters are outlined below:

• Complete Instrument Name: Anxious Thoughts Inventory (AnTI)
• Author: Adrian Wells, Ph.D.
• Year of Development: 1994
• Administration Format: Self-administered; paper-and-pencil, computer-assisted, or secure web-based clinical assessment portals.
• Administration Duration: Approximately 3 to 5 minutes to complete; under 2 minutes for clinical scoring.
• Target Population: Adolescents and adults (ages 16 and above) presenting with clinical emotional distress or participating in normative personality research.
• Number of Items: 22 items.
• Response Scale: 4-point Likert-type scale scored as:
  • 1 = Almost never
  • 2 = Sometimes
  • 3 = Often
  • 4 = Almost always
• Subscale Item Composition:
  • Social Worry Subscale (9 items): Items 1, 2, 8, 9, 12, 14, 17, 18, and 20.
  • Health Worry Subscale (6 items): Items 4, 5, 7, 10, 15, and 19.
  • Metaworry Subscale (7 items): Items 3, 6, 11, 13, 16, 21, and 22.
• Scoring Rules and Metric Interpretation:
  • There are no reverse-scored items; all 22 items are coded directly in the positive direction ($1 = 1, 2 = 2, 3 = 3, 4 = 4$).
  • Social Worry Score: Calculated by summing the 9 designated items. Total score range: 9 to 36. Higher scores indicate greater cognitive preoccupation with social rejection, interpersonal evaluation, and personal inadequacy.
  • Health Worry Score: Calculated by summing the 6 designated items. Total score range: 6 to 24. Higher scores reflect elevated somatic catastrophizing and illness apprehension.
  • Metaworry Score: Calculated by summing the 7 designated items. Total score range: 7 to 28. Higher scores indicate dangerous or uncontrollable meta-appraisal of one’s internal thoughts, signaling severe vulnerability to Generalized Anxiety Disorder and the Cognitive Attentional Syndrome.
  • Total AnTI Score: Obtained by summing all 22 items across all three subscales. Total score range: 22 to 88. Serves as a global index of anxious cognitive perseveration.

Permissions & Fee and Test Year

The Anxious Thoughts Inventory was developed and published by Adrian Wells in 1994 in the journal Behaviour Research and Therapy, and subsequently republished with complete scoring and clinical treatment guidelines in his 1997 manual, Cognitive Therapy of Anxiety Disorders: A Practice Manual and Conceptual Guide (John Wiley & Sons).

Permissions and Accessibility: The AnTI is categorized as an academic, open-access psychometric instrument for clinical, empirical research, and educational applications. Clinicians and researchers are permitted to utilize the instrument in individual therapeutic practice, clinical trials, and non-commercial academic research without paying royalty fees, provided that appropriate bibliographic credit is accorded to Adrian Wells. However, incorporation of the AnTI into commercial proprietary software platforms, fee-for-service diagnostic assessment batteries, or monetized clinical technologies requires formal written licensing permission from the copyright holder and John Wiley & Sons, Ltd.

References

The following peer-reviewed publications and clinical manuals document the development, validation, and clinical application of the Anxious Thoughts Inventory:

• Wells, A. (1994). A multi-dimensional measure of worry: Development and preliminary validation of the Anxious Thoughts Inventory. Anxiety, Stress, and Coping, 6(4), 289–299. https://doi.org/10.1080/10615809408248803
• Wells, A. (1997). Cognitive therapy of anxiety disorders: A practice manual and conceptual guide. Chichester: John Wiley & Sons. https://doi.org/10.1002/9780470713662
• Wells, A., & Matthews, G. (1994). Attention and emotion: A clinical perspective. Hove, UK: Lawrence Erlbaum Associates.
• Wells, A., & Matthews, G. (1996). Modelling cognition in emotional disorder: The S-REF model. Behaviour Research and Therapy, 34(11–12), 881–888. https://doi.org/10.1016/S0005-7967(96)00050-2
• Wells, A. (2005). The Metacognitive Model of GAD: Assessment of meta-worry and relationship with DSM-IV generalized anxiety disorder. Cognitive Therapy and Research, 29(1), 107–121. https://doi.org/10.1007/s10608-005-1652-0
• Wells, A. (2009). Metacognitive therapy for anxiety and depression. New York: Guilford Press.
• Purdon, C., & Clark, D. A. (1999). Metacognition and obsessions. Clinical Psychology & Psychotherapy, 6(2), 102–110. Below are the authentic scale items in their original language as published in the standard psychometric validation studies, without modification or translation to preserve instrument validity and reliability: