Bifactor Model Research Articles

Psychometric Properties of the Modified Scale for Suicidal Ideation-Self-Report–Revised With High-Risk Adolescents

In treatment-seeking samples, approximately 40% of adolescents with suicidal ideation transition to engaging in suicidal behavior. Thus, assessment of the severity of suicidal ideation during care is critically important. However, few free, self-report, evidence-based measures that assess for severity of suicidal ideation exist. To address this need, the present study examined several psychometric properties of a revised version of the Modified Scale for Suicidal Ideation–Self Report (MSSI-SR-R) with adolescents. A sample of 284 adolescents (Mage = 15.30; SD = 1.42; range = 12–18; 62.7% White; 75.7% Non-Hispanic/Latinx) enrolled in a partial hospital program (PHP) completed the MSSI-SR-R at the beginning of treatment. The factor structure of the MSSI-SR-R was assessed using bifactor exploratory structural equation modeling. Measurement invariance across sexual identity was also examined. Internal consistency was calculated using omega. Validity was assessed by comparing scores on the MSSI-SR-R to well-validated self-report measures of suicidal thought and behaviors as well as depressive symptoms. Results supported a bifactor model with a general factor and two specific factors (i.e., Desire/Ideation and Plans/Preparation), which was invariant across sexual identity. The MSSI-SR-R total score demonstrated good reliability and was positively associated with measures of suicidal ideation and behaviors as well as depressive symptom severity. Though additional psychometric work is needed, the MSSI-SR-R demonstrated good psychometric properties. It may offer utility in clinical settings as a brief, free, and validated assessment of severity of suicidal ideation.

Psychometric Properties of the Chinese Version of the Core Symptom Index: A Study among Chinese Parents of Children with Autistic Spectrum Disorders.

(1) Background: Parents of children with autism spectrum disorders often experience psychological distress, which can affect the quality of childcare they provide. It is crucial to screen for psychiatric symptoms among these parents. The core symptom index (CSI) is a widely recognized tool used to assess general symptoms, including depression, anxiety, and somatic issues. It has proven validity and reliability across diverse Thai populations. Given the cultural similarities between Thai and Chinese populations, the CSI has been successfully implemented within the Chinese population. Nevertheless, it is crucial to research its validity and reliability in the general Chinese population. This study aimed to investigate the psychometric properties of the Chinese version of the CSI among parents of children with autism spectrum disorders using confirmatory factor analysis (CFA). (2) Methods: A total of 794 Chinese parents raising children with autism participated in this study. All completed the CSI, along with the social inhibition subscale of the Interpersonal Problems Inventory and the Couple Satisfaction Index. Factorial validity was assessed using CFA to determine how well the bifactor three-factor model fits the data. Various structural models were compared using model fit indices. Convergent and discriminant validity were examined by exploring correlations with the social inhibition subscale and the Couple Satisfaction Index. Invariance testing of the CSI was conducted across multiple groups based on gender, age, and education using CFA. The reliability of the CSI was evaluated using McDonald's omega coefficients. (3) Results: The bifactor model emerged as the best-fitting model for the data, suggesting that the total score of the CSI adequately represents overall psychiatric symptoms. The CSI exhibited significant correlations with the social inhibition subscale (r = 0.41, p < 0.01) and smaller correlation coefficients with the Couple Satisfaction Index (r = -0.16, p < 0.05), indicating both convergent and discriminant validity. The invariant test results support scalar invariance levels based on gender and age but only partial invariance for education. The Chinese version of the CSI demonstrated high consistency, with McDonald's omega coefficients ranging between 0.86 and 0.95. (4) Conclusions: The bifactor model of the Chinese version of the CSI is validated, making it a suitable tool for measuring depression, anxiety, and somatization symptoms among parent(s) of children with autism spectrum disorders. Further research on other Chinese populations is encouraged.

Prevalence of celebrity worship: Development and application of the short version of the Celebrity Attitude Scale (CAS-7) on a large-scale representative sample.

Celebrity worship, defined as an excessive admiration towards celebrities, has generated considerable research and public interest. A widely used assessment instrument to measure celebrity worship is the 23-item Celebrity Attitude Scale. However, concerns have been raised regarding the measurement, including the inconsistent factor structure and lack of a cut-off point to identify "celebrity worshipers". The present study aims to address these concerns by testing the psychometric appropriateness of a short, 7-item version of the CAS (i.e., CAS-7) and estimating the prevalence of individuals with high-level celebrity admiration using a representative sample of Hungarian adults (between 18 and 64 years of age) according to gender, age, geographic location, and size of residence. The total sample comprised 2028 respondents, of which 769 valid responses were administered from participants who reported having a favorite celebrity (51.11% men, Mage = 36.38 years, SD = 13.36). Results indicated an excellent model fit for the two-factor and bifactor model of the CAS-7. Based on the suggested cut-off score of 26, the prevalence of high-level celebrity admiration is 4.53% in the Hungarian adult population (18-64 years of age) and 8.51% among young adults (18-34 years of age). Individuals with this high level of admiration towards a favorite celebrity reported more symptoms of problematic Internet use, depression, anxiety, and stress than individuals with general celebrity admiration levels. TheCAS-7 demonstrated sound psychometric properties, confirming its applicability in research and practice.

Can a global score be derived from the Integration of Stressful Life Experiences Scale-Short Form (ISLES-SF)? Empirical support for a bi-factor model

The Integration of Stressful Life Experiences Scale-Short Form (ISLES-SF) is a popular, two-factor measure of meaning-making. Although researchers have used this instrument to calculate a global index of meaning-making, there has been little evidence to support this practice. Therefore, the objective of this study was to examine the validity of this approach by analyzing data from 2,380 American adults during the beginning of the COVID-19 pandemic. A series of confirmatory factor analyses, reliability analyses, and concurrent validity analyses support a bi-factor model of meaning-making. These results affirm the practice of not only using the ISLES-SF to measure meaning-making at a global level, but also to interpret its two specific dimensions (Comprehensibility and Footing in the World) as well.

Accommodating and Extending Various Models for Special Effects Within the Generalized Partially Confirmatory Factor Analysis Framework.

Special measurement effects including the method and testlet effects are common issues in educational and psychological measurement. They are typically covered by various bifactor models or models for the multiple traits multiple methods (MTMM) structure for continuous data and by various testlet effect models for categorical data. However, existing models have some limitations in accommodating different type of effects. With slight modification, the generalized partially confirmatory factor analysis (GPCFA) framework can flexibly accommodate special effects for continuous and categorical cases with added benefits. Various bifactor, MTMM and testlet effect models can be linked to different variants of the revised GPCFA model. Compared to existing approaches, GPCFA offers multidimensionality for both the general and effect factors (or traits) and can address local dependence, mixed-type formats, and missingness jointly. Moreover, the partially confirmatory approach allows for regularization of the loading patterns, resulting in a simpler structure in both the general and special parts. We also provide a subroutine to compute the equivalent effect size. Simulation studies and real-data examples are used to demonstrate the performance and usefulness of the proposed approach under different situations.

Psychometric Characteristics of the Italian Version of the Revised Sociosexual Orientation Inventory

Sociosexuality refers to the tendency to engage in uncommitted sexual behavior and has been dissected into three domains: sociosexual behavior, attitudes, and desire (Penke & Asendorpf, 2008), which led to the revised Sociosexual Orientation Inventory (SOI-R), which was validated on a German sample. The current research aimed at translating and validating an Italian version (I-SOI-R), administered to three distinct Italian participant groups. In the first sample (N = 710, females = 521, age = 18–59 years), we found evidence for a bifactor model, articulated in a general sociosexuality factor and three specific factors (behavior, attitudes, desire). High internal consistency was established for total and subscale scores, alongside favorable test–retest reliability. A connection was found between relationship status and sociosexual desire, though not gender dependent. We found evidence for test–retest reliability in a second sample (N = 55, females = 37, age 20–58 years). In a third study (N = 305, females = 147, age = 19–60 years), the earlier findings were replicated, further confirming the I-SOI-R's construct, criterion, and nomological validity on an online sample. Combining data from the three studies revealed full configural, metric, and scalar invariance regarding gender. This allowed us to meaningfully compare the observed scores of women and men and replicated the finding that men display higher levels of unrestricted sociosexuality. In conclusion, the I-SOI-R may serve as a valuable tool to assess and enhance sexual health, albeit warranting future research on construct and criterion validity.

P-Factor(s) for Youth Psychopathology Across Informants and Models in 24 Societies

ABSTRACT Objective Although the significance of the general factor of psychopathology (p) is being increasingly recognized, it remains unclear how to best operationalize and measure p. To test variations in the operationalizations of p and make practical recommendations for its assessment, we compared p-factor scores derived from four models. Methods We compared p scores derived from principal axis (Model 1), hierarchical factor (Model 2), and bifactor (Model 3) analyses, plus a Total Problem score (sum of unit-weighted ratings of all problem items; Model 4) for parent- and self-rated youth psychopathology from 24 societies. Separately for each sample, we fitted the models to parent-ratings on the Child Behavior Checklist for Ages 6–18 (CBCL/6–18) and self-ratings on the Youth Self-Report (YSR) for 25,643 11–18-year-olds. Separately for each sample, we computed correlations between p-scores obtained for each pair of models, cross-informant correlations between p-scores for each model, and Q-correlations between mean item x p-score correlations for each pair of models. Results Results were similar for all models, as indicated by correlations of .973–.994 between p-scores for Models 1–4, plus similar cross-informant correlations between CBCL/6–18 and YSR Model 1–4 p-scores. Item x p correlations had similar rank orders between Models 1–4, as indicated by Q correlations of .957–.993. Conclusions The similar results obtained for Models 1–4 argue for using the simplest model – the unit-weighted Total Problem score – to measure p for clinical and research assessment of youth psychopathology. Practical methods for measuring p may advance the field toward transdiagnostic patterns of problems.

General psychopathology factor in Chinese adolescents and its correlation with trans-diagnostic protective psycho-social factors

BackgroundComorbidity in mental disorders is prevalent among adolescents, with evidence suggesting a general psychopathology factor (“p” factor) that reflects shared mechanisms across different disorders. However, the association between the “p” factor and protective factors remains understudied. The current study aimed to explore the “p” factor, and its associations with psycho-social functioning, in Chinese adolescents. Methods2052 students, aged 9–17, were recruited from primary and secondary schools in Shanghai, China. Multiple rating scales were used to assess psychological symptoms and psycho-social functioning. Confirmatory factor analysis was conducted to verify the fit of models involving different psychopathology domains such as externalizing, internalizing, and the “p” factor. Subsequently, structural equation models were used to explore associations between the extracted factors and psycho-social functioning, including emotion regulation, mindful attention awareness, self-esteem, self-efficacy, resilience, and perceived support. ResultsThe bi-factor model demonstrated a good fit, with a “p” factor accounting for 46 % of symptom variation, indicating that the psychological symptoms of Chinese adolescents could be explained by internalizing, externalizing, and the “p” factor. Psychologically, a higher “p” was positively correlated with emotion suppression and negatively correlated with mindful attention awareness, emotion reappraisal, self-esteem, and resilience. Socially, a higher “p” was associated with decreased perceived support. LimitationsOnly common symptoms were included as this study was conducted at school. Furthermore, the cross-sectional design limited our ability to investigate causal relationships. ConclusionsA “p” factor exists among Chinese adolescents. Individuals with higher “p” factor levels were prone to experience lower levels of psycho-social functions.

The Fear of COVID-19 Scale: Psychometric properties and measurement invariance across 10 months

The coronavirus disease 2019 (COVID-19) has led to various negative consequences including fear. The Fear of COVID-19 Scale (FCV-19S) has been widely used in diverse cultures, but no study has ever investigated its longitudinal measurement invariance and predictive validity. Therefore, we examined its longitudinal measurement invariance and predictive validity over 10 months. A sample of Chinese undergraduates (N = 682; first wave 842; 682 second wave) completed the FCV-19S as well as measures assessing depression, anxiety, and stress. Exploratory and confirmatory factor analyses were conducted along with measurement invariance testing. The results showed that the bifactor model fitted well, and significantly predicted stress and anxiety, but not depression. The FCV-19S demonstrated partial measurement invariance (i.e. configural and metric invariances) across time. These findings suggest that the Chinese version of FCV-19S is a reliable tool and could be used in evaluating the severity of fear of COVID-19 among Chinese young adults.

Unidimensional versus multidimensional: A bifactor factor structure of the Self-Stigma Scale-Short (SSS-S) among U.S. adults with psychiatric disabilities.

This study aimed to examine the psychometric properties of the English version of the Self-Stigma Scale-Short (SSS-S), a nine-item self-report self-stigma measurement, among U.S. adults with psychiatric disabilities. We obtained reliability and validity evidence from a sample of 275 adults with psychiatric disabilities. Exploratory factor analysis (EFA; n = 139) yielded a two-factor solution that accounts for 64.97% of the variance. Confirmatory factor analysis (CFA; n = 136) was conducted to compare alternative solutions, including a single-factor model, a two-correlated-factor model, a three-correlated-factor model, and a bifactorial model. The CFA results supported the bifactor S·I - 1 model as a superior latent factor structure for the SSS-S. The coefficient ω of the SSS-S was .94, indicating excellent internal reliability. Concurrent validity of the SSS-S was supported by significant positive correlations with societal stigma and psychiatric symptom severity, and negative correlations with psychiatric disability acceptance, general self-efficacy, and hope. The overall findings concluded that the English version of the SSS-S demonstrated reliable and valid scores and a primarily unidimensional structure of self-stigma among U.S. adults with psychiatric disabilities. Given the adverse impact of self-stigma and its relationships with recovery-related constructs shown in this study, the routine use of the SSS-S is recommended in psychiatric rehabilitation settings. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Competing Confirmatory Factor Analysis Models in Management Research: Bifactor Modeling of the Employee Work Assessment Tool

Competing Confirmatory Factor Analysis Models in Management Research: Bifactor Modeling of the Employee Work Assessment Tool

Do measures of social support and social distress share general factors associated with suicidal ideation and attempts?

Aspects of social relationships have variably been associated with suicidal ideation (SI) and suicide attempts (SAs). This study assessed whether social support and social distress measures have general factors versus measure-specific factors that are associated with suicide risk. Adults (N = 455, 60.0% female), admitted to psychiatric inpatient units following a recent suicide attempt or active SI, completed assessments of social support (emotional support, instrumental support, friendship, perceived support from significant others, friends, family) and social distress (loneliness, perceived rejection, perceived burdensomeness, thwarted belongingness). Bifactor modeling examined general and specific factors of social support and distress in relation to SI (week prior to hospitalization, via the Beck Scale for SI) and SAs (past 30 days, via the Columbia Suicide Severity Rating Scale). SI was significantly associated with the general social support (B = -1.51), the general social distress (B = 1.67), and the specific perceived burdensomeness (B = 1.57) factors. SAs were significantly associated with the specific Perceived Rejection (OR = 1.05) and Thwarted Belongingness (OR = 0.91) factors. General social support and social distress were associated with SI but not recent SAs. Specific social distress factors were also related to SI and SAs controlling for general social distress, suggesting areas for future interventions.

Minimal-Effect Testing, Equivalence Testing, and the Conventional Null Hypothesis Testing for the Analysis of Bi-Factor Models

A necessary step in applying bi-factor models is to evaluate the need for domain factors with a general factor in place. The conventional null hypothesis testing (NHT) was commonly used for such a purpose. However, the conventional NHT meets challenges when the domain loadings are weak or the sample size is insufficient. This article proposes using minimal-effect testing (MET) and equivalence testing (ET) to analyze bi-factor models. A key element in conducting MET and ET is the minimal size of factor loadings that can be regarded as noteworthy in practice, termed as minimal noteworthy size. This article presents two approaches to formulating the minimal noteworthy size and compares the pros and cons of MET, ET, and the conventional NHT. Analysis shows that MET, ET, and the conventional NHT are complementary. Combining them to test the noteworthiness of domain loadings can help researchers make a comprehensive judgment. Real and simulated datasets illustrate the applications of the three methods. Monte Carlo results show that MET and ET can control type I errors reasonably well while maintaining good statistical power.

Correction: Vispoel et al. Extending Applications of Generalizability Theory-Based Bifactor Model Designs. Psych 2023, 5, 545–575

In the original publication [...]

Psychometric Properties of the Spanish Motives for Online Gaming Questionnaire in a Sample of College Students.

This study investigates the psychometric properties of the Spanish version of the Motives for Online Gaming Questionnaire (MOGQ). We explored the factor structure and construct validity of the MOGQ through its relationships with gaming disorder symptoms (IGD-20) and impulsivity traits. We also analyzed if sociodemographic variables and gaming habits were related to gaming motives. An online cross-sectional survey was completed by 845 college students. Structure validity was examined using a combination of exploratory and confirmatory factor analyses, which supported a bifactor model composed of a general motivation factor and six uncorrelated factors (a mixed factor composed of escape and coping, competition, recreation, skill, social, and fantasy). Omega-hierarchical and omega coefficients were used to determine reliability of the MOGQ. The scale presented acceptable reliability for the general factor (ωh = .79) and the specific factor scores (social ω = .79, escape/coping ω = .81, competition ω = .79, skill ω = .84, fantasy ω = .82, and recreation ω = .70). Positive associations were observed between the MOGQ and the IGD-20 symptoms, with escape/coping (r = .48) and fantasy (r =.40) showing the strongest ones. Null or low correlations were observed with impulsivity traits. Motives to play varied significantly across genders. These findings provide evidence that the Spanish version of the MOGQ is a reliable and valid tool to assess motives to play online games.

Career-Related Teacher Support Scale: Validation Among Italian Adolescents and Assessment of the Effect on Career Decision Self-Efficacy Through Career Adaptability

Teachers have a key role in adolescents’ career development. Nevertheless, few studies examined their effect on career outcomes. The present study has two objectives. First, to test the factorial structure of the Italian version of the Career-related Teacher Support Scale (CRTSS) with Italian adolescents. Second, to test how career-related teacher support leads to career decision self-efficacy through career adaptability. In Study I, 469 adolescents were asked to complete the CRTSS. A first-order model, a second-order model (hierarchical), a bi-factor model (hierarchical), and a single-factor model were compared to check the best model solution. Additionally, the measurement invariance was examined across male and female. Results reveal the superiority of the second-order (hierarchical) model and the invariance between males and females. In the Study II took part 228 adolescents. A structural equation modeling approach assessed the relationship between career-related teacher support, career adaptability and career decision self-efficacy. A mediation model showed a significant indirect effect of career-related teacher support on decision self-efficacy through career adaptability. Practical implications for interventions were discussed.

Assessing the Factor Structure and Construct Validity of the Beck Depression Inventory (BDI-II) in a Korean Preschool Teacher Sample

Preschool teachers are at high risk of depression. Therefore, emphasis should be placed on evaluating a valid and reliable instrument to measure depressive symptomatology for this population. One such promising instrument is the Beck Depression Inventory (BDI-II), a 21-item self-report measure designed to assess the presence and severity of depression. Although the tool’s psychometric properties have been analyzed in different samples, they have not been validated with a Korean preschool teacher sample. Hence, this study sought to evaluate the psychometric properties of the Korean version of the BDI-II in a sample of Korean preschool teachers. The study participants comprised 252 Korean preschool teachers (243 women and 9 men) with an average age of 33.5 years (SD = 10.3, range = 21-59). The participants completed the Korean version of the BDI-II online. Confirmatory factor analysis was performed to examine the psychometric properties of the BDI-II. Six models based on the literature were tested, including single-factor, correlated two- and three-factor, and bifactor models. The reliability and validity of the BDI-II were also analyzed. The bifactor model demonstrated the best data fit with an overarching vital general depression factor that coexists with comparatively weak specific cognitive, affective, and somatic factors. Internal consistency, measured using McDonald’s omega coefficients, was adequate in all cases. Using a validated screening instrument such as the BDI-II may allow clinicians to better detect depression among preschool teachers because of its conciseness and satisfactory psychometric properties. Since a bifactor model confirmed the unidimensionality of the BDI-II, the scores of specific factors should not be used in isolation. However, clinical assessment may benefit from its subscales if they are used in conjunction with total scores.

Opposition to Same-Sex Parenting as Sexual Prejudice: Validity Evidence of a Model and Associated Factors in the Brazilian Context

The Beliefs about Children’s Adjustment in Same-Sex Families Scale (BCASSFS) is an instrument that measures rejection of the child-rearing and educational practices of same-sex parents. This scale evaluates beliefs that are organized in a bifactorial model referring to Individual Opposition and Normative Opposition. This study aimed to adapt the BCASSFS and examine its validity in the Brazilian context with a sexually diverse sample. A national survey was carried out with a sample of 1038 participants. It was used a sociodemographic questionnaire; a version of the BCASSFS adapted for the Brazilian context; a measure of sexual prejudice; and a measure of conservatism. Confirmatory factor analysis showed good fit indices for the model suggested for the original scale. Multigroup confirmatory factor analysis showed that the BCASSFS displayed measurement invariance between subsamples of different profiles, such as heterosexual-cisgender and LGBT + participants. Investigations of concurrent, convergent and criterion validity also showed that both scale factors were associated with theoretically related variables, in line with the literature. Together, these results show that the BCASSFS has good validity and can be used in the Brazilian context. By confirming the replication of the original model, this study contributes to broadening knowledge about these dimensions of sexual prejudice.

Decomposing the variance in early maladaptive schemas: the major role of one general factor, the minor role of domains, and their differential relations to facial emotion recognition.

Despite the growing interest in the early maladaptive schemas, the progress in understanding their impacts is decelerated by a lack of clear understanding of their structure. Different composite scores are calculated without a solid ground or a clarified meaning. Here we explain that the schema variance can be theoretically decomposed into three components: schema-specific, domain-specific due to the unmet core needs, and the common variance we call general susceptibility; each can differentially correlate with other substantive variables. Using this framework, we empirically examine the structure of schemas and their relationships to facial emotion recognition, a crucial ability that can widely affect our social interactions. A sample of adults completed an emotion recognition task and the Young Schema Questionnaire. Using different factor models, the specific and shared variance across schemas was analyzed. Then, the relation of these variance components to facial emotion recognition was explored. A general factor explained 27%, 40%, and 64% of the total variance in items, schemas, and domains, respectively. Partialling out the common variance, there was little domain-specific variance remained. Regarding facial emotion recognition, they were not correlated with specific schemas; however, the general susceptibility factor was correlated with anger recognition. The variance decomposition approach to schemas, which uses the bifactor model, may offer a clearer way to explore the impacts of schemas. While domain scores are widely used, their reliability, validity, and meaning are questionable. The generic factor, which is consistently extractable from empirical data, requires further attention.

Support for a Single Underlying Dimension of Self-Reported Health in a Sample of Adults with Low Back Pain in the United States

There is increasing interest in measuring “whole person” health and deriving an overall summary score. Underlying physical and mental health dimensions have been found consistently in prior studies of self-reported health, but it is unclear whether a single underlying health factor is supported across health domains. We examine the dimensionality of nine domains from the Patient-Reported Outcomes Measurement Information System (PROMIS®)-29 + 2 profile measure, the PROMIS social isolation scale, the Personal Wellbeing Index, and the EQ-5D-5L preference score in a sample of 1256 adults with back pain in the United States: mean age was 55 (range 18–94), 52% female, 74% non-Hispanic White, 61% were married or living with a spouse, and the highest level of education completed for 35% of the sample was a high school degree or general education diploma. The sample reported substantially more pain intensity, pain interference, and worse physical function than the U.S. general population. Product-moment correlations among the measures ranged from 0.25 to 0.83 (median correlation = 0.52). A bifactor model showed that a general health factor accounted for most of the covariation among measures, but physical function, pain interference, and pain intensity loaded slightly more on the physical health group factor than on the general health factor. The study provides some support for combining multiple aspects of self-reported health into an overall indicator of whole-person health.