Penn State Worry Questionnaire (PSWQ)

Abstract

The Penn State Worry Questionnaire (PSWQ) is widely recognized as the gold-standard, self-report psychometric instrument designed to assess the generality, excessiveness, and uncontrollability of pathological worry. Developed by Thomas D. Borkovec and colleagues at Pennsylvania State University in 1990, the instrument was engineered to measure worry as a continuous, trait-like cognitive construct rather than a transient emotional state. Unlike earlier anxiety inventories that primarily conflated somatic tension, autonomic arousal, or specific situational fears, the PSWQ isolates the cognitive activity of worry—defined as a chain of negatively valenced thoughts and images aimed at mental problem-solving under conditions of perceived threat. Comprising 16 items evaluated on a 5-point Likert scale (ranging from 1 = Not at all typical of me to 5 = Very typical of me), the instrument yields a total score ranging from 16 to 80, with higher scores reflecting elevated levels of pathological, pervasive worry.

Psychometrically, the PSWQ demonstrates exceptional internal consistency across diverse clinical and non-clinical populations, with Cronbach’s alpha coefficients consistently spanning between .88 and .95, alongside robust test-retest reliability across intervals ranging from two to ten weeks (r = .74 to .93). Factor analytic investigations have sparked meaningful psychometric inquiry: while originally conceptualized as strictly unidimensional, confirmatory factor analyses frequently reveal that a two-factor solution—differentiating substantive worry items from reverse-scored items—or a bifactor model best accounts for item variance, predominantly representing a method effect driven by reverse-worded phrasing. The instrument exhibits substantial convergent validity with measures of generalized anxiety, such as the Generalized Anxiety Disorder 7 (GAD-7) and the State-Trait Anxiety Inventory (STAI-T), while maintaining robust discriminant validity against pure indices of depressive symptom severity and autonomic somatic anxiety. As an essential diagnostic screener and outcome measure for Generalized Anxiety Disorder (GAD) under DSM-III-R, DSM-IV, and DSM-5 criteria, the PSWQ remains foundational in clinical assessment, cognitive-behavioral therapy evaluation, and neurobehavioral research.

Keywords

Penn State Worry Questionnaire, PSWQ, pathological worry, Generalized Anxiety Disorder, cognitive avoidance, psychometrics, factor structure, trait anxiety, uncontrollability, excessiveness, reliability, construct validity, cognitive behavioral therapy, diagnostic cut-off, anxiety assessment

Authors

The Penn State Worry Questionnaire was formulated, validated, and published in 1990 by a team of clinical psychologists and clinical psychometrics researchers affiliated with the Department of Psychology at the Pennsylvania State University (University Park, Pennsylvania, United States):

• Timothy J. Meyer, Ph.D. — Department of Psychology, The Pennsylvania State University. Dr. Meyer spearheaded initial scale construction, item generation, and preliminary empirical trials examining the differentiation of worry from general somatic and affective distress.
• Michael L. Miller, Ph.D. — Department of Psychology, The Pennsylvania State University. Dr. Miller contributed extensively to the psychometric validation, cross-sample reliability testing, and data analytic structuring of the original validation cohorts.
• Richard L. Metzger, Ph.D. — Department of Psychology, The Pennsylvania State University (later affiliated with the Department of Psychology, Mercer University). Dr. Metzger’s expertise in cognitive psychology and experimental assessment guided the methodological framework for quantifying repetitive thought dynamics.
• Thomas D. Borkovec, Ph.D. — Emeritus Professor of Psychology, The Pennsylvania State University. Dr. Borkovec is globally regarded as the preeminent pioneer in the scientific investigation of worry and Generalized Anxiety Disorder. His cognitive avoidance model of worry provided the theoretical architecture underlying the questionnaire’s focus on the non-somatic, functional, and uncontrollable aspects of repetitive negative cognition.

Purpose

The fundamental clinical and scientific impetus behind the creation of the Penn State Worry Questionnaire was the pressing need to develop a pure, content-independent measure of worry. Historically, empirical research in the 1970s and 1980s evaluated anxiety primarily through physiological indices (e.g., heart rate, galvanic skin conductance, muscle tension) or broad affective inventories such as the Taylor Manifest Anxiety Scale and the Spielberger State-Trait Anxiety Inventory. While these tools captured somatic arousal and general negative affectivity, they largely failed to measure the central cognitive component that defines Generalized Anxiety Disorder (GAD) as codified by the American Psychiatric Association in the Diagnostic and Statistical Manual of Mental Disorders: persistent, excessive, and uncontrollable worry.

Prior worry instruments frequently evaluated specific worry topics or contextual domains—such as financial distress, academic performance, health fears, or interpersonal conflict (e.g., the Worry Domains Questionnaire). However, this domain-dependent approach suffered from severe construct contamination: individuals might score low simply because their specific worries were omitted from the inventory, or score high merely due to living in high-stress, low-resource environments where realistic concerns do not represent psychopathology. The authors of the PSWQ bypassed this critical limitation by deliberately designing an instrument that measures the process, severity, frequency, and uncontrollability of worry, entirely divorced from the specific topical content of that worry.

In clinical practice, the PSWQ serves several core functions:

• Diagnostic Screening: Serving as a primary clinical screener to identify individuals who meet diagnostic thresholds for GAD, differentiating clinical worriers from non-clinical or normative worriers.
• Differential Diagnosis: Distinguishing GAD from Major Depressive Disorder (MDD), Panic Disorder, Social Anxiety Disorder, and Obsessive-Compulsive Disorder (OCD) by isolating transdiagnostic repetitive thinking from mood-congruent rumination or intrusive ego-dystonic obsessions.
• Treatment Outcome Monitoring: Functioning as a sensitive, responsive metric to evaluate treatment gains in cognitive-behavioral therapy (CBT), acceptance-based behavioral therapies, and pharmacotherapy targeting chronic anxiety.
• Basic Cognitive Research: Providing psychopathology researchers with an empirically robust paradigm to stratify participants into high-worry versus low-worry cohorts for experimental investigations into attentional bias, working memory load, autonomic inflexibility, and emotional processing.

Psychological Construct

The psychological construct evaluated by the PSWQ is pathological trait worry. In clinical psychology and psychopathology, worry is defined as a predominantly verbal-linguistic, future-oriented cognitive process characterized by repetitive thoughts concerning potential negative events, uncertainties, and catastrophic outcomes, accompanied by a perceived inability to modulate or terminate these cognitive intrusions.

The construct measured by the PSWQ is characterized by three core phenomenological dimensions:

1. Excessiveness and Pervasiveness

Normative worry is adaptive, circumscribed, and proportional to real-world threats, functioning as an anticipatory problem-solving mechanism. In contrast, pathological worry measured by the PSWQ is characterized by intense disproportionality: the individual worries extensively about everyday minor matters as well as catastrophic possibilities. Pervasiveness refers to the chronic dispersion of worry across multiple life domains simultaneously (e.g., family, finances, physical safety, work performance, punctuality). Items such as “Many situations make me worry” (Item 4) and “I am always worrying about something” (Item 7) specifically reflect this omnipresent, domain-general quality.

2. Uncontrollability

The defining hallmark of clinical worry—and the diagnostic criterion that most strongly predicts functional impairment—is the subjective experience of uncontrollability. Individuals with high scores on the PSWQ report that once a worrisome thought sequence initiates, they lack the executive or cognitive inhibitory control to halt its progression. This subjective helplessness is captured directly by items such as “I know I should not worry about things, but I just cannot help it” (Item 5) and “Once I start worrying, I cannot stop” (Item 14). Uncontrollability is closely linked to deficits in working memory updating and cognitive shifting, which prevent the disengagement of attention from internal threat-related stimuli.

3. Trait-Level Stability and Generalization

The PSWQ measures worry as an enduring, habitual cognitive disposition rather than an acute state elicited by an immediate stressor. Items like “I have been a worrier all my life” (Item 12) evaluate this chronological chronicity. Crucially, the construct is distinguished from depressive rumination. While rumination is predominantly past-oriented, focusing on themes of loss, failure, and self-worth (“Why did this happen to me?”), worry is relentlessly future-oriented, fixated on perceived threat, vulnerability, and catastrophic prevention (“What if something terrible happens?”). The PSWQ specifically isolates this anticipatory cognitive activity.

Theoretical Framework

The theoretical architecture undergirding the PSWQ is heavily rooted in cognitive-behavioral paradigms of emotional disorders, most notably Borkovec’s Cognitive Avoidance Model of Worry, as well as subsequent metacognitive and emotion regulation theories.

Borkovec’s Cognitive Avoidance Model of Worry

Pioneered by Thomas D. Borkovec (1994), this foundational theory posits that worry functions as an internal, negative reinforcement strategy designed to prevent deeper emotional processing of somatic arousal and catastrophic mental imagery. Borkovec observed that verbal-linguistic mentation (the verbal internal dialogue characteristic of worry) elicits significantly less autonomic physiological reactivity than does vivid visual imagery of feared events. Consequently, high worriers rapidly deploy abstract, verbal worry as a cognitive avoidance strategy to suppress distressing physiological sensations and painful somatic states associated with core fears.

Furthermore, because the catastrophic events that individuals worry about rarely occur in reality, the non-occurrence of the catastrophe is mistakenly attributed by the individual to the act of worrying itself (“Because I worried about it, it didn’t happen”). This superstitious reinforcement mechanism solidifies worry as a chronic, habitual coping mechanism, directly explaining why items on the PSWQ reflect an ongoing, uncontrollable impulse to worry across situations.

Wells’ Metacognitive Model

Adrian Wells extended this framework through the Metacognitive Model of GAD, which distinguishes between Type 1 worry (worry about external events or non-cognitive internal states) and Type 2 worry (meta-worry, or worry about worry itself). According to Wells, pathological worry develops when an individual adopts positive metacognitive beliefs (e.g., “Worrying helps me stay prepared and safe”), followed by the emergence of negative metacognitive beliefs concerning the uncontrollability and dangerousness of worry (e.g., “My worrying is out of control and could make me go crazy”). The PSWQ items, particularly those assessing perceived lack of control (e.g., “My worries overwhelm me”), tap directly into the experiential distress generated by these negative metacognitive appraisals.

The Contrast Avoidance Model

More recently, the Contrast Avoidance Model formulated by Michelle Newman and Sandra Llera (2011) provides additional theoretical grounding for the PSWQ’s focus on chronic worry. This theory posits that individuals with GAD are highly sensitive to sharp, sudden shifts in negative emotional states (such as an unexpected shock or devastating disappointment). By maintaining a chronically elevated, sustained level of anticipatory negative affect and worry, the individual avoids experiencing a sudden, intolerable negative emotional contrast. Trait worry, as quantified by the PSWQ, represents this sustained affective and cognitive baseline plateau.

Validity

The Penn State Worry Questionnaire has been subjected to some of the most rigorous psychometric evaluations in the history of clinical assessment. Its construct, convergent, discriminant, and criterion-related validity have been cross-validated across clinical, non-clinical, college, and cross-cultural cohorts.

Convergent Validity

The PSWQ demonstrates robust, statistically significant correlations with other self-report and clinician-rated measures of anxiety and repetitive thought:

• State-Trait Anxiety Inventory – Trait Scale (STAI-T): Moderate-to-high correlations typically ranging from r = .64 to .79, reflecting shared variance in general trait anxiety while confirming that the PSWQ specifically targets cognitive apprehension.
• Generalized Anxiety Disorder 7 (GAD-7): Strong correlations ranging from r = .71 to .82 in outpatient psychiatric settings, validating the questionnaire’s alignment with contemporary DSM GAD criteria.
• Intolerance of Uncertainty Scale (IUS): High correlations (r = .60 to .72), confirming that the cognitive intolerance of ambiguity and unpredictability directly fuels trait worry severity.
• Metacognitions Questionnaire (MCQ): Significant positive associations with the uncontrollability and danger subscales (r = .65 to .75).

Discriminant Validity

A primary psychometric achievement of the PSWQ is its discriminant capacity. Many legacy anxiety measures display excessive collinearity with depression inventories, capturing broad negative affectivity (neuroticism) rather than anxiety-specific constructs. In contrast:

• Separation from Depression: While the PSWQ correlates moderately with measures like the Beck Depression Inventory (BDI; typically r = .36 to .52), multiple regression and structural equation modeling demonstrate that when general negative affect is statistically controlled, the PSWQ continues to uniquely account for variance in GAD diagnosis, whereas depressive measures do not.
• Separation from Somatic Panic: Correlations with somatic anxiety instruments such as the Beck Anxiety Inventory (BAI) and the Anxiety Sensitivity Index (ASI) are noticeably lower (r = .30 to .45), demonstrating that the PSWQ successfully isolates cognitive worry from autonomic hyperarousal, trembling, tachycardia, and catastrophic misinterpretations of bodily sensations.
• Separation from OCD: Although both involve repetitive intrusive cognition, the PSWQ correlates only moderately with measures of obsessive compulsive symptoms (e.g., Yale-Brown Obsessive Compulsive Scale; r = .30 to .40), successfully distinguishing ego-syntonic, realistic catastrophic worries from bizarre, ego-dystonic obsessions.

Criterion and Predictive Validity

The instrument displays exceptional diagnostic accuracy for Generalized Anxiety Disorder. In receiver operating characteristic (ROC) curve analyses, the PSWQ consistently yields an Area Under the Curve (AUC) between .85 and .93 in discriminating individuals with GAD from healthy controls and non-GAD psychiatric outpatients. Optimal clinical cutoff scores established in the literature include:

• Score of 45: Frequently used in non-clinical or community screening environments to flag individuals with elevated worry requiring further clinical evaluation (high sensitivity: > .90).
• Score of 60 to 62: Standard clinical cutoff utilized in clinical trials and diagnostic evaluations. A score of 62 or higher optimizes both sensitivity (.83) and specificity (.85) for identifying a clinical DSM diagnosis of GAD (Fresco et al., 2003; Brown et al., 1992).
• Score of 65+: Indicates severe, deeply disabling pathological worry, commonly observed in severe psychiatric outpatient cohorts.

Reliability

Empirical evaluations consistently document that the PSWQ possesses exemplary reliability characteristics across varied administrative conditions, demographics, and clinical diagnostic categories.

Internal Consistency

Across the published literature, internal consistency for the 16-item scale is exceptionally high. In the original seminal validation study by Meyer, Miller, Metzger, and Borkovec (1990), Cronbach’s alpha was reported at α = .91 in college student samples and α = .95 in a clinical sample of patients diagnosed with Generalized Anxiety Disorder. Subsequent investigations across diverse worldwide populations have mirrored these findings:

• Non-clinical adult community samples: α = .88 to .93.
• Clinical psychiatric outpatient cohorts (GAD, MDD, Panic Disorder): α = .92 to .96.
• Older adult populations: α = .89 to .92, demonstrating the instrument’s reliability across the lifespan (PSWQ-A variants also corroborate these indices).
• Corrected item-total correlations typically exceed .50 for all substantive items, with the vast majority ranging between .60 and .82.

Test-Retest Reliability

Given that the PSWQ conceptualizes worry as a stable personality-like trait, temporal stability is critical. Studies evaluating test-retest reliability across diverse intervals demonstrate remarkable stability in the absence of active clinical intervention:

• 2- to 4-Week Interval: Stability coefficients range from r = .88 to .93, reflecting minimal measurement error and high test-retest consistency.
• 8- to 10-Week Interval: Meyer et al. (1990) documented a test-retest correlation of r = .74 to .78 across a 10-week unmanipulated period.
• Long-term Stability: Over periods exceeding six months, longitudinal non-clinical correlations remain between r = .65 and .72, confirming the enduring, trait-like nature of the construct.

Standard Error of Measurement and Sensitivity to Change

The Standard Error of Measurement (SEM) for the PSWQ is generally estimated between 3.2 and 4.1 points. Using the Jacobson and Truax Reliable Change Index (RCI), a score change of 7 to 9 points or greater post-treatment is recognized as reflecting statistically reliable, clinically meaningful therapeutic improvement rather than standard measurement artifact.

Factor Analysis

The latent structure of the Penn State Worry Questionnaire has been the subject of extensive psychometric and statistical scrutiny using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA).

The Unidimensional Hypothesis vs. Two-Factor Models

Meyer and colleagues (1990) originally reported a unidimensional structure, with a single dominant general factor accounting for the vast majority of shared item variance. However, subsequent empirical examinations (e.g., Brown, Antony, & Barlow, 1992; Beck, Stanley, & Zebb, 1995) regularly demonstrated that a two-factor model demonstrated superior statistical fit in standard CFA testing. These two factors consistently separated as:

• Factor 1 (Worry Engagement / Trait Worry): Comprising the 11 positively phrased items (Items 2, 4, 5, 6, 7, 9, 12, 13, 14, 15, and 16). Item loadings on this substantive factor are consistently strong, usually ranging between .65 and .88.
• Factor 2 (Absence of Worry): Comprising the 5 reverse-scored items (Items 1, 3, 8, 10, and 11), with loadings ranging from .55 to .78.

Method Effect and Bifactor Structural Models

The discovery of this two-factor split raised a critical psychometric debate: Does the PSWQ measure two clinically meaningful, distinct psychological constructs, or is the second factor merely a statistical artifact caused by item phrasing?

Groundbreaking work by Hazlett-Stevens, Ullman, and Craske (2004), alongside structural evaluations by Brown (2003), confirmed that the two-factor solution is a methodological artifact resulting from reverse-worded items (a classic method effect), rather than substantive multidimensionality. When researchers specified a bifactor model—featuring one general substantive trait worry factor alongside an orthogonal method factor for reverse-scored items—the model exhibited exceptional fit indices across both clinical and non-clinical populations:

• Comparative Fit Index (CFI): > .96
• Tucker-Lewis Index (TLI): > .95
• Root Mean Square Error of Approximation (RMSEA): < .05 (90% CI [.041, .059])
• Standardized Root Mean Square Residual (SRMR): < .04

Because the general factor in bifactor analyses accounts for over 85% to 90% of the common variance (Omega Hierarchical, ω_h > .85), psychometric consensus firmly supports retaining and interpreting the PSWQ as an essentially unidimensional measure. Clinicians and researchers are advised to use the unified full-scale total score rather than calculating separate subscales.

Instrument / Measurement Tool

The operational administration parameters and structural specifications of the PSWQ are detailed below:

• Instrument Name: Penn State Worry Questionnaire (PSWQ)
• Assessment Type: Self-report psychometric rating inventory; pencil-and-paper or digital administration
• Target Population: Adults and adolescents aged 16 years and older (a specialized version, the PSWQ-C, exists for children aged 7–16; the PSWQ-A is an abbreviated 8-item version often used for older adults)
• Administration Time: Approximately 3 to 5 minutes
• Total Number of Items: 16 questions
• Response Scale: 5-point Likert scale formatted as:
  • 1 = Not at all typical of me
  • 2 = Rarely typical of me
  • 3 = Somewhat typical of me
  • 4 = Often typical of me
  • 5 = Very typical of me
• Reverse-Scoring Rules: Five items are framed in a reverse orientation to attenuate acquiescence bias. Specifically, Items 1, 3, 8, 10, and 11 must be reverse-scored prior to calculating the composite score:
  • Response 1 becomes 5
  • Response 2 becomes 4
  • Response 3 remains 3
  • Response 4 becomes 2
  • Response 5 becomes 1
• Scoring Mechanics & Range: Following item reversal, all 16 items are summed. Total scores range from a minimum of 16 to a maximum of 80 points.
  • 16 – 39: Low or non-pathological worry (normative control range)
  • 40 – 59: Moderate worry (elevated apprehension, potential subthreshold GAD)
  • 60 – 80: High / Pathological worry (characteristic of clinical Generalized Anxiety Disorder; 62+ is standard diagnostic benchmark)

Permissions & Fee and Test Year

The Penn State Worry Questionnaire was formally published in 1990 in the peer-reviewed journal Behaviour Research and Therapy. Under academic convention and fair-use dissemination standards established by the primary author, Dr. Thomas D. Borkovec, the instrument is generally accessible for non-profit academic research, scientific investigations, and direct clinical practice without the payment of per-use royalty fees, provided the questionnaire is cited accurately using its original 1990 citation.

However, when utilized within commercial settings, large-scale pharmaceutical trials, proprietary digital health platforms, or electronic medical record (EMR) software systems for commercial monetization, licensing verification and explicit publisher permissions may be required through the copyright holder (Elsevier / Behaviour Research and Therapy). Researchers and clinicians wishing to reproduce or translate the questionnaire into other languages must adhere to standard cross-cultural translation back-translation protocols and obtain necessary formal publisher clearances.

References

• Beck, J. G., Stanley, M. A., & Zebb, B. R. (1995). Psychometric properties of the Penn State Worry Questionnaire and the Worry Domains Questionnaire in a clinical anxiety disorder sample. Journal of Psychopathology and Behavioral Assessment, 17(1), 33–46. https://doi.org/10.1007/BF02229201
• Borkovec, T. D. (1994). The nature, functions, and origins of worry. In G. C. L. Davey & F. Tallis (Eds.), Worrying: Perspectives on theory, assessment and treatment (pp. 5–33). John Wiley & Sons.
• Brown, T. A. (2003). Confirmatory factor analysis of the Penn State Worry Questionnaire: Multiple latent factors or method effects? Behaviour Research and Therapy, 41(12), 1411–1426. https://doi.org/10.1016/S0005-7967(03)00059-7
• Brown, T. A., Antony, M. M., & Barlow, D. H. (1992). Psychometric properties of the Penn State Worry Questionnaire in a clinical anxiety disorders sample. Behaviour Research and Therapy, 30(1), 33–37. https://doi.org/10.1016/0005-7967(92)90093-V
• Fresco, D. M., Mennin, D. S., Heimberg, R. G., & Turk, C. L. (2003). Using the Penn State Worry Questionnaire to identify individuals with generalized anxiety disorder: A receiver operating characteristic analysis. Journal of Anxiety Disorders, 17(3), 283–303. https://doi.org/10.1016/S0887-6185(02)00204-6
• Hazlett-Stevens, H., Ullman, J. B., & Craske, M. G. (2004). Factor analysis of the Penn State Worry Questionnaire: Examination of a method effect. Assessment, 11(4), 361–370. https://doi.org/10.1177/1073191104269872
• Meyer, T. J., Miller, M. L., Metzger, R. L., & Borkovec, T. D. (1990). Development and validation of the Penn State Worry Questionnaire. Behaviour Research and Therapy, 28(6), 487–495. https://doi.org/10.1016/0005-7967(90)90135-6
• Newman, M. G., & Llera, S. J. (2011). A novel theory of experiential avoidance in generalized anxiety disorder: A review and synthesis of research supporting a contrast avoidance model of worry. Clinical Psychology Review, 31(3), 371–382. https://doi.org/10.1016/j.cpr.2011.01.008
• Wells, A. (1995). Meta-cognition and worry: A cognitive model of generalized anxiety disorder. Behavioural and Cognitive Psychotherapy, 23(3), 301–320. https://doi.org/10.1017/S1352465800015897

13. Items of the Scale (Questionnaire)