Certainty About Mental States Questionnaire (CAMSQ)

1. Abstract

The Certainty About Mental States Questionnaire (CAMSQ) is a psychometrically robust, multidimensional self-report measurement tool developed to quantify an individual’s perceived capacity for mentalizing, defined as the reflective mental capacity to perceive, interpret, and infer intentional mental states—including feelings, beliefs, intentions, desires, and goals—in both oneself and other people. Constructed simultaneously in parallel English and German linguistic forms using rigorous consensus-based item selection and iterative back-translation, the instrument originated from an initial pool of 40 potential items that underwent psychometric refinement down to an optimized 20-item final structure. The questionnaire comprises two primary correlated yet distinct subscales: Self-Certainty (10 items), measuring an individual’s subjective clarity and perceived certainty concerning their own internal mental processes, and Other-Certainty (10 items), evaluating perceived accuracy and certainty regarding the inferred mental states, emotions, and intentions of social interaction partners. Additionally, the instrument allows the computational derivation of an Other-Self Discrepancy metric.

Psychometric evaluation across large-scale adult community samples from the United States and Germany demonstrates high structural stability and internal consistency. In empirical investigations, McDonald’s omega coefficient (ω) reached .90 (U.S.) and .88 (Germany) for Self-Certainty, and .91 (U.S.) and .89 (Germany) for Other-Certainty. Two-week test-retest reliability assessments confirmed temporal stability with test-retest coefficients (r_tt) of .85 for Self-Certainty, .78 for Other-Certainty, and .82 for Other-Self Discrepancy. Exploratory and confirmatory factor analyses verified that a two-dimensional oblique model accurately captures the underlying data structure, exhibiting satisfactory fit indices across dynamic cutoff benchmarks. Rigorous multi-group invariance testing established both metric (ΔCFI = −.002) and scalar (ΔCFI = −.010) cross-national invariance across English-speaking American and German-speaking European cohorts. Convergent and discriminant validity were established against established psychometric measures of reflective functioning and social cognition, demonstrating that the CAMSQ provides a sound, cross-culturally validated approach to measuring mentalizing capacity in social, personality, and clinical research contexts.

2. Keywords

Mentalizing, Mental States, Perceived Mentalizing Capacity, Self-Certainty, Other-Certainty, Other-Self Discrepancy, Reflective Functioning, Social Cognition, Psychometrics, Measurement Invariance, Assessment

3. Authors

The Certainty About Mental States Questionnaire (CAMSQ) was conceptualized, operationalized, and psychometrically validated by researchers based at the University of Kassel in Germany:

• Sascha Müller, M.Sc., Ph.D. candidate / Postdoctoral Researcher.
  Affiliation: University of Kassel, Department of Psychology, Kassel, Germany.
  ORCID: 0000-0002-8663-4543.
  Email: [email protected].
  Postal Address: Department of Psychology, Holländische Straße 36-38, 34127 Kassel, Germany.
• Leon P. Wendt, M.Sc., Researcher.
  Affiliation: University of Kassel, Department of Psychology, Kassel, Germany.
  Email: [email protected].
• Johannes Zimmermann, Ph.D., Full Professor of Psychodiagnostics and Psychological Assessment.
  Affiliation: University of Kassel, Department of Psychology, Kassel, Germany.
  ORCID: 0000-0001-6975-2356.

4. Purpose

The primary purpose of the Certainty About Mental States Questionnaire (CAMSQ) is to provide an empirically grounded, cross-culturally equivalent self-report metric designed to assess perceived mentalizing capacities across two primary domains: self-oriented mentalizing and other-oriented mentalizing. Within contemporary clinical psychology, social psychology, and psychiatric diagnostics, mentalizing—frequently operationalized within psychodynamic traditions as reflective functioning—represents a foundational metacognitive and affective skill. It allows human beings to navigate complex social environments by conceptualizing their own and other individuals’ behavior as driven by underlying internal psychological phenomena such as goals, feelings, hopes, doubts, and core values.

Historically, evaluating mentalizing has relied extensively on structured and semi-structured performance interviews, most notably the Adult Attachment Interview (AAI) scored using the Reflective Functioning Scale (RFS), or labor-intensive narrative coding systems. Although psychometrically rich, narrative coding requires extensive training, exhibits inter-rater variability, and presents significant administrative costs that restrict large-scale clinical trials, epidemiological research, and standard psychiatric intake evaluations. Conversely, several existing self-report scales either capture solely a global, undifferentiated construct of mentalization, conflate the motivation to mentalize with perceived competence, or fail to systematically contrast how individuals appraise their internal psychological life versus how they interpret interpersonal cues from others. The CAMSQ was created directly to address these psychometric and conceptual limitations.

From a theoretical and diagnostic standpoint, the CAMSQ enables investigators and clinicians to examine not only overall perceived competence but also the structural balance or imbalance between self-focused certainty and other-focused certainty. In conditions such as borderline personality disorder, narcissistic personality structures, and severe affective dysregulation, individuals often exhibit stark dissociations between introspective self-awareness and social mentalizing. Some patients report feeling completely baffled by their own intense emotional reactions while paradoxically claiming absolute certainty regarding the hostile intentions of others (a phenomenon known in contemporary psychiatric literature as hypermentalizing or pseudomentalizing). In contrast, individuals presenting with alexithymia, avoidant traits, or certain neurodevelopmental profiles may display pervasive hypomentalizing, expressing fundamental uncertainty across both domains.

In clinical practice, the CAMSQ is designed for use in initial psychiatric evaluations, baseline personality disorder profiling under frameworks such as the Alternative Model for Personality Disorders (AMPD) in the DSM-5 and the ICD-11, and treatment progress monitoring within evidence-based modalities like Mentalization-Based Treatment (MBT). In experimental, social, and personality psychology, the CAMSQ facilitates investigations into interpersonal communication dynamics, attachment-related cognitive biases, conflict resolution mechanisms, and social-cognitive changes across adulthood.

5. Psychological Construct

The construct assessed by the CAMSQ is perceived mentalizing capacity, operationalized as an individual’s subjective sense of epistemic certainty regarding internal mental states. Mentalizing is inherently multidimensional, encompassing implicit (automatic, non-conscious) and explicit (controlled, reflective) processing, cognitive versus affective facets, internal versus external foci, and self versus other targets. The CAMSQ specifically targets explicit, declarative beliefs regarding certainty in mental state attribution, structured along the self-other axis.

Self-Certainty

The Self-Certainty subscale (10 items) operationalizes the extent to which an individual experiences their own thoughts, emotional dynamics, physiological-affective reactions, motivations, and desires as identifiable, comprehensible, and internally coherent. High scores reflect epistemic clarity regarding one’s inner world—a person knows why they feel upset, understands what they desire in a given dilemma, and perceives their emotional states as orderly and decipherable. Conversely, low Self-Certainty characterizes states of emotional confusion, pervasive self-alienation, and alexithymic dissociation, wherein affective states appear erratic, unpredictable, and uninterpretable.

Other-Certainty

The Other-Certainty subscale (10 items) captures the extent to which an individual feels confident in inferring, interpreting, and predicting the internal mental states, motivations, feelings, and beliefs of other social agents. A person scoring high on Other-Certainty feels capable of decoding facial expressions, vocal inflections, interpersonal pauses, and contextual cues to ascertain what a peer or romantic partner is undergoing internally. Low Other-Certainty reflects chronic social disorientation, perceived incapacity to read social signals, and profound skepticism about ever truly understanding what others think or feel.

Other-Self Discrepancy

In addition to the two primary dimensions, the CAMSQ framework conceptualizes the algebraic difference between these subscales as Other-Self Discrepancy (computed by subtracting Self-Certainty scores from Other-Certainty scores). This derived construct provides clinically valuable insights into asymmetrical mentalizing profiles. A marked positive discrepancy (Other-Certainty considerably higher than Self-Certainty) often characterizes projective or hypermentalizing pathologies: the person struggles to regulate or comprehend their own identity and emotions, yet projects rigid, unverified certainties onto the minds of others. Conversely, a negative discrepancy (Self-Certainty considerably higher than Other-Certainty) reflects an individual who feels introspectively self-aware but fundamentally disconnected from or bewildered by external social contexts, a profile observed in certain avoidant, schizoid, or neurodivergent populations.

6. Theoretical Framework

The theoretical framework grounding the CAMSQ synthesizes developmental attachment theory, contemporary psychodynamic mentalization theory, and social-cognitive epistemic models. The primary theoretical foundation is rooted in the work of Peter Fonagy, Anthony Bateman, Mary Target, and George Gergely, who formalized the mentalizing approach to developmental psychopathology and psychotherapy.

Fonagy and colleagues argue that mentalizing is not an innate, pre-packaged capacity, but rather a developmental achievement that unfolds within early caregiver-infant attachment interactions. Through contingent, marked parental mirroring, the infant learns that internal affective experiences are separate from physical reality and can be symbolically represented. When caregivers fail to provide accurate, marked mirroring, mentalizing development can become compromised, resulting in non-mentalizing modes of thinking:

• Psychic Equivalence: Internal thoughts and feelings are equated with external reality (e.g., “If I feel you hate me, you objectively do hate me”). Epistemic certainty is absolute and rigid.
• Pretend Mode: Thoughts and mental concepts are decoupled from physical reality to an excessive degree, resulting in intellectualized, empty discourse devoid of emotional contact.
• Teleological Stance: Mental states are acknowledged only when expressed through concrete, physical actions (e.g., a person only believes they are loved if given material gifts or physical acts of reassurance).

The CAMSQ directly operationalizes the degree of subjective certainty an individual adopts when functioning in reflective modes versus non-mentalizing modes. Furthermore, the questionnaire aligns with metacognitive theories of epistemic trust developed by Fonagy and Allison, which examine how individuals evaluate the reliability of information coming from self-awareness versus external social cues. In adaptive mentalizing, an individual displays flexible certainty: possessing sufficient confidence to make sense of self and others, while retaining humility and openness to revisions based on new social feedback. By measuring certainty on a calibrated continuum, the CAMSQ maps both hypo-certainty (epistemic freezing, confusion, and total doubt) and hyper-certainty (epistemic rigidity, dogmatic mind-reading, and hypermentalizing).

7. Validity

The empirical validation of the CAMSQ was conducted across two large, independent adult samples from the United States (N ≈ 800) and Germany (N ≈ 800), providing extensive cross-cultural evidence for its construct, convergent, and discriminant validity.

Convergent Validity

The correlational patterns observed across both the American and German cohorts proved highly uniform, confirming the stability of the nomological network across different languages and cultural settings:

• Self-Certainty Convergence: The Self-Certainty subscale exhibited strong, statistically significant positive correlations with the Mentalization Scale (MentS; Dimitrijević et al., 2018) Self subscale (r ≈ .65 to .72), the Mentalization Questionnaire (MZQ; Hausberg et al., 2012), and established measures of adult reflective functioning. These associations demonstrate that individuals reporting high Self-Certainty consistently demonstrate elevated self-directed emotional awareness and introspective clarity.
• Other-Certainty Convergence: The Other-Certainty subscale demonstrated robust convergent correlations with the MentS Other subscale (r ≈ .60 to .68), which measures an individual’s perceived competence in inferring and attending to others’ thoughts, beliefs, and emotional nuances.

Discriminant Validity

Discriminant validity was established by comparing correlations against theoretically distinct constructs:

• Both CAMSQ subscales (Self-Certainty and Other-Certainty) displayed noticeably weaker, negligible-to-modest correlations with the MentS Motivation to Mentalize dimension (r values typically ranging between .15 and .25). This empirical divergence is theoretically critical, demonstrating that an individual’s desire or motivation to understand mental states is psychometrically distinct from their perceived epistemic certainty and competence in actually decoding those states.
• The scales showed distinct, divergent correlational profiles with standard measures of general distress, neuroses, and cognitive styles, establishing that the CAMSQ does not merely capture general psychological well-being or socially desirable responding.

8. Reliability

The reliability of the CAMSQ has been comprehensively evaluated across multiple empirical indices, demonstrating internal consistency and temporal stability.

Internal Consistency

Because traditional Cronbach’s alpha often assumes tau-equivalence (an assumption rarely met in multidimensional psychometric scales), the authors evaluated internal consistency primarily via McDonald’s omega coefficient (ω), supplemented by standard alpha metrics:

• Self-Certainty Subscale: Demonstrated an omega value of ω = .90 in the United States sample and ω = .88 in the German validation sample.
• Other-Certainty Subscale: Yielded an omega value of ω = .91 in the United States sample and ω = .89 in the German validation sample.

These values demonstrate that both 10-item subscales possess high internal consistency without reaching levels of redundant item formulation (ω > .95).

Test-Retest Reliability

Temporal stability was evaluated in a longitudinal sub-study comprising 100 participants reassessed across a 2-week time interval under uniform testing conditions. Stability coefficients (r_tt) demonstrated consistency over time:

• Self-Certainty: r_tt = .85, indicating high temporal stability for self-directed mental state evaluation.
• Other-Certainty: r_tt = .78, indicating adequate temporal stability while remaining sensitive to fluctuations in social-environmental contexts.
• Other-Self Discrepancy: r_tt = .82, confirming that relative differences between self- and other-mentalizing certainty represent stable trait-like individual differences.

9. Factor Analysis

The structural composition of the CAMSQ was validated using both Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA), complemented by dynamic fit index evaluation and measurement invariance testing.

Exploratory Factor Analysis (EFA)

During preliminary scale development on the initial 40-item candidate pool, exploratory factor analysis with oblique rotation was conducted. The scree plot, parallel analysis, and eigenvalues consistently pointed toward a clear two-factor solution. The factor loadings for retained items cleanly separated into the two theorized dimensions—Self-Certainty and Other-Certainty—with high primary factor loadings (> .55) and negligible cross-loadings (< .20). Ten items per construct were selected to form the final 20-item scale, balancing structural integrity, content breadth, and brevity.

Confirmatory Factor Analysis (CFA)

Confirmatory factor analyses were estimated using robust maximum likelihood estimators to model the two-factor oblique structure across independent U.S. and German cohorts:

• United States Sample: The baseline model fit was χ²(701) = 1,679.77, CFI = .90, RMSEA = .05, SRMR = .04. Using dynamic cutoff criteria tailored to model characteristics (which established cutoffs of CFI ≥ .903, RMSEA ≤ .079, and SRMR ≤ .072), the empirical data demonstrated good model fit, yielding refined fit indices of CFI = .934, RMSEA = .054, and SRMR = .055.
• German Sample: The baseline model fit was χ²(701) = 1,700.46, CFI = .89, RMSEA = .05, SRMR = .04. Tailored dynamic cutoff criteria were established at CFI ≥ .934, RMSEA ≤ .059, and SRMR ≤ .052. The empirical data supported the hypothesized model, producing fit indices of CFI = .935, RMSEA = .049, and SRMR = .052.

Measurement Invariance

To determine whether the CAMSQ functions equivalently across different cultural and linguistic settings, hierarchical multi-group confirmatory factor analysis (MGCFA) was executed testing configural, metric (weak), and scalar (strong) invariance between the U.S. and German cohorts:

• Metric Invariance: Constraining factor loadings across groups yielded a negligible change in comparative fit (ΔCFI = −.002), well within the recognized psychometric cutoff (ΔCFI ≤ .010). This indicates that the latent constructs have identical meaning across both linguistic versions.
• Scalar Invariance: Constraining item intercepts across groups produced a change in comparative fit of ΔCFI = −.010. Achieving scalar invariance confirms that mean differences in latent factor scores between English- and German-speaking respondents can be interpreted meaningfully as true psychological differences rather than measurement artifacts.

10. Instrument / Measurement Tool

• Test Type: Original Multi-Scale Self-Report Psychological Inventory / Questionnaire.
• Item Count: 20 items (10 items on the Self-Certainty subscale, 10 items on the Other-Certainty subscale).
• Response Format: Items are rated using an authentic 7-point Likert-type scale, where:
  • 1 = never
  • 2 = almost never
  • 3 = sometimes
  • 4 = half of the time
  • 5 = often
  • 6 = almost always
  • 7 = always
• Administration Method: Self-administered paper-and-pencil questionnaire or computerized / online psychometric testing platforms.
• Administration Duration: Approximately 5 to 7 minutes.
• Target Population: General and clinical adult populations (aged 18 years and older), encompassing young adulthood (18–29), thirties (30–39), middle age (40–64), and older adulthood (65+).
• Languages Available: English and German (developed simultaneously with verified conceptual and scalar equivalence).
• Scoring Rules:
  • Self-Certainty Scale Score: Mean or sum of the 10 Self-Certainty items (higher scores indicate greater perceived certainty and clarity regarding one’s own internal mental states).
  • Other-Certainty Scale Score: Mean or sum of the 10 Other-Certainty items (higher scores indicate greater perceived certainty and confidence in interpreting the mental states and intentions of other people).
  • Other-Self Discrepancy Index: Computed by subtracting the standardized or raw mean Self-Certainty score from the Other-Certainty score (Other-Certainty − Self-Certainty). Positive scores denote an other-dominant certainty profile (potential hypermentalizing/projection), whereas negative scores reflect a self-dominant certainty profile.

11. Permissions & Fee and Test Year

• Test Publication Year: 2023 (published in the journal Assessment).
• Authors / Copyright Holders: Sascha Müller, Leon P. Wendt, and Johannes Zimmermann.
• Commercial Use: Non-commercial research and educational use only. Commercial exploitation, distribution within commercial software packages, or fee-charging clinical diagnostics requires prior explicit written permission.
• Fee: Free of charge for non-commercial academic research, empirical theses, and educational purposes.
• Permissions / Licensing: Researchers and clinicians wishing to use, reproduce, or adapt the CAMSQ must contact the corresponding author, Sascha Müller, via email ([email protected]) or consult the original publication in Assessment. The full set of 20 items in both English and German is published in the Appendix (page 670) of the original open/licensed source paper.

12. References

• Dimitrijević, A., Hanak, N., Altaras Dimitrijević, A., & Jolić Marjanović, Z. (2018). The Mentalization Scale (MentS): A self-report measure for the assessment of mentalizing capacity. Journal of Personality Assessment, 100(3), 268–280. https://doi.org/10.1080/00223891.2017.1310730
• Fonagy, P., Gergely, G., Jurist, E. L., & Target, M. (2002). Affect regulation, mentalization, and the development of the self. Other Press.
• Hausberg, M. C., Schulz, H., Pundt, I., Feldmann, M., Völlm, B. A., & Fabritz, R. (2012). Development and validation of the Mentalization Questionnaire (MZQ). Zeitschrift für Psychosomatische Medizin und Psychotherapie, 58(4), 330–340. https://doi.org/10.13109/zptm.2012.58.4.330
• Müller, S., Wendt, L. P., & Zimmermann, J. (2023). Development and validation of the Certainty About Mental States Questionnaire (CAMSQ): A self-report measure of mentalizing oneself and others. Assessment, 30(3), 651–674. https://doi.org/10.1177/10731911211061280

13. Items of the Scale