2019 NASSPD Conference Oral Presentation
- 1. Construction & Development of a Self-report Measure of Subjective Emptiness 1Stephanie L. Price, 1Heike I.M. Mahler, 2Christopher J. Hopwood 1California State University San Marcos, 2University of California, Davis
- 2. Acknowledgments Special thanks to the subject matter experts who evaluated the content validity of the Subjective Emptiness Scale (SES) and to Marie Thomas in the Department of Psychology at California State University San Marcos for suggestions regarding the content of this document. This research was supported in part by a grant from the National Institute of General Medical Sciences (NIH/NIGMS).
- 3. Associated with various forms of psychopathology • Depression • Anxiety • Agoraphobia • Schizophrenia • Eating disorders • Personality disorders • Narcissistic • Antisocial • Borderline • greater impairment across the broadest range of psychosocial domains • one of the slowest remitting symptoms
- 4. Significant psychosocial impairement • High risk complications • Self-harm • Suicidality • higher rates of psychiatric hospitalization • Addiction • Sexual compulsivity • Substance use • Daily problems • emotion dysregulation • social dysfunction • problems with identity and self-concept • absenteeism at work • reliance on disability benefits
- 5. Treatment resistant •preliminary evidence supports efficacy of: • electroconvulsive therapy • group schema therapy •lack of efficacy found for: • all pharmacological interventions • most psychotherapeutic approaches
- 6. Limitations of existing measures •single items from instruments designed to measure BPD • less reliable than multi-item scales •multi-item measures • inadequete evidence of construct validity • ostensible construct irrelavent variance
- 7. Mixed method construct validation approach (Leovinger, 1957) Qualitative stage • Phase 1 - definition • item generation • literature review • thematic analysis • interviews • content validity • item reviews by experts and patients
- 8. Phase 1: Inductive + deductive item generation *Borderline, narcissistic, antisocial, schizoid, avoidant, and schizotypal personality disorders Method Personality Disorders* Mood Disorders Schizophrenia/ Psychotic disorders Anxiety Disorders PTSD Literature review X X X X Thematic analysis (e.g. Psych Forums) X X X Phenomenological Interviews (n = 18) X X X X X
- 9. Emptiness conceptualization & table of specifications • INTERNAL: a state of profound hollowness (33%) • EXTERNAL: disconnection from the external world (33%) • PERVASIVE: chronic unfulfillment (33%) Phase 1: Inductive + deductive item generation
- 10. Phase 1: Content validity survey Sample • Subject matter experts (n = 13) • Interviewees from previous study (n = 10) Procedure •Rated each item according to its content relevance (CVR; Lawshe, 1975) •Edit and suggest new items •Qualitative feedback
- 12. Phase 1: Content validity survey Sample • Subject matter experts (n = 13) • Interviewees from previous study (n = 10) Procedure •Rated each item according to its content relevance (CVR; Lawshe, 1975) •Edit and suggest new items •Qualitative feedback Results •Experts (95.45%) and patients (93.18%) reliabily rated most of the items as useful •Inclusion of construct irrelevant content •Better content domain coverage
- 13. Mixed method construct validation approach (Leovinger, 1957) Qualitative stage • Phase 1 - definition • item generation • literature review • thematic analysis • interviews • content validity • item reviews by experts and patients Quantitative Stage • Phase 2 - internal structure • Dimensionality • PCA • CFA/IRT • Homogeneity • Internal consistency
- 14. Phase 2: PCA •Participants: Sample 1 (n = 543) •417 women, 126 men •18 to 46 years, M = 20.22, SD = 3.17 •44.5% were Hispanic/Latino, 26.7% Caucasian, 14.3% Asian/Pacific Islander, 10.8% Other, and 3.1% African-American •Thirty-five (6.4%) reported taking psychiatric medication and 39 participants (7.2%) attended therapy.
- 15. Phase 2: PCA 53-item model • Sample 1 • Kaiser-Meyer Olkin measure of sampling adequacy suggested that sample 1 was factorable (KMO = .97) and Bartlett's test of sphericity was significant (X2 (1,378) = 19,630.09, p < .001) • Loadings from this model ranged from .48 to .83 • First component accounted for 48.63% of the total item covariance 17-item model • Sample 1 • Kaiser-Meyer Olkin measure of sampling adequacy suggested that the data were factorable (KMO = .96) and Bartlett's test of sphericity was significant (X2 (136) = 5,826.63, p < .001) • Loadings from this model ranged from .69 to .84 • First component accounted for 55.53% of the total item covariance
- 16. Phase 2: CFA & Internal consistency • Participants: Sample 2 (n = 1,067) • 716 women, 342 men • 18 to 77 years old (M = 29.78, SD = 11.49) • 81.8% Caucasian, 7.6% Other, 4.1% Hispanic/Latino, 3.2% Asian/Pacific Islander, and 2.2% African-American • 42.3% completed some college, 25.1% possessed a bachelor’s degree, 15.3% earned a high school diploma/G.E.D., 13.2% earned a postgraduate degree, and 3.7% did not graduate from high school • Six-hundred and ninety-three (64.9%) reported taking psychiatric medication and 520 participants (48.7%) attended therapy • Participants: Sample 3 (n = 1,016) • 572 women, 403 men, and 39 other • 18 to 76 years old (M = 27.50, SD = 10.13) • 81.5% Caucasian, 6.9% Other, 5.0% Asian/Pacific Islander, 4.1% Hispanic/Latino, and 2.0% African-American o • 41.4% had completed some college, 22.9% possessed a bachelor’s degree, 18.4% earned a high school diploma/G.E.D., 12.5% earned a postgraduate degree, and 4.3% did not graduate from high school • Five-hundred and three (49.5%) participants reported taking psychiatric medication and 397 participants (39.1%) attended therapy
- 17. Model 1 (Sample 2)
- 18. Model 2 (Sample 2)
- 19. Model 3 (Sample 2) α = .91
- 20. Model 3 replication (Sample 3) α = .93
- 21. •Phase 2: IRT for polytomous items
- 22. •Phase 2: IRT for polytomous items
- 23. Phase 2: IRT for polytomous items
- 24. Phase 2: IRT for polytomous items
- 25. Phase 2: IRT for polytomous items
- 26. Mixed method construct validation approach (Leovinger, 1957) Qualitative stage • Phase 1 - definition • item generation • literature review • thematic analysis • interviews • content validity • item reviews by experts and patients Quantitative Stage • Phase 2 - internal structure • Dimensionality • PCA • CFA/IRT • Homogeneity • Internal consistency • Phase 3 - associations • Extreme groups validity • Convergent and discriminant validity
- 27. Phase 3: Extreme groups validity Figure 2. Student participants scored significantly lower across all seven items than clinical participants in samples two, F (7, 1,575) = 117.20, p < .001; Wilk's Λ = 0.658, partial η2 = .34, and three, F (7, 1,533) = 126.38, p < .001; Wilk's Λ = 0.634, partial η2 = .37
- 28. Phase 3: Convergent & discriminant validity Samples 1 and 2 measures • The Zanarini Rating Scale for Borderline Personality Disorder • Sample 1 α = .82 • Sample 2 α = .83 • The Purpose in Life test short form • Sample 1 α = .83 • Sample 2 α = .85 • The Barratt Impulsiveness Scale–Brief • Sample 1 α = .72 • Sample 2 α = .84 • The Center for Epidemiologic Studies Short Depression Scale • Sample 1 α = .81 • Sample 2 α = .83
- 29. Phase 3: Convergent & discriminant validity Sample 3 measures • The Self-concept and Identity Measure • Internal consistency of the SCIM instrument (α = .92) • identity disturbance subscale (α = .88), • identity consolidation subscale (α = .83) • lack of identity subscale (α = 0.89) • The Personality Inventory for DSM-5 Short Form
- 33. Conclusion Strengths •Large samples •Multiple samples •Rigorous content validity testing •Numerous discriminant and convergent measures Weaknesses •High dropout in clinical samples •Lack of ethnic diversity
- 34. Implications •Brief, unidimensional measure, the Subjective Emptiness Scale (SES) •Useful for future research
- 35. Key Points •The Subjective Emptiness Scale (SES) is brief 7-item self-report measure of emptiness •High internal consistency •Unidimensional •Strong evidence of construct validity • Extreme groups • Discriminant • Convergent
- 36. References Loevinger, J. (1957). Objective tests as instruments of psychological theory. Psychological Reports, 3(3), 635-694.
- 38. Table 4. Correlations between 7-item Subjective Emptiness Scale (SES) and validation measure scores among sample 3. PID-5 Scale Correlation Scale Correlation Anhedonia .80** Irresponsibility .27** Anxiousness .39** Manipulativeness .06 Attention seeking -.07* Perceptual dysregulation .40** Callousness .19** Perseveration .38** Deceitfulness .18** Restricted affectivity .21** Depressivity .80** Rigid perfectionism .24** Distractibility .26** Risk taking .11** Eccentricity .31** Separation insecurity .26** Emotional lability .35** Submissiveness .18** Grandiosity .08* Suspiciousness .43** Hostility .29** Unusual beliefs and experiences .25** Impulsivity .18** Withdrawal .42** Intimacy avoidance .29** Self-Concept and Identity Measure (SCIM) Total .61** Identity disturbance .35** Lack of identity .77** Identity consolidation -.48** Note: * Significant at p = .05; ** significant at p = .01