Personal Feelings Questionnaire-2 (PFQ-2)

Abstract

The Personal Feelings Questionnaire-2 (PFQ-2) is a widely utilized psychometric instrument developed by David W. Harder and S. J. Lewis (1987) designed to assess continuous trait dispositions toward the two fundamental self-conscious moral emotions: shame proneness and guilt proneness. Originating as a revision of the original Personal Feelings Questionnaire (Harder & Ritzler, 1979), the PFQ-2 addresses the psychometric necessity of distinguishing between shame—an acute, global devaluation of the self—and guilt—a dysphoric tension centered on a specific behavioral transgression. The instrument consists of 22 self-report items comprising 10 shame-related affective descriptors, 6 guilt-related affective descriptors, and 6 filler items measuring other emotional states (e.g., sadness, depression, euphoria, enjoyment) to mask the underlying testing objective and reduce response bias.

Respondents evaluate the intensity or frequency with which they experience each emotional state using a 5-point Likert-type scale ranging from 0 (“I do not experience the feeling”) to 4 (“I experience the feeling very strongly”). Extensive psychometric evaluations across clinical and non-clinical cohorts have demonstrated sound internal consistency, with Cronbach’s alpha coefficients typically ranging from .76 to .82 for the Shame subscale and .68 to .78 for the Guilt subscale. Test-retest reliability across multi-week intervals has demonstrated strong temporal stability (typically exceeding .75). Factor analytic investigations confirm a robust two-factor structure delineating shame and guilt, while construct validity studies highlight shame’s potent associations with internalizing psychopathology, depression, social anxiety, and low self-esteem, compared to guilt’s more differentiated relationship with moral concern, empathy, and constructive interpersonal repair.

Keywords

Personal Feelings Questionnaire-2, PFQ-2, shame proneness, guilt proneness, self-conscious emotions, moral affect, psychometrics, affective assessment, Harder, self-concept, depressive vulnerability, personality assessment.

Authors

The Personal Feelings Questionnaire-2 was developed by David W. Harder and S. J. Lewis in 1987 as part of empirical research programs evaluating self-conscious emotions at Tufts University and affiliated research institutions. Dr. David W. Harder is an American clinical psychologist and academic affiliated with Tufts University (Department of Psychology) and McLean Hospital/Harvard Medical School, whose research focused heavily on psychodynamics, object relations, ego development, and the distinct clinical manifestations of shame versus guilt. Dr. S. J. Lewis collaborated with Harder during the late 1970s and 1980s to refine the psychometric properties of trait affect checklists. Subsequent key psychometric validation studies on the PFQ-2 were led by David W. Harder in collaboration with colleagues including Leslie Rockart and Lisa Cutler (1993).

Purpose

The fundamental purpose of the Personal Feelings Questionnaire-2 is to provide a brief, psychometrically sound, and phenomenologically direct measure of an individual’s susceptibility to experiencing shame and guilt as personality traits. In both theoretical psychology and empirical clinical research, distinguishing shame from guilt was historically impeded by conceptual conflation and semantic ambiguity; laypersons and researchers alike frequently used the terms interchangeably despite their divergent phenomenological, psychological, and behavioral trajectories.

The PFQ-2 addresses this problem by avoiding hypothetical vignettes or situational narratives (such as those later popularized by the Test of Self-Conscious Affect [TOSCA]), relying instead on a direct affective adjective rating format. By presenting phenomenologically grounded affective markers, the PFQ-2 captures the raw subjective frequency and intensity with which individuals encounter these emotional states in their day-to-day lives. This design makes the instrument particularly well-suited for research into personality traits, psychopathology, self-concept vulnerabilities, and emotional regulation.

Clinically, the PFQ-2 serves as an essential assessment tool for diagnosing affective vulnerabilities across a spectrum of psychiatric disorders. Elevated shame proneness has been identified as an underlying transdiagnostic vulnerability factor in major depressive disorder, generalized social phobia, borderline personality disorder, narcissistic personality disorder (especially covert or vulnerable narcissism), post-traumatic stress disorder (PTSD), and substance use disorders. In substance abuse treatment—such as alcohol misuse rehabilitation—the PFQ-2 allows clinicians to identify whether self-destructive consumption patterns are driven by debilitating shame (global feelings of worthlessness and defectiveness) or remorseful guilt, which can be channeled toward therapeutic amends and self-forgiveness.

Psychological Construct

The psychological constructs measured by the PFQ-2 rest upon the empirical taxonomy of self-conscious moral emotions, which require self-reflection, self-awareness, and self-evaluation against internalized societal, cultural, or personal standards, ideals, and rules.

The Shame Construct

Within the PFQ-2 framework, shame is conceptualized as an acute, deeply painful, and debilitating negative evaluation focused on the entire global self. When experiencing shame, the individual perceives the core self as inherently flawed, deficient, laughable, ridiculous, or disgusting. The phenomenological experience is characterized by intense self-consciousness, feelings of exposure, autonomic blushing, and an overwhelming impulse to hide, withdraw, or disappear. The 10 shame items on the PFQ-2 capture these somatic and psychological dimensions:

• Global Defectiveness and Cognitive Paralysis: Captured by items such as Feeling “stupid” (Item 10) and Feeling helpless, paralyzed (Item 14). These reflect the cognitive disruptions and feelings of operational incapacity that accompany profound shame.
• Ridicule and Social Exposure: Captured by Embarrassment (Item 1), Feeling ridiculous (Item 3), Self-consciousness (Item 6), and Feeling laughable (Item 18). These capture the agonizing awareness of being perceived negatively by an imagined or real observing audience.
• Interpersonal Degradation and Somatosensory Markers: Captured by Feeling humiliated (Item 7), Feeling “childish” (Item 12), Feelings of blushing (Item 16), and Feeling disgusting to others (Item 21). These reflect visceral self-loathing, physical signs of shame, and perceived total loss of social status.

The Guilt Construct

In contrast, guilt is conceptualized as a dysphoric emotional tension centered upon a specific behavior, action, or failure to act, rather than the core self. In guilt, the self remains structurally intact; the focus of condemnation is directed outward toward the transgression and its consequences for others. Phenomenologically, guilt involves feelings of remorse, tension, worry over harm inflicted, and a motivational press toward reparative actions, confession, apology, and making amends. The 6 guilt items on the PFQ-2 represent this construct:

• Affective Tension and Moral Distress: Reflected in Mild guilt (Item 2) and Intense guilt (Item 8), assessing the perceived intensity of the pure emotional affect.
• Interpersonal Concern and Empathy: Reflected in Worry about hurting or injuring someone (Item 4), highlighting the empathic and other-oriented nature of adaptive guilt.
• Cognitive Accountability and Reparative Drive: Reflected in Regret (Item 11), Feeling you deserve criticism for what you did (Item 17), and Remorse (Item 22), capturing the conscious recognition of behavioral culpability.

Filler and Masking Affects

The PFQ-2 deliberately includes 6 filler items—Sadness (Item 5), Euphoria (Item 9), Mild happiness (Item 13), Depression (Item 15), Rage (Item 19), and Enjoyment (Item 20). These items are not calculated into either the shame or guilt scale scores. They serve primarily to prevent pattern response sets, reduce demand characteristics, and mask the specific moral focus of the instrument.

Theoretical Framework

The theoretical bedrock of the PFQ-2 is rooted in psychodynamic theory and modern affective cognitive science, most notably the seminal conceptual framework formulated by psychoanalyst Helen Block Lewis in her landmark 1971 work, Shame and Guilt in Neurosis. Lewis postulated that although both affects are moral emotions triggered by self-evaluation, their cognitive structural organization, ego involvement, and relational dynamics are fundamentally different.

Helen Block Lewis’s Structural Distinction

Lewis asserted that in shame, the self is both the subject and the object of judgment: “I (subject) feel ashamed of who I am (object).” The ego is fragmented, overwhelmed by feelings of worthlessness, inferiority, and exposure before an idealized or condemning other. Consequently, shame elicits passive, defensive behaviors: shrinking, hiding, avoidance, or retaliatory rage (what Lewis called “shame-rage spirals”). Conversely, in guilt, the ego remains functioning and unthreatened in its core identity; instead, judgment is focused onto an act: “I (subject) committed an immoral or destructive deed (act).” The focus remains on the violation of internalized standards (the superego), generating tension that motivates corrective action.

Integration with Psychodynamic and Personality Theory

Harder and Lewis (1987) translated Lewis’s qualitative psychodynamic insights into quantitative measurement. Prior psychodynamic theorists (e.g., Sigmund Freud, Erik Erikson, Heinz Kohut) had often prioritized guilt (linking it to the Oedipus complex, conscience, and castration anxiety) while marginalizing shame as an infantile precursor. However, Kohut’s self psychology later situated shame at the center of narcissistic injury—the painful deflation that occurs when the grandiose self meets cold, unresponsive environmental mirrors. Harder utilized these formulations to argue that trait shame represents structural ego weakness and susceptibility to narcissistic injury, whereas trait guilt reflects superego vigilance.

Cognitive Attributional Parallels

The theoretical underpinnings of the PFQ-2 also parallel cognitive attribution theory (e.g., Bernard Weiner). Shame is driven by internal, stable, and global attributions for negative events (“I failed because I am inherently flawed”), which leads to hopelessness and withdrawal. Guilt stems from internal, unstable, and specific attributions (“I failed because I did not prepare adequately”), which preserves self-worth while promoting corrective and reparative behavioral adjustments.

Validity

The construct, convergent, and discriminant validity of the PFQ-2 has been examined across numerous empirical investigations involving college student cohorts, community samples, and diverse psychiatric patient populations (Harder, 1990; Harder, Rockart, & Cutler, 1993).

Convergent Validity

Studies evaluating the PFQ-2 have consistently demonstrated high convergent validity between the Shame subscale and measures of internalizing psychopathology, psychological distress, and neuroticism. In validation studies by Harder et al. (1993), the PFQ-2 Shame scale demonstrated significant positive correlations with:

• Depressive Symptomatology: Correlating strongly with the Beck Depression Inventory (BDI) (r values typically ranging from .45 to .62).
• Symptom Distress: Exhibiting strong positive correlations with the Global Severity Index (GSI) of the Symptom Checklist-90-Revised (SCL-90-R) (r ≈ .50 to .65).
• Social Anxiety and Shyness: Showing robust relationships with measures of social evaluation anxiety and fear of negative evaluation (r ≈ .40 to .58).
• Borderline and Vulnerable Narcissistic Personality Organization: Demonstrating substantial associations with borderline pathology and identity diffusion.

The PFQ-2 Guilt scale similarly demonstrates convergent validity with validated moral and personality inventories, exhibiting moderate positive correlations with measures of interpersonal concern, empathy, internalized standards, and conscientiousness.

Discriminant Validity

Discriminant validity is a critical area of empirical focus in self-conscious emotion research. Because shame and guilt are both negative moral affects, self-report measures of trait shame and trait guilt frequently demonstrate moderate intercorrelations (typically r = .40 to .55). However, when partial correlations or multiple regressions are applied to control for the shared affective variance:

• Self-Esteem: PFQ-2 Shame correlates strongly and negatively with global self-esteem as measured by the Rosenberg Self-Esteem Scale (zero-order r ≈ -.50 to -.65; partial r remains strongly negative), whereas PFQ-2 Guilt, once shame is partialled out, typically exhibits negligible or even slightly positive relationships with self-esteem.
• Assertiveness and Anger Expression: Shame correlates inversely with assertiveness (e.g., the Rathus Assertiveness Schedule) and positively with unexpressed, suppressed resentment or indirect hostility. Guilt, when decoupled from shame, is unrelated to passive-aggressive or explosive hostility.
• Substance Abuse and Forgiveness: Research evaluating addiction populations (e.g., Scherer, 2010) reveals that high PFQ-2 Shame scores predict chronic relapse, avoidance coping, and inability to self-forgive, whereas PFQ-2 Guilt is predictive of constructive remorse, therapeutic engagement, and successful self-forgiveness interventions.

Reliability

The reliability of the Personal Feelings Questionnaire-2 has been substantiated through rigorous assessments of internal consistency (Cronbach’s alpha) and temporal stability (test-retest reliability).

Internal Consistency

In the primary normative and validation samples reported by Harder and Lewis (1987) and Harder (1990):

• Shame Subscale (10 items): Cronbach’s alpha coefficients consistently fall within the range of .76 to .82 in university undergraduate samples, and frequently reach .80 to .85 in clinical cohorts. These coefficients reflect strong homogeneity among the shame descriptors.
• Guilt Subscale (6 items): Cronbach’s alpha coefficients typically range from .68 to .78. While slightly lower than the Shame subscale, this range is deemed psychometrically acceptable given the brevity of the subscale (only 6 items) and the nuanced breadth of guilt manifestations (spanning cognitive remorse to empathetic worry).

Test-Retest Stability

Because the PFQ-2 is intended to measure proneness (dispositional trait susceptibility) rather than transient emotional state fluctuations, temporal stability is essential. Longitudinal testing by Harder, Rockart, and Cutler (1993) across a two-week interval yielded test-retest correlation coefficients of r = .84 (p < .001) for the Shame subscale and r = .74 (p < .001) for the Guilt subscale. Across longer test-retest windows ranging from 5 to 7 weeks, the stability coefficients remained solid (Shame r ≈ .75; Guilt r ≈ .69), demonstrating that the instrument successfully indexes enduring personality traits rather than temporary mood shifts.

Factor Analysis

The factorial validity of the PFQ-2 has been evaluated via both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) across diverse samples.

Exploratory Factor Analysis (EFA)

Early psychometric analyses conducted by Harder and colleagues utilized principal components analysis with varimax and oblimin rotations on the 16 target items (excluding the 6 filler items). These analyses repeatedly confirmed a clear two-factor solution corresponding directly to the theoretical constructs of Shame and Guilt:

• Factor 1 (Shame Proneness): Accounts for the largest percentage of common variance (typically between 28% and 35%). Items such as Feeling humiliated, Feeling “stupid”, Feeling ridiculous, Self-consciousness, and Feeling laughable exhibit high factor loadings (typically ranging from .55 to .78) on this dimension, with minimal cross-loadings on the Guilt factor.
• Factor 2 (Guilt Proneness): Accounts for an additional 12% to 18% of the common variance. Items such as Remorse, Intense guilt, Mild guilt, and Feeling you deserve criticism for what you did demonstrate strong primary loadings (.50 to .75) on this factor.

Confirmatory Factor Analysis (CFA) and Model Fit

Subsequent structural equation modeling and confirmatory factor analytic studies have formally compared a one-factor model (general negative moral affect) against a two-factor correlated model (distinct Shame and Guilt dimensions). The two-factor model consistently shows superior fit to the empirical data:

• Chi-Square / Degrees of Freedom Ratio (χ²/df): Typically < 2.5, indicating good model parsimony.
• Comparative Fit Index (CFI): Generally exceeds .90 to .93 in non-clinical samples.
• Root Mean Square Error of Approximation (RMSEA): Ranges between .05 and .07, confirming an acceptable-to-close fit.

While the Shame and Guilt factors remain moderately correlated (inter-factor correlations typically range between .42 and .56), attempts to collapse all items into a single undifferentiated construct result in significant deterioration of model fit indices, affirming that shame and guilt are distinct constructs.

Instrument / Measurement Tool

• Test Type: Self-report psychological questionnaire / trait affect rating scale.
• Format: Continuous adjective/descriptor rating scale.
• Administration Time: Approximately 5 to 10 minutes.
• Target Population: Adolescents and adults (ages 16 and older); adaptable for clinical, organizational, and research environments.
• Total Item Count: 22 items total.
  • Shame Subscale (10 items): Items 1, 3, 6, 7, 10, 12, 14, 16, 18, 21.
  • Guilt Subscale (6 items): Items 2, 4, 8, 11, 17, 22.
  • Filler / Masking Items (6 items): Items 5, 9, 13, 15, 19, 20.
• Response Format: 5-point Likert-type scale scored as:
  • 0 = I do not experience the feeling
  • 1 = I experience the feeling a little bit
  • 2 = I experience the feeling moderately
  • 3 = I experience the feeling strongly
  • 4 = I experience the feeling very strongly
• Scoring Guidelines:
  • Shame Score: Calculated by summing the ratings of the 10 shame items (score range: 0 to 40) or by computing the mean item score (range: 0.0 to 4.0). Higher scores indicate greater shame proneness.
  • Guilt Score: Calculated by summing the ratings of the 6 guilt items (score range: 0 to 24) or by computing the mean item score (range: 0.0 to 4.0). Higher scores reflect greater guilt proneness.
  • Filler Items: Items 5, 9, 13, 15, 19, and 20 are omitted from all subscale scoring.
  • Reverse Scoring: There are no reverse-coded items on the PFQ-2.

Permissions & Fee and Test Year

The Personal Feelings Questionnaire-2 was officially published in 1987 by David W. Harder and S. J. Lewis in the book chapter titled “The Assessment of Shame and Guilt” within Volume 2 of the academic serial Advances in Personality Assessment (edited by C. D. Spielberger and J. N. Butcher, published by Lawrence Erlbaum Associates).

The PFQ-2 is generally available for non-commercial academic, psychological research, and clinical educational purposes without licensing fees, provided that appropriate scholarly attribution is given to the original authors. Formal copyright resides with the authors and the original publisher. Researchers planning large-scale institutional, clinical trial, or commercial applications are advised to consult the primary developer (David W. Harder) or refer to rights holders for permissions.

References

• Harder, D. W. (1990). Additional construct validity evidence for the Harder Personal Feelings Questionnaire measure of shame and guilt proneness. Psychological Reports, 67(1), 288–290. https://doi.org/10.2466/pr0.1990.67.1.288
• Harder, D. W., & Lewis, S. J. (1987). The assessment of shame and guilt. In C. D. Spielberger & J. N. Butcher (Eds.), Advances in personality assessment (Vol. 6, pp. 89–114). Lawrence Erlbaum Associates.
• Harder, D. W., & Ritzler, V. L. (1979). The Personal Feelings Questionnaire. Unpublished manuscript, Tufts University, Medford, MA.
• Harder, D. W., Rockart, L., & Cutler, L. (1993). Additional validity evidence for the Harder Personal Feelings Questionnaire-2 (PFQ2): A measure of shame and guilt proneness. Journal of Clinical Psychology, 49(4), 584–604. http://scholarscompass.vcu.edu/cgi/viewcontent.cgi?article=3108&context=etd
• Tangney, J. P., Wagner, P., & Gramzow, R. (1992). Proneness to shame, proneness to guilt, and psychopathology. Journal of Abnormal Psychology, 101(3), 469–478. https://doi.org/10.1037/0021-843X.101.3.469

13. Items of the Scale (Questionnaire)