No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria Versus New Alternatives
Abstract
This article examines the adequacy of the “rules of thumb” conventional cutoff criteria and several new alternatives for various fit indexes used to evaluate model fit in practice. Using a 2‐index presentation strategy, which includes using the maximum likelihood (ML)‐based standardized root mean squared residual (SRMR) and supplementing it with either Tucker‐Lewis Index (TLI), Bollen's (1989) Fit Index (BL89), Relative Noncentrality Index (RNI), Comparative Fit Index (CFI), Gamma Hat, McDonald's Centrality Index (Mc), or root mean squared error of approximation (RMSEA), various combinations of cutoff values from selected ranges of cutoff criteria for the ML‐based SRMR and a given supplemental fit index were used to calculate rejection rates for various types of true‐population and misspecified models; that is, models with misspecified factor covariance(s) and models with misspecified factor loading(s). The results suggest that, for the ML method, a cutoff value close to .95 for TLI, BL89, CFI, RNI, and Gamma Hat; a cutoff value close to .90 for Mc; a cutoff value close to .08 for SRMR; and a cutoff value close to .06 for RMSEA are needed before we can conclude that there is a relatively good fit between the hypothesized model and the observed data. Furthermore, the 2‐index presentation strategy is required to reject reasonable proportions of various types of true‐population and misspecified models. Finally, using the proposed cutoff criteria, the ML‐based TLI, Mc, and RMSEA tend to overreject true‐population models at small sample size and thus are less preferable when sample size is small.
... .95 (Hu & Bentler, 1999). ...
... For NARQ, we report 95/5 Level-1 thresholds by McNeish and Wolf (2023). For CNI, the dynamic package did not return a solution, hence we inspected results relying on Hu and Bentler's (1999) cutoffs. ...
A wide range of trait characteristics describe narcissism, but there is active debate about which features are most central to this construct. To explore the narcissism spectrum, the current study used a community sample of adults (N = 555) and applied psychometric network analysis to investigate the core elements of narcissistic manifestations. By doing so, we administered three distinct narcissism questionnaires comprising four facets of narcissism (i.e., agentic, antagonistic, neurotic, and communal facets). Given that narcissistic trait expressions can vary systematically for men and women, we additionally examined potential gender differences in network structure. Results revealed that agentic, antagonistic, communal, and neurotic narcissism each play a central role in defining narcissism, with distinct connectivity patterns observed across the facets. Network-comparison testing revealed male-female network invariance, but males and females showed no substantial differences in narcissistic configurations.
... All CFA analyses were conducted in MPlus Software Version 8 (Muthén & Muthén, 2017) using full-information maximum-likelihood estimation. Global model fit was evaluated by exact-fit test (chi-square test) and consideration of the approximate fit indices including Steiger-Lind rootmean-square error of approximation (RMSEA; with values of 0.08 or less acceptable; Browne & Cudeck, 1993;Steiger, 1990), comparative fit index (CFI; with values of .95 or above preferred; Hu & Bentler, 1999), and standardized root-mean-square residual (SRMR; with values of .08 or less preferred; Browne & Cudeck, 1993). To assess local fit for each model, we reviewed the residual variance patterns for large residual values indicative of areas of misfit (as recommended by Kline, 2023), including special cases such as Heywood cases. ...
Purpose
This study investigated the dimensionality of language in bilingual children using measures of semantics and morphosyntax in English and Spanish.
Method
Participants included 112 Spanish–English bilingual children ages 4–8 years from a wide range of language abilities and dominance profiles. Using measures of semantics and morphosyntax from both norm-referenced assessments and language samples, we evaluated the structure of language in bilingual children. We used confirmatory factor analysis (CFA) to estimate dimensionality, comparing seven primary models that represented different theoretical structures of language in bilinguals.
Results
Although none of the models analyzed yielded good fit across all indices evaluated, the best-fitting CFA model was a two–correlated factor model with separate factors for Spanish and English, which included measures from only norm-referenced assessments.
Conclusions
Language in Spanish–English children seems to represent two related but distinct constructs, even in bilinguals from a wide range of language abilities and dominance profiles. Clarifying how language in bilinguals is conceptualized and impacted by the concurrent development of two languages is an area that requires further research. Understanding the dimensionality of language in bilinguals can further assist our knowledge of how language develops in bilingual children.
Supplemental Material
https://doi.org/10.23641/asha.28687466
... Although the present study extends research on the role of implicit sexual beliefs in the face of an imagined disclosure of sexual dysfunction in newly formed relationships, it is not without its limitations. First, our model fit was adequate but not good, leaving room for further improvement (see Hu & Bentler, 1999). Although controlling for the experimental condition and type of sexual dysfunction slightly improved model fit, these changes were minor. ...
Guided by implicit theories of relationships and sexual satisfaction and the theory of resilience and relational load (TRRL), our study investigated the mediating role of sexual communal strength on associations between implicit beliefs regarding sexual compatibility and relational thriving in new romantic couples (< 6 months duration) facing a hypothetical scenario of sexual dysfunction (i.e., vulvovaginal pain or erectile dysfunction). Participants in new romantic relationships (N = 461) completed an experimental survey in which they imagined their partner was experiencing sexual dysfunction. We explored indirect associations among implicit beliefs about sexual compatibility, specifically sexual growth (i.e., sexual [in]compatibility requires work) and sexual destiny (i.e., sexual [in]compatibility as being reflective of a [poor] romantic match) beliefs, and couples’ anticipated relational thriving, mediated by sexual communal strength. Our hypothesized model was tested via structural equation modeling in SPSS AMOS 29. The findings indicated that sexual growth beliefs, but not sexual destiny beliefs, were indirectly associated with anticipated relational thriving through sexual communal strength. Specifically, sexual growth beliefs were positively associated with sexual communal strength, which in turn was associated with and greater anticipated relational thriving. Although the indirect effect for sexual destiny beliefs was not significant, sexual destiny beliefs were significantly associated with less anticipated relational thriving. We extend theorizing about implicit (sexual) beliefs by examining a new relational context—relationship initiation—as well as offer practical implications for partnerships experiencing sexual dysfunction.
... indicated good fit (Browne & Cudeck, 1992), while RMSEA below .08 and CFI above .90 indicated acceptable fit (Hu & Bentler, 1999). Since occupational stress can differ based on gender (Spielberger & Reheiser, 1994), we controlled for gender by including it as a predictor of all latent factors. ...
To keep up with fast-paced societal changes, teachers must continuously commit valuable time and effort to their professional learning. However, teaching is already a demanding profession, leading to high levels of work-related stress - possibly impeding commitment to professional learning. Data were collected among 151 secondary school teachers for 15 consecutive work days. Models comparing teachers and daily changes were tested, accounting for the dynamic nature of stress emotions and the profession. The results showed ambiguous findings, emphasizing the differences in responses to the daily and cross-sectional measures. These findings and their relevance for practice are discussed.
... A következő értékhatároknak megfelelően a modell a mintára megfelelően illeszkedőnek tekinthető: CFI > 0,95; TLI > 0,95; RMSEA < 0,06. Ugyanakkor elfogadhatónak tekinthető az illeszkedés, ha a CFI és TLI értékek 0,90 felettiek, míg az RMSEA esetében a 0,80 alatti érték tekinthető még elfogadhatónak (Hu & Bentler, 1999). A skálák megbízhatóságát Cronbach-alfa értékekkel jellemeztük. ...
A magas szintű szövegértelmezés számos, egymással szorosan összefüggő tudáselem elsajátítását és gyakorlását feltételezi, amelyek egyikét az olvasási stratégiák jelentik. A magyar tanulók olvasásistratégia-használatáról és az azokat befolyásoló tényezőkről viszonylag keveset tudunk, ami minden bizonnyal összefügg a magyar tanulók stratégiahasználatának megismerésére alkalmazható megfelelő eszközök alacsony számával. Az említett hiányosságok mérséklése érdekében munkánk célja (1) a nemzetközi szakirodalomban széles körben alkalmazott Metacognitive Awareness of Reading Strategies Inventory (MARSI, Mokhtari & Reichard, 2002) átdolgozott változatának, a Metacognitive Awareness of Reading Strategies Inventory – Revised (MARSI-R, Mokhtari et al., 2018) kérdőívnek a kipróbálása felső tagozatos tanulók körében; (2) a tanulók stratégiahasználati jellemzőinek feltárása; valamint (3) az olvasási stratégiák és a rendelkezésre álló háttérváltozók közötti kapcsolatok elemzése. Keresztmetszeti kutatásunkban 5–8. évfolyamos tanulók vettek részt, összesen 926 fő. Kutatásunk eredményei alapján a MARSI-R kérdőív alkalmazható a magyar felső tagozatos tanulók körében, bár az 5. és 7. évfolyamosok esetében további fejlesztések szükségesek az egyes stratégiacsoportok strukturális validitásának javítása érdekében. A reliabilitásértékek a teljes mintán és az egyes évfolyamokat tekintve mindhárom stratégiacsoport kapcsán elfogadhatóak. A problémamegoldó stratégiák használatában nincs eltérés az évfolyamok között, néhány átfogó és támogató stratégiát viszont a 8. évfolyamosok használnak inkább. Az olvasási stratégiák és a háttérváltozók közötti kapcsolatok általában gyengék voltak. Kutatásunk illeszkedik ahhoz a mintázathoz, mely szerint a magyar tanulók körében az olvasási stratégiák mérésére adaptált, elsősorban angol nyelven kidolgozott kérdőívek faktorstruktúrája teljes mértékben nem reprodukálható.
... each of the constructs has indicators that allow measurement of the overall latent construct. this method is preferred because it estimates the multiple and interrelated dependence in a single analysis (hu & Bentler, 1999). ...
The current study analyzes if higher purchase intention can result from the communication, compliance with, and reporting of sustainable initiatives, and the perceived quality of products in a business-to-business (B2B) context. The theory of consumption values, and the triple bottom line theory help set the study’s framework. Data was gathered through a Qualtrics panel survey of 174 B2B buyers. Structural Equation Modeling was used as the methodology. Results suggest that sustainability and reporting efforts from the seller can positively impact purchase intention. The perceived quality of the products sold moderates the relationship between the seller and the buyer’s intent to purchase sustainable products. Managers should avoid creating voluntary sustainability measures. The company instead could have invested resources into enhanced communication and reporting sustainable initiative results as motivators for buyers to purchase high-quality green products. This study promotes novel discourse on sustainable communication and the perceived quality of sustainable products. As the seller appeals to a buyer in a societal context by offering environmentally friendly products, sustainability becomes central to revenue generation through the triple bottom line perspective.
... The mean of the slope factor represents the average growth rate of PSMU across various time points, and the variance of the slope factor reflects the differences in growth rate among individuals' PSMU. Furthermore, model fit was evaluated using the following indices (Hu & Bentler, 1999): the comparative fit index (CFI; acceptable >0.90), Tucker-Lewis index (TLI; acceptable >0.90), root mean square error of approximation (RMSEA; acceptable <0.08), and standardized root mean square residual (SMRM; acceptable <0.08). Given that the rates of missing data for all variables in the current study were less than the recommended cut-off of 5% (Moon, 1996), Expectation-Maximization (EM) was applied to address the missing data. ...
Background and aims
Researchers have suggested that subtypes of problematic social media use (PSMU) should be identified for purposes of prevention and intervention. However, most studies have overlooked the heterogeneous characteristics of PSMU trajectories, and no research has systematically examined which interpersonal factors could predict these trajectories. In the present study, we identified classes of developmental trajectories of PSMU and examined differences across classes in adolescents' interpersonal functioning in family, school, and peer contexts.
Methods
Participants were 357 Chinese adolescents enrolled in two middle schools in China (52.1% girls, aged 12–15 years). The students completed questionnaires in their classrooms over the course of one year in a three-wave longitudinal study.
Results
Latent growth mixture modeling (LGMM) revealed three developmental trajectory classes of PSMU based on the intercepts and slopes of PSMU scores over time: high risk-gradual increase group (37%), low risk-sharp increase group (39%), and low risk-stable group (24%). Parent-adolescent attachment (family context), teacher-student relationships (school context), and deviant peer affiliation (peer context) were associated with variations in developmental trajectories.
Conclusions
The findings can inform the design of prevention and intervention programs for specific subgroups of adolescents who show problematic social media use.
... The measurement model was tested, and structural model analysis was conducted to examine the relationships between variables and analyze the hypotheses. Model evaluation utilized fit indices including chi-square/degrees of freedom (χ 2 /df), root mean square error of approximation (RMSEA), comparative fit index (CFI), Tucker-Lewis index (TLI), and standardized root mean square residual (SRMR) (Kline, 2015), with χ 2 /df < 3, SRMR and RMSEA < 0.08, and CFI and TLI > 0.90 indicating good model fit (Hu & Bentler, 1999;Kline, 2015). Furthermore, the bootstrap method was employed to test the significance of the effect of artificial intelligence literacy on fear of innovation through lifelong learning (Preacher & Hayes, 2004). ...
The study aims to identify the profiles of university students regarding Artificial Intelligence Literacy, Lifelong Learning, and Fear of Innovation using cluster analysis and to examine the relationships among these variables. Cluster analysis and structural equation modeling were conducted with valid responses from 402 university students. The cluster analysis identified three distinct student profiles: the highly adaptive group (Profile 1), the needs improvement group (Profile 2), and the high support required group (Profile 3). Structural equation modeling revealed that artificial intelligence literacy positively affects the tendency for lifelong learning and negatively impacts the fear of innovation. Lifelong Learning Trends also negatively influences the fear of innovation. Furthermore, artificial intelligence literacy was found to indirectly reduce fear of innovation through lifelong learning tendencies. The findings from cluster analysis and structural equation modeling provide significant insights into understanding university students’ artificial intelligence literacy, lifelong learning tendencies, and fear of innovation. Developing customized education and support programs tailored to each profile’s characteristics and the relationships among the study variables can help students enhance their competencies in these areas.
... y SRMR ≤ .10 (Hu & Bentler, 1999;Mehmetoglu & Jakobsen, 2016). Dado que el instrumento incluye ítems de puntuación inversa, se exploró la posible influencia de un factor de método asociado a este tipo de puntuaciones. ...
El objetivo del estudio fue traducir el Bedtime Procastination Scale (BPS) al español y evaluar sus propiedades psicométricas. Participó una muestra de 419 estudiantes de tres universidades peruanas. Se realizó un análisis factorial confirmatorio para evaluar el modelo original de la escala, el cual inicialmente mostró índices de ajuste pobres. Se obtuvieron mejores índices de ajuste después de eliminar los ítems con puntuación inversa, χ²(5) = 9.240, CFI = .998, RMSEA = .045, TLI = .997, y SRMR = .013. La consistencia interna de la escala, examinada mediante el coeficiente omega, resultó ser satisfactorio (ω = .86). Asimismo, el análisis de invarianza de la medición demostró la invarianza de género, lo que indica que la escala funciona de manera equivalente en participantes masculinos y femeninos. Además, se evaluó la validez convergente examinando las correlaciones entre la BPS y el insomnio, medido por la Escala de Insomnio de Atenas (AIS-5), encontrando una correlación moderada (r = .44). En resumen, la traducción al español de la BPS demostró adecuadas propiedades psicométricas en universitarios peruanos y una estructura unifactorial después de la eliminación de los ítems con puntuación inversa.
... the root-mean-square error of approximation (values of .06 or less), and the standardized root-mean-square residual (value less than .08) (Hu & Bentler, 1999). Measuring Barriers to Physical Activity Support in Parents of Children With Disabilities 29 한국지체 · 중복 · 건강장애교육학회 Ⅲ ...
Purpose: The purposes of the current study were: 1) to develop a questionnaire examining
barriers to parental physical activity (PA) support and 2) to examine how barriers are associated
with intention and parental PA support in parents of children with disabilities.
Method: This study was conducted in three phases. Phase I focused on item development,
while Phase II addressed content validity. Phase III involved field testing. The Questionnaire
for Barriers to Parental PA Support in Parents of Children with Disabilities (QBPPS-PCD) was
initially developed, comprising three constructs and 16 items. Additionally, parents’ intentions
to support their child’s physical activity and their actual support behaviors were measured to
examine their association with perceived barriers.
Results: A final model with nine items (three items in each construct) was developed, which revealed an excellent model fit([df = 24, p = .147], CFI = 0.99, TLI = 0.93. RMSEA = 0.05). Pathway analysis indicated that competence barriers were significantly and negatively
associated with intention (β = -.28, p = .03), which in turn was significantly and positively
associated with parental PA support (β = .31, p < .001).
Conclusion: The QBPPS for PCD consists of three constructs, differently associated with
intention and parental PA support. The questionnaire developed in this study can serve as a
tool for assessing parents’ competence barriers.
... The established measuring model was not found to fit well. (Hu & Bentler, 1999;Sumer, 2000) RMSEA .06 .05 ...
Introduction.
Adolescence is a critical period of rapid growth and change during which family harmony and attachment styles play a significant role in shaping behavior. This study examines how these factors influence social media use disorder among adolescents using structural equation modeling. Methods.I n this study, 355 high school students aged 14-18 (53.5% girls) studying at various levels of high school affiliated with the Ministry of National Education in Turkey participated in the 2023-2024 academic year. The study data were collected by the convenience sampling method in the classrooms where the students were enrolled. Results. According to the results obtained in the study, secure, avoidant, and anxious attachment styles have a mediating role in the relationship between family harmony and social media use disorder. An analysis of the effect values ob-tained from structural equation modeling shows that family harmony negatively predicts anxious and avoidant attachment styles. On the other hand, family harmony positively predicts a se-cure attachment style, while a secure attachment style negatively predicts social media use disor-der. The model in this study shows that anxious and avoidant attachment positively affects so-cial media use disorder. The results indicate that family harmony negatively predicts anxious and avoidant attachment styles while positively pre-dicting secure attachment, which in turn reduces social media use disorder. Anxious and avoid-ant attachment styles were found to mediate the relationship between family harmony and social media use disorder. These findings emphasize the need for parental education programs to strengthen family unity and mitigate negative social media use among adolescents.
Conclusions. Adolescents' harmful use of social media affects attachment to parents and family harmony. For this reason, providing family unity along with parent education will reduce the level of adolescents’ exposure to negative social media. Limitations of the study. This study was conducted in one of the eastern provinces of Turkey. Therefore, the sample was limited to students enrolled in high schools in this province. In addition, data were collected only from adolescents, and no data were collected from parents or teachers. Another limitation is that data were collected at one time instead of at different times from the same individuals.
Key words: family harmony, attachment styles, social media use disorder
... The objective is to ascertain whether the model is acceptable, based on the following fitting metrics. It is commonly accepted that a CMIN/ DF < 3.000、RMSEA < 0.080 an RMR value of less than 0.050, and CFI、GFI、NFI、TLI、IFI > 0.900 are indicative of an appropriate model fit [18]. The evaluation of the scale's validity was conducted through the assessment of its convergent and discriminant validity. ...
Background
Suicide represents a significant public health concern at the global level and is a major area of concern for mental health professionals. Nurses are positioned to identify and manage individuals at risk of suicide or suicidal ideation. It is widely acknowledged that ensuring nurses are adequately trained to assess and manage suicidal patients is of paramount importance in the prevention of suicide. The objective of this study was to examine the reliability and structural validity of the Suicide Management Competency Scale (SMCS) in clinical nurses population.
Methods
A total of 452 clinical nurses in a third-class hospital in Liaoning Province were selected using convenience sampling. The survey was conducted using the general demographic questionnaire and the Suicide Management Competency Scale (SMCS). The reliability and acceptability of the scale were assessed by checking the consistency of the scale part, the split-half reliability coefficient, and the correlation between each item and the score of the total scale. Confirmatory factor analysis, convergent validity and discriminant validity were used to determine the dimensional structure and validity of the scale.
Results
The internal consistency of the scale, as indicated by the Cronbach’s α coefficient, was 0.902. The split-half reliability coefficient was 0.771. The results of the confirmatory factor analysis were as follows: CMIN/DF = 2.609, RMSEA = 0.060, RMR = 0.040, CFI = 0.957, GFI = 0.930, NFI = 0.933, TLI = 0.949, IFI = 0.958. All of the model fitting indexes were within the acceptable range. The average variance extracted (AVE) of the three subscales ranged from 0.500 to 0.583. The combined reliability values (CR) range from 0.848 to 0.888, indicating that the SMCS scale exhibits good convergence validity. The analysis of the correlation coefficient between the subscale and the total scale revealed that the AVE square root value of each subscale is between 0.707 and 0.763, which is greater than the correlation coefficient between the two indicators. This indicates that there are significant differences between the subscales of the SMCS, that the internal structure of the questionnaire is highly differentiated, and that the discrimination validity is good.
Conclusion
To the best of our knowledge, this is the inaugural study to report the utilisation of the Suicide Management Competency Scale (SMCS) in the context of clinical nursing. The findings offer preliminary support for the utilisation of the SMCS in clinical nurses. In the future, nursing managers will be able to effectively evaluate clinical nurses’ ability to manage and prevent suicide, as well as train nurses in this area, with the ultimate goal of saving as many lives as possible.
... It has been suggested (O'Connor, 2000) that the MAP test might underestimate the number of factors, and the parallel analysis may overestimate it; therefore, we conducted both analyses. Following Hu and Bentler (1999), model fit was tested using the following indices: chi-square statistic (χ 2 ), confirmatory fit index (CFI), root-mean-square error of approximation (RMSEA), and standardized root mean residual (SRMR). CFI values > 0.95 and 0.90, RMSEA values < 0.06 and 0.10, and SRMR values < 0.06 and 0.09 were considered good and adequate fits, respectively. ...
Objectives
Based on current research, mindfulness interventions have a positive impact on both the mental and physical health of adults and adolescents. However, there is a growing need for additional research to develop properly validated measures. The objective of this study was to validate the Polish version of Child and Adolescent Mindfulness Measure (CAMM).
Methods
A sample of 325 adolescents aged 12–15 years (M = 13.39, SD = 1.05, 50% females) completed the Polish version of CAMM together with measures of mental well-being, psychological distress, psychological inflexibility and thought suppression. Factor structure using confirmatory factor analysis (CFA), internal consistency and convergent validity of the CAMM were assessed.
Results
The results provide evidence for the original one-factor structure of the CAMM, with good internal consistency with Cronbach’s α = 0.92 and McDonald’s ω = 0.92. As predicted, CAMM scores correlated positively with mental well-being and negatively with psychological distress, psychological inflexibility and thought suppression.
Conclusions
The Polish 9-item version of CAMM shows good psychometric qualities, being a promising tool to measure mindfulness in Polish-speaking children and adolescents.
... TLI and CFI > .95, and SRMR < .08 (Hu & Bentler, 1999). Models were performed both for the general sample and for each country. ...
Teachers’ burnout, a psychological syndrome that emerges in response to chronic interpersonal stressors on the job, can have serious consequences on children and staff in schools. The Maslach Burnout Inventory – Educators Survey (MBI-ES) is the most widely used scale to measure burnout among educational staff. However, this scale has revealed divergent results in factorial validity analysis across countries, calling for further cross-cultural research. This study examines the cross-cultural factorial validity of MBI-ES among a sample of 391 early childhood education professionals from four countries – Greece, Cyprus, Portugal, and Romania. Results of multi-group confirmatory factor analysis suggest that a 1-factor solution, exclusively focused on measuring emotional exhaustion, holds partial scalar invariance across countries. In light of previous research on this scale, the conceptualization of burnout, and the specificities of working in early childhood education and care contexts, these findings are discussed.
... SRMR = .042) according to common standards (Hu & Bentler, 1999), and better than the nine-factor model (χ 2 = 3555.524, p < .001, ...
... Following the two-index presentation format suggested in prior literature (L. Hu & Bentler, 1999), the root mean square error of approximation (RMSEA) and the standardized root mean square residual (SRMR) were used to evaluate model fit. The cutoffs for model acceptability were an RMSEA of .06 or lower and an SRMR of .09 ...
... Subsequently, a confirmatory factor analysis (CFA) was conducted to assess validity. The evaluation of the CFA model fit was based on standard fit indices (Hu & Bentler, 1999). The results of the CFA are provided in Table A1 in the Appendix. ...
The integration of arts into science education is increasingly recognized as a means to enhance both cognitive and non-cognitive learning outcomes. This study examines whether the incorporation of art elements into STEM lessons influences pupils' motivation, self-efficacy, flow experience, and subject-specific interest. A longitudinal design was employed, with data collected from 471 pupils across eight schools at two measurement points. The findings indicate a significant increase in self-efficacy, subject-specific STEM interest, extrinsic regulation and interest in art among pupils who participated in the intervention. These results suggest that artistic activities not only make STEM subjects more engaging but also foster a broader appreciation for the arts. The study highlights the potential of interdisciplinary teaching strategies in enhancing motivation and self-efficacy, offering valuable implications for educational practice and future research.
... When all factor loadings exceed 0.6 and are statistically significant, the data are considered acceptable (Ashill & Jobber, 2010). CFI and TLI are evaluated against a standard threshold of 0.9, while RMSEA and SRMR are assessed against a threshold of 0.8 (Hu & Bentler, 1999). ...
Online anonymity refers to individuals to interact online without fully revealing their identity. To understand the motivations behind this choice in the Chinese context, we translated the Online Anonymity Questionnaire (OAQ) and assessed it with 1,405 college students in Guangxi, China, followed by a retest with 245 participants a month later. The study evaluated the reliability and validity of the Chinese version of the OAQ, assessing internal consistency, test-retest reliability, confirmatory factor analysis, and measurement invariance. The questionnaire, featuring two dimensions anonymous self-expression and anonymous toxicity-demonstrated good internal reliability and test-retest reliability for both subscales. Confirmatory factor analysis showed the two-factor model was a good fit (CFI ¼ 0.98, TLI ¼ 0.97, RMSEA ¼ 0.05), and partial scalar invariance was confirmed. These findings validate the Chinese version of the OAQ for studying online anonymity motivations among Chinese college students. PUBLIC SIGNIFICANCE STATEMENTS This study successfully translated and validated the Online Anonymity Questionnaire (OAQ) for Chinese college students, confirming the dual motivations for seeking online anonymity-self-expression and toxicity. These findings deepen our understanding of the psychometric functioning of the OAQ in China and support its use as a psychometrically sound tool for assessing online anonymity behaviors in Chinese contexts.
Background
Secure attachment bonds between children and their parents are crucial for their positive developmental outcomes. In attachment-based treatment, there are few instruments available for identifying insecure child-caregivers attachment connection, particularly in terms of primary caregiver’s perceptions. This perspective is an important clinical component to consider in parenting therapies for preschool children. This main goal of this study was to analyze the psychometric properties of the Chinese version of the Attachment Insecurity Screening Inventory 2–5 Years (hereafter referred to as AISI 2–5 years), a parent-report questionnaire designed for assessing attachment insecurity in children aged 2 to 5 years.
Methods
The AISI 2–5 years underwent rigorous translation and back-translation processes. The sample comprised 486 preschoolers (Mage = 47.83 months, SD = 8.58; 49.79% girls) and their mothers (Mage = 36.30 years, SD = 3.82) from Shanghai, China. Confirmatory factor analysis (CFA) was used to explore internal structure, while Cronbach’s alphas were utilized to evaluate score reliability.
Results
The findings revealed a three-factor model, encompassing avoidant attachment, disorganized attachment, and ambivalent/resistant attachment, which demonstrated a good fit for the 18-item AISI 2–5 years. Additionally, these findings demonstrated sufficient internal consistency, reliability, and convergent validity. Consequently, the AISI 2–5 years proves suitable for assessing attachment insecurity among Chinese preschoolers.
Conclusion
The results contribute to the advancement of research on children’s insecure attachment within the Chinese context. Furthermore, future research is needed to replicate the present findings and enhance the proof supporting the appropriateness of the AISI 2–5 years for preschoolers in China. This will demonstrate its utility in evaluating parents’ perception of their children’s attachment insecurity.
Background
Mental health providers’ attitudes toward evidence-based practice are likely to influence what interventions they learn, implement, and sustain over time. A 36-item version of the Evidence-Based Practice Attitude Scale (EBPAS) was recently developed to assess provider attitudes in 12 domains. Research suggests the EBPAS-36 is a promising tool, though inconsistencies across studies signal the need to reexamine its validity and reliability along with the correlates of provider attitudes.
Methods
This study assessed the factorial structure of the EBPAS-36, the intercorrelations and reliabilities of its subscales, and correlates of practice attitudes in a U.S. sample of 445 practitioners who received training in trauma-focused cognitive behavioral therapy.
Results
A confirmatory factor analysis (CFA) verified that the EBPAS-36 fits a 12-factor model representing each of its subscales. Reinforcing prior results, the subscales of the EBPAS-36 were weakly to moderately correlated, indicating that the 12 domains are related yet distinct. A hypothesized second-order CFA model with three overarching latent factors was not validated, but an alternative second-order model with two factors fit the data adequately. Most subscales demonstrated good-to-excellent internal consistency, though values for certain subscales ranged from marginally acceptable to poor. Provider attitudes varied by gender, professional experience, and discipline. Practitioners who more frequently assessed client trauma symptoms reported more positive EBP attitudes, and those who expressed greater concerns that trauma assessments may cause harm reported more negative attitudes.
Conclusions
Taken together with previous findings, the results show the EBPAS-36 performs well overall, though some subscales may benefit from refinement. Further validation tests of the EBPAS-36 in diverse samples are warranted.
Background
This study aimed to investigate the relationship between personality traits, academic burnout, and academic engagement among dental students with emotion regulation as a mediating role. It sought to identify personality predictors within the HEXACO model and compare how these traits directly and indirectly affect students’ academic engagement and burnout.
Methods
The participants were 228 dental students from School of Dentistry. Data were collected using the HEXACO personality inventory, the Emotion Regulation Questionnaire (ERQ), the Maslach Burnout Inventory-Student Survey (MBI-SS), and the Korean Academic Engagement Inventory (KAEI). The study employed structural equation modeling to explore the direct and mediated relationships between personality traits, emotion regulation, academic burnout, and academic engagement.
Results
The findings indicated that specific personality traits, notably Extraversion, directly and indirectly influence both academic burnout and engagement, with emotion regulation serving as a mediating factor. Extraversion affected engagement directly and indirectly through cognitive reappraisal. Moreover, the burnout model revealed that, besides Extraversion, Agreeableness and Openness to experience also had a direct impact on burnout, suggesting a broader range of personality traits influencing burnout compared to engagement.
Conclusions
The study underscores the significant role of personality traits, particularly Extraversion, in determining academic burnout and engagement, with emotion regulation often playing a mediating role. The findings of this study indicate that personality traits can be approached in a distinct manner when developing strategies to mitigate burnout and increase engagement among dental students. These insights could inform targeted interventions aimed at improving academic engagement and performance.
Bu çalışma, bireylerin iş ve aile yaşantılarındaki uyumsuz rollerinden kaynaklı yaşadıkları duyguların yoğunluğunu ölçme amacıyla McElwain (2008) tarafından geliştirilen İş-Aile Suçluluğu Ölçeği’nin Türkçe’ye uyarlanmasını ve psikometrik özelliklerinin incelenmesini içermektedir. Birinci çalışmada faktör analizi ve güvenirlik analizleri için 235 kadın ve 195 erkek olmak üzere toplamda 430 kişi (%9.1 özel sektör; %89.3 kamu sektörü ve %1.6 kendi işi) araştırma sürecine dâhil edilmiştir. İkinci çalışmada ise doğrulayıcı faktör analizi ve ölçüt geçerliliğinin test edilmesi amacıyla 133 kadın ve 85 erkek olmak üzere toplamda 218 kişi (%14.2 özel sektör; %84.4 kamu sektörü ve %1.4 kendi işi) ile veriler analiz edilmiştir. Ölçüt geçerliliği için İş-Aile Çatışması Ölçeği kullanılmıştır. Yapılan analizler sonucunda İASÖ-TR formunun orijinal ölçekte olduğu gibi suçluluğun yönünü belirleyen iş-aile suçluluğu ve aile-iş suçluluğu şeklinde iki temel boyuta sahip olduğu görülmektedir. Her bir alt boyutun da fiziksel, duygusal ve psikolojik suçluluğun türünü tanımlayan üç alt faktörden oluştuğu bulgulanmıştır. Dolayısıyla İASÖ-TR formu, beyaz yakalı çalışanların yaşayabilecekleri suçluluk duygusunu tanımlayabilmeleri açısından alan yazına katkı sunmaktadır.
Bu çalışma, sağlık sektöründe hem çalışanların iyi oluşunu hem de organizasyonel performansı artırmak için bilimsel temelli öneriler geliştirme ihtiyacından doğmuştur. Özellikle hemşireler gibi kritik meslek gruplarının yaşadığı rol belirsizliği ve yabancılaşma sorunlarının daha iyi anlaşılması, sektörde sürdürülebilir çözümler üretmek adına önem taşımaktadır. Bu doğrultuda, çalışma hemşirelerin rol belirsizliğinin işe yabancılaşma üzerindeki etkisini, örgütsel muhalefet ve işle ilgili kaygı gibi aracı değişkenler üzerinden inceleyerek, iş yaşamında karşılaştıkları problemlerin nedenlerini daha derinlemesine anlamayı ve uygulayıcılara yenilikçi çözüm önerileri sunmayı hedeflemektedir. Bu amaç doğrultusunda Hatay’da orta ölçekli bir hastanede hemşire olarak görev yapan 456 kişiye kolayda örnekleme yöntemi ile ulaşılarak araştırma modeli sınanmıştır. Kesitsel olarak tasarlanan araştırma tanımlayıcı-ilişki arayıcı tiptedir. Araştırma analizleri sonucunda hemşirelerin rol belirsizlikleri ile işe yabancılaşmaları, işle ilgili kaygıları ve örgütsel muhalefetleri arasında pozitif ve anlamlı ilişkiler tespit edilmiştir. Ayrıca hemşirelerin rol belirsizlikleri ile işe yabancılaşmaları arasındaki anlamlı ilişkide ayrı ayrı olarak hem örgütsel muhalefetin hem de işle ilgili kaygının aracı rollerinin olduğu tespit edilmiştir. Son aşamada gerçekleştirilen sıralı aracılık analizlerine göre de hemşirelerin rol belirsizlikleri ile işe yabancılaşmaları arasındaki anlamlı ilişkide işle ilgili kaygı ve örgütsel muhalefetin sıralı aracılık rolünün de olduğu bulgulanmıştır.
Purpose
Higher educational institutions (HEIs) have become dynamic marketing labs to gain international prominence. Built on the dynamic capability theory and the resource-based view, this study aims to explore the application of dynamic capability in the HEIs of the United Arab Emirates. Specifically, this study analyzes relationships between dynamic capabilities, internal marketing orientation (IMO) and educational competitiveness.
Design/methodology/approach
Data were collected through an online survey from 349 faculty members across 35 universities, including eight public and 27 private institutions. The model was tested using structural equation modeling in AMOS 24.0.
Findings
Empirical findings reveal that the direct influence of dynamic capabilities on competitiveness was insignificant. However, results reveal a positive relationship between dynamic capabilities and IMO. IMO positively impacts educational competitiveness and mediates the relationship between dynamic capabilities and educational competitiveness, emphasizing the role of internal organizational structures.
Practical implications
The findings suggest that fostering dynamic capabilities and cultivating an IMO can enhance the competitiveness of HEIs. This implies the need for HEIs to focus on internal organizational processes and strategies to maintain competitiveness in the rapidly evolving higher educational landscape.
Originality/value
This study contributes to the literature by exploring the dynamics between dynamic capabilities, IMO and educational competitiveness in the context of HEIs in the UAE. By extending the literature boundaries, it extends the literature boundaries and offers a strategic roadmap for such institutions.
Rapid advancements in digital technologies have revolutionized care delivery. The medical device market, driven by the rise of wearable devices, continuously develops and sees increased end-user acceptance. However, healthcare markets often have uneven power dynamics between patients and providers and high uncertainty levels, leading to potential information asymmetry. Drawing on signaling theory and consumer innovativeness, we examine their effects on end-user acceptance. We explore perceived value and hospital corporate image as potential predictors of the intention to use wearable devices. Using a sample of 201 respondents, including both prior and non-prior users of wearable devices, we test a conceptual model to predict usage intention. The findings confirm significant direct relationships between perceived value, consumer innovativeness, and the intention to use smart wearable devices. They also highlight the importance of a hospital’s corporate image to resolve information asymmetries. This confirms its positive indirect effect on the intention to use these devices. This study contributes to the literature on consumer behavior and technology adoption. To the best of the authors’ knowledge, it is the first to demonstrate the direct and indirect effects of brand image on the intention to use smart wearable devices.
Dysfunctional metacognitive beliefs have been established as a transdiagnostic correlate of anxiety and depression in adults and are considered a central mechanism of persistent emotional distress according to the metacognitive model of psychological disorders. However, the importance of metacognitions for distress and emotional disorder in adolescence is far less researched, and to investigate this further there is a need for reliable and valid assessment tools. The Metacognitions Questionnaire-30 (MCQ-30) might be a suitable and valid assessment tool in adolescents with the implication that it can be used to research the role of metacognitions in youth mental health and track how metacognitions change and relate to outcomes over time from adolescence to adulthood. We therefore aimed to examine the psychometric properties of the MCQ-30 in an Upper Secondary School-sample of 494 Norwegian adolescents, aged 16- to 18-years old. A confirmatory factor analysis indicated a good fit for the proposed five-factor structure and evidence for measurement invariance was supported across sex and groups of anxiety severity. The factors showed acceptable to good internal consistency and there was support for convergent validity. In conclusion, these findings indicate that the adult version of the MCQ-30 can be applied in Norwegian adolescents from 16-years old.
This study adopted a three‐wave random intercept cross‐lagged panel model (RI‐CLPM) to explore the longitudinal reciprocal relationships between social support and engagement in EFL learning among Chinese high school students. Two samples, including 3505 students (Sample 1 = 1794; Sample 2 = 1711), were tracked with their perceived social support and engagement levels in EFL learning over three academic semesters. Using a multiple‐indicator RI‐CLPM, the study demonstrated a reciprocal relationship between perceived social support and engagement in EFL learning maintained consistently across samples. Perceived social support in EFL learning positively predicted students' subsequent learning engagement. Meanwhile, students' engagement in EFL learning positively predicted their subsequent perceived social support. While prior cross‐sectional research mostly showed how social support predicts student engagement, not the reverse prediction, the current study provides initial evidence for the possibly mutually reinforcing relationship over time.
This study examines the extent if any by which the effect of low selfcontrol on crime is mediated by assessments of the costs and benefits of crime. It examines this idea with self-report survey data on intimate partner violence. The findings reveal that the effect of low self-control is partially mediated by the respondents’ cost considerations but not by their assessments of the anticipated benefits of intimate partner violence. While expected utility also partially mediated the effect of low self-control, we conclude that this is due primarily through the element of cost considerations.
This study was conducted to examine the relationship between parental interest levels and children's psychological well‐being and emotion regulation skills in 5–6‐years‐old (60–72 months) children. Structural equation modeling, a quantitative research method, was employed. A convenience sampling method was used to determine the study group. The study group consisted of 60–72‐months‐old children and their parents attending private and public schools in Sivas city center. Child participants were reached through individual interviews at their schools, while parents were reached through online forms. The study group consisted of a total of 441 individuals: 147 children aged 5–6, 147 mothers of these children, and 147 fathers of these children. The Parental Interest Towards Children Scale, the Psychological Well‐Being Scale for 5–6 Year Old Children, and the Preschool Emotion Regulation Scale were used as data collection tools. Data analysis processes were conducted in the R program. CTT, lavaan, dplyr, psych, semPlot, and foreign packages were used in performing the analyses. According to the research results, although the Self‐Regulation sub‐dimension of the Psychological Well‐Being Scale did not have a direct effect on the level of Emotion Regulation Skills, it was observed to have an indirect effect through the mediating variables of the Mother's and Father's Interest Towards School.
Self-efficacy, the conviction in one’s ability to succeed in particular tasks, is crucial in academic performance and professional development, especially in higher education programs. Although it has been widely studied in STEM disciplines, research on gender differences in self-efficacy in business education is scarce, especially in the Chilean context. This study aims to fill this gap by examining self-efficacy beliefs and their association with academic performance among business students at a Chilean university. Using a quantitative, cross-sectional design, data were collected from 239 students via the validated Academic Self-Efficacy Scale (ESAA). Confirmatory factor analysis confirmed the scale’s psychometric robustness in measuring situational and personal efficacy beliefs. No significant associations were found between self-efficacy and academic performance. However, gender differences emerged in personal efficacy beliefs, with men reporting higher levels than women, while no differences were found in situational efficacy. These findings reinforce the multidimensional nature of academic self-efficacy and reveal persistent gender disparities in how students perceive their individual academic capacities. This asymmetry may limit women’s engagement, confidence, and long-term professional development in male-dominated fields such as business. The results point to the need for gender-sensitive educational strategies that intentionally foster personal efficacy beliefs among female students, thereby promoting more equitable academic experiences and professional trajectories.
The aim of this research was to explore the concept of virtual personality. Traditional personality models were developed based on characteristics of the real (or offline) world. Given the evidence suggesting that people may behave differently in virtual environments, it is important to explore whether there is a virtual way of being: a virtual personality. To address this objective, three consecutive studies were conducted using various samples of Spanish-speaking adults from the general population. In Study 1, a list of adjectives that people reported using to describe how they and others are in virtual environments was developed. In Study 2, the list was psychometrically analyzed to generate both a theoretical model and a measurement instrument. A three-factor/traits model was obtained: authenticity, cautiousness, and agreeableness and sociability. The virtual personality model was named ACAS. In Study 3, its associations with normal traits and its predictive power over internet related behaviors were analyzed. ACAS’ traits are related to normal personality traits but do not constitute the same construct and present incremental validity when predicting internet related behaviors. This virtual personality model constitutes an initial approach and should be studied in greater depth in other samples and in longitudinal studies that analyze its temporal stability.
College students worldwide are encountering signi cant mental health challenges and positive psychological approaches are effective in addressing these issues. Psychological capital as a construct having origin from organizational psychology plays an important role in young people's education and mental health. The dimensions of psychological capital and spiritual well-being play an important role in manifesting emotional experiences, both positive and negative, among young individuals. In the contemporary scenario, where social and economic factors act as deciding factors in deciphering and in uencing emotions, there is a dire need to focus on the positive developmental aspects of human behavior. We, thus, explore the interconnection between psychological capital and spiritual well-being and quantitatively investigate how both of them might in uence the emotional experiences of college students (N = 214). Results from the regression analysis suggest that psychological capital predicts positive emotions, whereas spiritual well-being mediates this relationship. One of the practical implications of our results suggest that these three variables have good interventional potential for well-being promotion studies.
This research focused on the narrative relating to the learning process of secondary students, especially those who are less affluent. The scope of this narrative deals with the possible roles and benefits of indirect feedback to enhance secondary students’ learning. Thus, the research aimed to examine the impacts of their past experiences with indirect feedback (in interactions with people outside school), attitude, and social expectation on secondary students’ paradigm and mindset. The survey was based on the Double-loop Learning concept and Theory of Planned Behavior. A total of 1473 students from eight secondary schools located in the central Java region (Indonesia) participated in the survey. Based on the findings, two factors (i.e., attitude and indirect feedback) significantly influenced students’ paradigm and mindset on learning. Key implications to help sustain learning for secondary students include the continuous importance of an effective student–teacher relationship and the potential integration of indirect feedback from people outside school. In other words, the research suggests more classroom openness for indirect feedback when enhancing secondary students’ learning process. Finally, the limitations and future studies are discussed.
Racial and ethnic minority women with HIV experience higher levels of HIV-related stigma, have poorer adherence to antiretroviral therapy (ART), and lower viral suppression rates than men and white women with HIV. Using structural equation modeling, we examined the direct and indirect associations between race and ethnicity, ART adherence and viral suppression through HIV-related stigma dimensions (anticipated, internalized, and enacted) among 542 racial and ethnic minority women with HIV (37% Black [excluding Hispanic and Haitian]; 34% Hispanic [of any race]; 29% Haitian [of any race]) who completed a survey about women-centered HIV care. All paths from each racial and ethnic group to the mediators and outcomes were estimated in comparison to the overall sample estimates. Approximately 62% of participants were adherent to ART, and 91% were virally suppressed. Haitian ethnicity was associated with increased anticipated stigma (β = 0.33, 95% CI: 0.19, 0.47), decreased internalized stigma (β = – 0.16, 95% CI: – 0.31,– 0.02), and decreased viral suppression (β = – 0.48, 95% CI: – 0.91, – 0.15). Hispanic ethnicity was associated with increased viral suppression (β = 0.43, 95% CI: 0.13, 0.85) and decreased anticipated stigma (β = – 0.25, 95% CI: – 0.37, – 0.13). Black race was not associated with any of the variables examined. None of the HIV-related stigma dimensions had a significant mediating effect. Our findings highlight the need for in-depth qualitative research to understand the unique cultural beliefs/practices and perceptions about HIV within the Haitian population that drive HIV-related stigma and decreased viral suppression.
This study examined how three broad uncertainty-related dispositions—Affective Discomfort with Uncertainty, Cognitive Rigidity and Closure, and Need for Predictability versus Complexity—relate to individual responses to the unprecedented global uncertainty triggered by the COVID-19 pandemic. We collected data from 446 participants (61% female, ages 18–62) during Italy’s government-mandated lockdown in April 2020. Participants completed validated scales for uncertainty-related traits and a specially designed survey assessing pandemic responses, including Pandemic Anxiety and Routine Disruption—capturing fear of infection, distress from lifestyle changes, and difficulty processing conflicting information—and Proactive Compliance and Normative Trust, which reflects engagement in protective behaviors and endorsement of official guidelines. Structural equation modeling revealed that Cognitive Rigidity and Closure was the strongest predictor of Pandemic Anxiety and Routine Disruption, as well as Proactive Compliance and Normative Trust (PCNT), overshadowing the influence of Affective Discomfort with Uncertainty. Need for Predictability versus Complexity positively predicted Proactive Compliance and Normative Trust but was unrelated to Pandemic Anxiety and Routine Disruption. These findings highlight the twofold role of cognitive rigidity in amplifying emotional distress and fostering compliance during crises. The study emphasized the interplay between emotional and cognitive uncertainty-related traits in predicting pandemic-related distress and adherence to public health measures.
The exploration of opportunities is a major precondition for idea generation. Despite its relevance, theoretical approaches to innovation have paid scant attention to its conceptualization, measurement, antecedents and outcomes. To fill this void, this paper adopts a behavioural approach to delve into the micro‐foundations of this early stage of the innovation process by introducing and operationalizing the construct of individual exploratory behaviours (IEBs), which are the actions undertaken by individuals to boost idea generation through opportunity search. After reviewing earlier research on the exploratory stage of the ideation process, we propose and validate IEBs as an original multidimensional construct encompassing three behavioural components ( questioning the status quo , sensing the environment and experimenting ). Results from structural equation modelling corroborate IEBs' positive relationship with idea generation and radical innovation as outcomes and with risk‐taking and self‐confidence as antecedents. Our research adds to the literature by focusing the attention on the critical role of the beginning of the innovation process and analysing its different components according to a behavioural perspective, which is an aspect that has been neglected in favour of its cognitive/psychological dimension. In so doing, this research provides a formalization of the unstructured process of opportunity exploration and preliminary evidence of the relationship between IEBs and organizational innovative performance. The findings offer insights for both educational and organizational settings that can adopt IEBs in their assessment and development processes to boost individuals' ability to promote innovation at the organizational level.
Objective
Haircare cosmetic products are commonly reported to have a positive impact on psychological wellbeing. These effects are attributed to increased feelings of confidence and improved self‐esteem facilitated by improved hair and scalp condition. However, the causal relationship between hair and scalp health and psychological wellbeing is under‐researched. This paper reports the results of an extensive survey of haircare consumers in diverse populations using an exploratory Hair & Scalp CARE (Condition and Affective Response Evaluations) questionnaire.
Method
Participants ( N = 1184) completed an online 23‐item questionnaire designed to capture hair and scalp‐related wellbeing as an initial exploratory validation of Hair & Scalp CARE . For the analysis, the data were randomly split into 2 equal samples; Sample 1 provided the data for the initial exploratory factor analysis, and Sample 2 was used for confirmatory factor analysis. Participants also provided demographic information and completed the Sleep Health Index ( SHI ) and Perceived Stress Scale ( PSS ) to investigate sleep health and perceived stress.
Results
Factor analysis provided a one‐factor solution, explaining 55% of the variance. The final version of the Hair & Scalp CARE questionnaire consisted of 21 items. The one‐factor structure was supported by confirmatory factor analysis. Correlational analyses demonstrated that higher scores on Hair & Scalp CARE were also associated with lower PSS scores and higher SHI scores.
Conclusion
The Hair & Scalp CARE questionnaire is a valid tool for the assessment of the impact of hair and scalp condition on psychological wellbeing. The present data also suggest a relationship between hair and scalp wellbeing and other psychological wellbeing indicators, as healthier hair and scalp was also linked to lower levels of perceived stress and good sleep health. Hair & Scalp CARE could be used within a variety of further research designs to demonstrate the positive impact of cosmetic haircare products on wellbeing.
Purpose
The purpose of this study was to propose a theoretical model to predict government compliance on COVID-19 vaccination targeting Millennials and Generation Z (generation MZ) by expanding the situational theory of problem solving (STOPS) model with government trust in the COVID-19 pandemic. The presented model found that situational motivation in problem solving on COVID-19 vaccines and active communicative action in problem solving (i.e. information seeking, forfending and forwarding on vaccines) increased government compliance. In addition, public trust in government was a critical factor in generation MZ being active in following government instructions on the COVID-19 vaccine. As a post hoc analysis, results of indirect effect revealed that generation MZ is less likely to follow government instructions when their communicative behaviors are increased. Theoretical and practical implications are as follows.
Design/methodology/approach
Focusing on young adults (i.e. generation MZ) in South Korea, a pairing of two groups – Millennials (born 1981–1995) and Generation Z (born 1996–2005) ( N = 1,200) – this study applied the STOPS and trust in government to predict when and how the younger generation complies with government instructions regarding COVID-19 vaccination.
Findings
The results of this study showed that people were motivated to employ active communication behaviors to seek, forfend and forward information related to COVID-19. In addition, we found that situational motivation in problem solving, active communication behaviors and government trust were positively associated with government compliance regarding COVID-19 vaccination.
Originality/value
Many countries around the world have faced extraordinary challenges in the effort to effectively slow the spread of coronavirus disease (COVID-19) and maintain national sustainability since the emergence of COVID-19. To reduce the spread of COVID-19, governments have asked their citizens to follow health instructions during the COVID-19 pandemic. It is known that vaccination can lower an individual’s risk of contracting and spreading COVID-19. Despite the amount of research on COVID-19, few scholars have studied younger generations and their preventive behaviors, including the following of government instructions (i.e. government compliance).
Purpose
Excessive social media use has emerged as a pressing issue in recent years, prompting concerns about its adverse effects on users’ physical and mental well-being. In response to these concerns, the Chinese short video platform Douyin has implemented healthy use reminders to mitigate excessive and problematic use. This study aims to investigate the effectiveness of Douyin healthy use reminders on users’ perceptions and behaviors.
Design/methodology/approach
We conducted an online survey on Chinese college students who were Douyin users. We measured users’ perceived intrusiveness toward the reminder messages, perceived freedom threat, psychological reactance, attitudes toward the reminder messages, and their compliance behavior in the survey.
Findings
Results from structural equation modeling showed that users’ perception of the reminders’ intrusiveness negatively impacted their attitude, subsequently affecting compliance behavior. Additionally, results showed that this effect could be explained by psychological reactance theory, as perceived intrusiveness heightened users’ sense of freedom threat and psychological reactance, further influencing attitude and compliance behavior.
Implications
Our findings imply that social media platforms like Douyin should take into account user experience and acceptance of the healthy use reminders to enhance the persuasive effects. Possible strategies to enhance the communication effectiveness of the healthy use reminder messages are discussed.
Originality/value
Although recent literature has paid more attention to the problematic and excessive use of social networks, these studies primarily focused on exploring users’ dispositional traits to explain the association between excessive social media use and users’ well-being. The present study is the first to explore the psychological processes that explain the effectiveness of platform-initiated healthy use intervention. We also extend the theoretical understanding of the psychological reactance theory in social media use in a non-English context.
Advances in artificial intelligence (AI) have changed the relationships between humans and machines. AI can exhibit human‐like characteristics in its behavior, prompting humans to view it as a semi‐human. Companies embed humanlike traits in AI to facilitate interaction with consumers. As humanlike traits in AI increase, people may perceive it as eerie. Yet, there is no consensus on the effect of search modalities (text vs. voice) on humanlike traits in AI. In addition, it is not clear which humanlike traits in AI influence eeriness and purchase intention. To address the gaps, we utilized anthropomorphism theory and human psychological needs to identify why people perceive more autonomy, competence, and social presence in voice‐based AI compared with text‐based AI. We employed a quasi‐experiment (one group) consisting of 385 participants, and two experiments: pretest 2 consisted of 130 participants (two groups: experimental and control groups) and the main experiment consisted of 528 participants (two groups: experimental and control groups). We used Amazon Mechanical Turk and recruited participants in the USA who completed between 500 to 1000 HITS with a 90% approval rate. Results reveal that voice searching in AI evokes greater autonomy, competence, and social presence. Utilizing the uncanny valley effect, results showed that perceived competence lowers perceived eeriness and thereby enhances purchase intention. Although social presence enhances perceived eeriness, it increases purchase intention. In addition, perceived autonomy triggers eeriness, yet it increases consumers' purchase intention for low‐involvement products. Additionally, we examined the interaction effects of self‐construal and autonomy, competence, and social presence on eeriness. We also suggest practical implications for AI developers and research limitations, and directions for future research.
This study aims to explore the relationship between LGBTQ+ culture in video games and brand resonance, aiming to understand to what extent the use of LGBTQ+ cultural assets can affect brand awareness, quality, and loyalty between LGBTQ+ and non-LGBTQ+ consumers. An online survey collected data from 488 people, and a proposed model was tested using SEM multigroup analysis. Findings show that the impact of the “LGBTQ+ cultural assets” differs between LGBTQ+ and non-LGBTQ+ consumers. The results of this study suggest the LGBTQ+ culture in a video game is the key predictor of brand loyalty among LGTBQ+ consumers.
Monte Carlo computer simulations were used to investigate the performance of three χ–2 test statistics in confirmatory factor analysis (CFA). Normal theory maximum likelihood χ–2 (ML), Browne's asymptotic distribution free χ–2 (ADF), and the Satorra-Bentler rescaled χ–2 (SB) were examined under varying conditions of sample size, model specification, and multivariate distribution. For properly specified models, ML and SB showed no evidence of bias under normal distributions across all sample sizes, whereas ADF was biased at all but the largest sample sizes. ML was increasingly overestimated with increasing nonnormality, but both SB (at all sample sizes) and ADF (only at large sample sizes) showed no evidence of bias. For misspecified models, ML was again inflated with increasing nonnormality, but both SB and ADF were underestimated with increasing nonnormality. It appears that the power of the SB and ADF test statistics to detect a model misspecification is attenuated given nonnormally distributed data. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Covariance structure analysis uses χ–2 goodness-of-fit test statistics whose adequacy is not known. Scientific conclusions based on models may be distorted when researchers violate sample size, variate independence, and distributional assumptions. The behavior of 6 test statistics was evaluated with a Monte Carlo confirmatory factor analysis study. The tests performed dramatically differently under 7 distributional conditions at 6 sample sizes. Two normal-theory tests worked well under some conditions but completely broke down under other conditions. A test that permits homogeneous nonzero kurtoses performed variably. A test that permits heterogeneous marginal kurtoses performed better. A distribution-free test performed spectacularly badly in all conditions at all but the largest sample sizes. The Santorra-Bentler scaled test performed best overall. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
This article compares two structural equation modeling fit indexes—Bentler's ( 1990; Bentler & Bonett, 1980) Confirmatory Fit Index (CFI) and Steiger and Lind's (1980; Browne & Cudeck, 1993) Root Mean Square Error of Approximation (RMSEA). These two fit indexes are both conceptually linked to the noncentral chi‐square distribution, but CFI has seen much wider use in applied research, whereas RMSEA has only recently been gaining attention. The article suggests that use of CFI is problematic because of its baseline model. CFI seems to be appropriate in more exploratory contexts, whereas RMSEA is appropriate in more confirmatory contexts. On the other hand, CFI does have an established parsimony adjustment, although the adjustment included in RMSEA may be inadequate.
Testing fit is a critical and controversial step in structural equation modeling (SEM), and alternative fit indices have been proposed. To provide guidance, a simulation study was conducted to evaluate the effects of estimation method, number of indicators per factor (p/r ratio), sample size, and loading size on six SEM fit indices: chi-square/degree of freedom ratio (chi(2)/df), Normed Fit Index (NFI), Nonnormed Fit Index (NNFI), Centrality m index, Relative Noncentrality Index (RNI), and Comparative Fit Index (CFI). When improper solutions occurred, the effects of constraining out-of-bounds estimates to a reasonable range (the method used in EQS software) on the fit indices was evaluated. Four levels of sample size (50, 100, 200, and 500), three levels of loading size (0.50, 0.70, and 0.90), five levels of p/r ratio (2, 3, 4, 5, and 6), and two levels of estimation method (Generalized Least Squares [GLS] and Maximum Likelihood [ML]) were examined. The results of this study indicated that: (a) improper solutions occurred frequently when p/r = 2 and sample size was small or loadings were low; (b) GLS produced more improper solutions than ML in general; (c) there was no effect of improper solutions (constrained to some boundary) found for any of the fit indices but NFI (more downward bias occurred for NFI when improper solutions were present); (d) four incremental fit indices (NFI, NNFI, RNI, and CFI) were negatively affected by increasing the p/r ratio; (e) NFI was affected much more seriously than the other three fit indices and should not be used; (f) all fit indices except NNFI were found to be significantly affected by estimation method (less bias occurred for GLS than for ML); and (g) interaction effects between estimation method, p/r ratio, sample size, and loading size also occurred. Recommendations regarding selection of a fit index are made based on the findings.
Anumber of goodness-of-fit indices for the evaluation of multivariate structural models are expressed as functions of the noncentrality parameter in order to elucidate their mathematical properties and, in particular, to explain previous numerical findings. Most of the indices considered are shown to vary systematically with sample size. It is suggested that H. Akaike's (1974; see record
1989-17660-001) information criterion cannot be used for model selection in real applications and that there are problems attending the definition of parsimonious fit indices. A normed function of the noncentrality parameter is recommended as an unbiased absolute goodness-of-fit index, and the Tucker–Lewis (see record
1973-30255-001) index and a new unbiased counterpart of the Bentler–Bonett (see record
1981-06898-001) index are recommended for those investigators who might wish to evaluate fit relative to a null model. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
P. M. Bentler and D. G. Bonett (1980) argue that it is often useful to compare a hypothesized covariance structure model or set of models to a nested null model using fit coefficients and they propose both generic null models for a variety of cases and 2 new measures of fit
extends the work of Bentler and Bonett in 2 ways / 1st, we provide general analytic conditions for ascertaining whether their generic null models are nested under a substantive model of interest, an issue they do not address clearly and completely / 2nd, we show that the null models they propose are inappropriate in all but the purely exploratory case / in other cases, we argue that the comparison should be developed in terms of baseline models that reflect the state of prior theory and knowledge, unlike the null models of Bentler and Bonett (PsycINFO Database Record (c) 2012 APA, all rights reserved)
[Correction Notice: An erratum for this article was reported in Vol 75(1) of
Journal of Applied Psychology (see record
2008-10492-001). An error exists in Figure 2 and the accompanying text of the article. The corrected information is included in the erratum.] The problem of assessing fit of structural equation models is reviewed, and two sampling studies are reported that examine the effects of sample size, estimation method, and model misspecification on fit indices. In the first study, the behavior of indices in a known-population confirmatory factor analysis model is considered. In the second study, the same problem in an empirical data set is examined by looking at antecedents and consequences of work motivation. The findings across the two studies suggest that (a) as might be expected, sample size is an important determinant in assessing model fit; (b) estimator-specific, as opposed to estimator-general, fit indices provide more accurate indications of model fit; and (c) the studied fit indices are differentially sensitive to model misspecification. Some recommendations for the use of structural equation model fit indices are given. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Factor analysis, path analysis, structural equation modeling, and related multivariate statistical methods are based on maximum likelihood or generalized least squares estimation developed for covariance structure models (CSMs). Large-sample theory provides a chi-square goodness-of-fit test for comparing a model (M) against a general alternative M based on correlated variables. It is suggested that this comparison is insufficient for M evaluation. A general null M based on modified independence among variables is proposed as an additional reference point for the statistical and scientific evaluation of CSMs. Use of the null M in the context of a procedure that sequentially evaluates the statistical necessity of various sets of parameters places statistical methods in covariance structure analysis into a more complete framework. The concepts of ideal Ms and pseudo chi-square tests are introduced, and their roles in hypothesis testing are developed. The importance of supplementing statistical evaluation with incremental fit indices associated with the comparison of hierarchical Ms is also emphasized. Normed and nonnormed fit indices are developed and illustrated. (43 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
In confirmatory factor analysis, hypothesized models reflect approximations to reality so that any model can be rejected if the sample size is large enough. The appropriate question is whether the fit is adequate to support the model, and a large number of fit indexes have been proposed for this purpose. In the present article, we examine the influence of sample size on different fit indexes for both real and simulated data. Contrary to claims by Bentler and Bonett (1980), their incremental fit index was substantially affected by sample size. Contrary to claims by Joreskog and Sorbom (1981), their goodness-of-fit indexes provided by LISREL were substantially affected by sample size. Contrary to claims by Bollen (1986), his new incremental fit index was substantially affected by sample size. Hoelter's (1983) critical N index was also substantially affected by sample size. Of the more than 30 indexes considered, the Tucker-Lewis (1973) index was the only widely used index that was relatively independent of sample size. However, four new indexes based on the same form as the Tucker-Lewis index were also relatively independent of sample size., (C) 1988 by the American Psychological Association <2>
Bentler and Bonett's nonnormed fit index is a widely used measure of goodness of fit for the analysis of covariance structures. This note shows that contrary to what has been claimed the nonnormed fit index is dependent on sample size. Specifically for a constant value of a fitting function, the nonnormed index is inversely related to sample size. A simple alternative fit measure is proposed that removes this dependency. In addition, it is shown that this new measure as well as the old nonnormed fit index can be applied to any fitting function that measures the deviation of the observed covariance matrix from the covariance matrix implied by the parameter estimates for a model.
We describe a general procedure by which any number of parameters of the factor analytic model can be held fixed at any values and the remaining free parameters estimated by the maximum likelihood method. The generality of the approach makes it possible to deal with all kinds of solutions: orthogonal, oblique and various mixtures of these. By choosing the fixed parameters appropriately, factors can be defined to have desired properties and make subsequent rotation unnecessary. The goodness of fit of the maximum likelihood solution under the hypothesis represented by the fixed parameters is tested by a large samplex
2 test based on the likelihood ratio technique. A by-product of the procedure is an estimate of the variance-covariance matrix of the estimated parameters. From this, approximate confidence intervals for the parameters can be obtained. Several examples illustrating the usefulness of the procedure are given.
Normed and nonnormed fit indexes are frequently used as adjuncts to chi-square statistics for evaluating the fit of a structural model. A drawback of existing indexes is that they estimate no known population parameters. A new coefficient is proposed to summarize the relative reduction in the noncentrality parameters of two nested models. Two estimators of the coefficient yield new normed (CFI) and nonnormed (FI) fit indexes. CFI avoids the underestimation of fit often noted in small samples for Bentler and Bonett's (1980) normed fit index (NFI). FI is a linear function of Bentler and Bonett's non-normed fit index (NNFI) that avoids the extreme underestimation and overestimation often found in NNFI. Asymptotically, CFI, FI, NFI, and a new index developed by Bollen are equivalent measures of comparative fit, whereas NNFI measures relative fit by comparing noncentrality per degree of freedom. All of the indexes are generalized to permit use of Wald and Lagrange multiplier statistics. An example illustrates the behavior of these indexes under conditions of correct specification and misspecification. The new fit indexes perform very well at all sample sizes.
Assessing overall model fit is an important problem in general structural equation models. One of the most widely used fit measures is Bentler and Bonett's (1980) normed index. This article has three purposes: (1) to propose a new incremental fit measure that provides an adjustment to the normed index for sample size and degrees of freedom, (2) to explain the relation between this new fit measure and the other ones, and (3) to illustrate its properties with an empirical example and a Monte Carlo simulation. The simulation suggests that the mean of the sampling distribution of the new fit measure stays at about one for different sample sizes whereas that for the normed fit index increases with N. In addition, the standard deviation of the new measure is relatively low compared to some other measures (e.g., Tucker and Lewis's (1973) and Bentler and Bonett's (1980) nonnormed index). The empirical example suggests that the new fit measure is relatively stable for the same model in different samples. In sum, it appears that the new incremental measure is a useful complement to the existing fit measures.
Marsh and Balla (1986) and Marsh, Balla, and McDonald (1988) proposed an index of fit called χI2, but McDonald and Marsh (1990) subsequently demonstrated that the index is biased and recommended that it not be used. Bollen (1989) independently proposed Δ2 which is the same as χI2 (hereafter referred to as χI2‐Δ2), indicating that it adjusts for sample size and degrees of freedom (df). Gerbing and Anderson (1992), apparently based on the assumption that the χI2‐Δ2 index is unbiased and appropriately corrects for df (penalizes a lack of parsimony), recommended its use, and the index is routinely presented by major computer programs (e.g., EQS and LISREL 8). However, a more critical evaluation of the χI2‐Δ2 index reveals that: (a) it is systematically biased (i.e., its value varies systematically with N) although the size of the bias may be small; (b) the adjustment for df is inappropriate in that it penalizes model parsimony instead of model complexity; and (c) the inappropriate penalty for model parsimony is larger for small N. Because of these undesirable properties, the χI2‐Δ2 index is not recommended for routine use.
Many mechanistic rules of thumb for evaluating the goodness of fit of structural equation models (SEM) emphasize model parsimony; all other things being equal, a simpler, more parsimonious model with fewer estimated parameters is better than a more complex model. Although this is usually good advice, in the present article a heuristic counterexample is demonstrated in which parsimony as typically operationalized in indices of fit may be undesirable. Specifically, in simplex models of longitudinal data, the failure to include correlated uniquenesses relating the same indicators administered on different occasions will typically lead to systematically inflated estimates of stability. Although simplex models with correlated uniquenesses are substantially less parsimonious and may be unacceptable according to mechanistic decision rules that penalize model complexity, it can be argued a priori that these additional parameter estimates should be included. Simulated data are used to support this claim and to evaluate the behavior of a variety of fit indices and decision rules. The results demonstrate the validity of Bollen and Long's (1993) conclusion that "test statistics and fit indices are very beneficial, but they are no replacement for sound judgment and substantive expertise" (p. 8).
This article reviews proposed goodness-of-fit indices for structural equation models and the Monte Carlo studies that have empirically assessed their distributional properties. The cumulative contributions of the studies are summarized, and the variables under which the indices are studied are noted. A primary finding is that many of the indices used until the late 1980s, including Joreskog and Sorbom's (1981) GFI and Bentler and Bonett's (1980) NFI, indicated better fit when sample size increased. More recently developed indices based on the chi-square noncentrality parameter are discussed and the relevant Monte Carlo studies reviewed. Although a more complete understanding of their properties and suitability requires further research, the recommended fit indices are the McDonald (1989) noncentrality index, the Bentler (1990)-McDonald and Marsh (1990) RNI (or the bounded counterpart CFI), and Bollen's (1989) DELTA2.
A major problem encountered in covariance structure analyses involves decisions concerning whether or not a given theoretical model adequately represents the data used for its assessment. Given that X2 goodness-of-fit tests are joint functions of the difference between theoretical and empirical covariance structures and sample size, gauging the impact of sample size on such tests is essential. In this paper, we propose a simple index (critical N) and tentative acceptance criterion, which, by focusing on sample size, provide an improved method for assessing goodness-of-fit. Both small- and large-sample examples are presented, illustrating the utility of the proposed method for assessing theoretical models.
The large-sample statistical theory for latent-variable structural equation models offers little solace to the developmental psychologist, who is often confronted with less than optimally large sample sizes. This article reviews previously proposed alternatives to the sample-size and goodness-of-fit issue in latent-variable structural equation models. Various nonparametric fit indices for latent-variable systems are reviewed with their strengths and weaknesses discussed. An alternative estimation strategy called ME2 estimation is introduced as a possible alternative solution to the small-sample problem.
This article considers single sample approximations for the cross-validation coefficient in the analysis of covariance structures. An adjustment for predictive validity which may be employed in conjunction with any correctly specified discrepancy function is suggested. In the case of maximum likelihood estimation under normality assumptions the coefficient obtained is a simple linear function of the Akaike Information Criterion. Results of a random sampling experiment are reported.
This study examined the role language plays in mediating the influence of verbal descriptions of persons on trait ratings of those persons. Subjects were given written descriptions of the behavior of fictitious persons in a work situation and were asked to rate them on fifteen trait- adjective scales. In one condition of the experiment, specific information about certain traits was withheld, forcing subjects to rate persons on traits for which they had no direct behavioral clues. In the other two conditions, the specific information was provided. Providing specific information about a trait directly influenced ratings on that trait even when sufficient general information on that trait was given. In one condition, the influence on the ratings of the additional behavioral clues was such that a new latent variable representing an additional component of meaning was called for in the structural equation model.
This paper examines methods for comparing the suitability of alternative models for covariance matrices. A cross-validation procedure is suggested and its properties are examined. To motivate the discussion, a series of examples is presented using longitudinal data.
In a typical study involving covariance structure modeling, fit of a model or a set of alternative models is evaluated using several indicators of fit under one estimation method, usually maximum likelihood. This study examined the stability across estimation methods of incremental and non incremental fit measures that use the information about the fit of the most restricted (null) model as a reference point in assessing the fit of a more substantive model to the data. A set of alternative models for a large empirical dataset was analyzed by asymptotically distribution-free, generalized least squares, maximum likelihood, and ordinary least squares estimation methods. Four incremental and four nonincremental fit indexes were com pared. Incremental indexes were quite unstable across estimation methods—maximum likelihood and ordinary least squares solutions indicated better fit of a given model than asymptotically distribution-free and generalized least squares solu tions. The cause of this phenomenon is explained and illustrated, and implications and recommenda tions for practice are discussed.
Let S represent the usual unbiased estimator of a covariance matrix, Σ0, whose elements are functions of a parameter vector . A generalized least squares (G.L.S) estimate, of may be obtained by minimizing where V is some positive definite matrix. Asymptotic properties of the G.L.S. estimators are investigated assuming only that satisfies certain regularity conditions and that the limiting distribution of S is multivariate normal with specified parameters. The estimator of which is obtained by maximizing the Wishart likelihood function (M.W.L. estimator) is shown to be a member of the class of G.L.S. estimators with minimum asymptotic variances. When is linear in a G.L.S. estimator which converges stochastically to the M.W.L. estimator involves far less computation. Methods for calculating estimates of , estimates of the dispersion matrix of , and test statistics, are given for certain linear models.
J. S. Tanaka and G. J. Huba (see record
1986-10882-001) introduced a general fit index for covariance structure models under generalized least squares estimation that in some cases specialized to the fit indices presented by K. G. Jöreskog and D. Sörbom (1981). For a wide class of models, the general form of this fit index can be expressed as a weighted coefficient of determination. This coefficient is given as the ratio of weighted trace functions of predicted and observed covariance matrix elements. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
consider the following issues: (a) the usefulness of the χ[superscript]2 statistic based on various estimation methods for model evaluation and selection; (b) the conceptual elaboration of and selection criteria for fit indexes; and (c) identifying some crucial factors that will affect the magnitude of χ[superscript]2 statistics and fit indexes / review previous research findings as well as report results of some new, unpublished research (PsycINFO Database Record (c) 2012 APA, all rights reserved)
The structure of the covariance matrix of sample covariances under the class of linear latent variate models is derived, using properties of cumulants. This derivation is employed to provide a general framework for robustness of statistical inference in the analysis of covariance structures arising from linear latent variate models. Conditions for normal theory estimators and test statistics to retain each of their usual asymptotic properties under nonnormality of latent variates are given. Factor analysis and LISREL analysis are discussed as examples. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
A general fit index for covariance structure models is obtained for all estimators that can be considered under generalized least squares (GLS) approaches including asymptotically efficient, robust, and resistant methods of estimation. This fit index is expressed as a function of the ratio of two trace functions. Normal theory ordinary least squares (OLS) and maximum likelihood (ML) fit indices previously given by Jöreskog and Sörbom can be derived from this framework. Fit indices for normal theory and generic GLS approaches including robust/resistant estimation methods are also obtained.
A condition is given by which optimal normal theory methods, such as the maximum likelihood methods, are robust against violation of the normality assumption in a general linear structural equation model. Specifically, the estimators and the goodness of fit test are robust. The estimator is efficient within some defined class, and its standard errors can be obtained by a correction formula applied to the inverse of the information matrix. Some special models, like the factor analysis model and path models, are discussed in more detail. A method for evaluating the robustness condition is given.
The purpose of the present investigation is to examine the influence of sample size (N) and model parsimony on a set of 22 goodness-of-fit indices including those typically used in confirmatory factor analysis and some recently developed indices. For sample data simulated from two known population data structures, values for 6 of 22 fit indices were reasonably independent ofN and were not significantly affected by estimating parameters known to have zero values in the population: two indices based on noncentrality described by McDonald (1989; McDonald and Marsh, 1990), a relative (incremental) index based on noncentrality (Bentler, 1990; McDonald & Marsh, 1990), unbiased estimates of LISREL's GFI and AGFI (Joreskog & Sorbom, 1981) presented by Steiger (1989, 1990) that are based on noncentrality, and the widely known relative index developed by Tucker and Lewis (1973). Penalties for model complexity designed to control for sampling fluctuations and to address the inevitable compromise between goodness of fit and model parsimony were evaluated.
In covariance structure analysis, alternative methods of estimation are now regularly available. A variety of statistics, such as estimators, test statistics, and residuals, are computed. The sampling variability of these statistics is known to depend on a matrix Γ which is based on the fourth-order moments of the data. Estimates of these fourth-order moments are expensive to compute, require a lot of computer storage, and have high sampling variability in small to moderate samples. By exploiting the linear relations that typically generate the covariance structure, we have developed conditions under which a matrix Γ*, which depends only on second-order moments of the data, can be used as a substitute for Γ to obtain correct asymptotic distributions for the statistics of interest. In contrast to related work on asymptotic robustness in covariance structure analysis, our theory is developed in the general setting of arbitrary discrepancy functions and addresses a broader class of statistics that include, for instance, goodness of fit statistics that are not necessarily asymptotically χ2 distributed, and statistics based on the residuals. Basically, our theory shows that the normal theory form Γ* for Γ can be used whenever an independence assumption (not only uncorrelatedness), which will always hold under normality, carries over to the model with nonnormal variables. This theory is spelled out in sufficient detail and simplicity so that it can be used in every day practice.
A framework for hypothesis testing and power analysis in the assessment of fit of covariance structure models is presented. We emphasize the value of confidence intervals for fit indices, and we stress the relationship of confidence intervals to a framework for hypothesis testing. The approach allows for testing null hypotheses of not-good fit, reversing the role of the null hypothesis in conventional tests of model fit, so that a significant result provides strong support for good fit. The approach also allows for direct estimation of power, where effect size is defined in terms of a null and alternative value of the root-mean-square error of approximation fit index proposed by J. H. Steiger and J. M. Lind (1980). It is also feasible to determine minimum sample size required to achieve a given level of power for any test of fit in this framework. Computer programs and examples are provided for power analyses and calculation of minimum sample sizes., (C) 1996 by the American Psychological Association
Asymptotic properties of estimators for the confirmatory factor analysis model are discussed. The model is identified by restrictions on the elements of the factor loading matrix; the number of restrictions may exceed that required for identification. It is shown that a particular centering of the maximum likelihood estimator derived under assumed normality of observations yields an asymptotic normal distribution that is common to a wide class of distributions of the factor vectors and error vectors. In particular, the asymptotic covariance matrix of the factor loading estimator derived under the normal assumption is shown to be valid for the factor vectors containing a fixed part and a random part with any distribution having finite second moments and for the error vectors consisting of independent components with any distributions having finite second moments. Thus the asymptotic standard errors of the factor loading estimators computed by standard computer packages are valid for virtually any type of nonnormal factor analysis. The results are extended to certain structural equation models.
Three types of asymptotic goodness-of-fit tests derived under the normal assumption have been used widely in factor analysis. Asymptotic behavior of the test statistics is investigated here for the factor analysis model with linearly or nonlinearly restricted factor loadings under weak assumptions on the factor vector and the error vector. In particular the limiting result for the three tests is shown to hold for the factor vector, either fixed or random with any distribution having finite second-order moments, and for the error vector with any distribution having finite second-order moments, provided that the components of the error vector are independent, not just uncorrelated. As special cases the result holds for exploratory and confirmatory factor analysis models and for certain nonnormal structural equation (LISREL) models.
A class of latent variable models which includes the unrestricted factor analysis model is considered. It is shown that minimum discrepancy test statistics and estimators derived under normality assumptions retain their asymptotic properties when the common factors are not normally distributed but the unique factors do have a multivariate normal distribution. The minimum discrepancy test statistics and estimators considered include the usual likelihood ratio test statistic and maximum likelihood estimators.
The information criterion AIC was introduced to extend the method of maximum likelihood to the multimodel situation. It was obtained by relating the successful experience of the order determination of an autoregressive model to the determination of the number of factors in the maximum likelihood factor analysis. The use of the AIC criterion in the factor analysis is particularly interesting when it is viewed as the choice of a Bayesian model. This observation shows that the area of application of AIC can be much wider than the conventional i.i.d. type models on which the original derivation of the criterion was based. The observation of the Bayesian structure of the factor analysis model leads us to the handling of the problem of improper solution by introducing a natural prior distribution of factor loadings.
A procedure for computing the power of the likelihood ratio test used in the context of covariance structure analysis is derived. The procedure uses statistics associated with the standard output of the computer programs commonly used and assumes that a specific alternative value of the parameter vector is specified. Using the noncentral Chi-square distribution, the power of the test is approximated by the asymptotic one for a sequence of local alternatives. The procedure is illustrated by an example. A Monte Carlo experiment also shows how good the approximation is for a specific case.
Current practice in structural modeling of observed continuous random variables is limited to representation systems for first and second moments (e.g., means and covariances), and to distribution theory based on multivariate normality. In psychometrics the multinormality assumption is often incorrect, so that statistical tests on parameters, or model goodness of fit, will frequently be incorrect as well. It is shown that higher order product moments yield important structural information when the distribution of variables is arbitrary. Structural representations are developed for generalizations of the Bentler-Weeks, Jöreskog-Keesling-Wiley, and factor analytic models. Some asymptotically distribution-free efficient estimators for such arbitrary structural models are developed. Limited information estimators are obtained as well. The special case of elliptical distributions that allow nonzero but equal kurtoses for variables is discussed in some detail. The argument is made that multivariate normal theory for covariance structure models should be abandoned in favor of elliptical theory, which is only slightly more difficult to apply in practice but specializes to the traditional case when normality holds. Many open research areas are described.
Maximum likelihood factor analysis provides an effective method for estimation of factor matrices and a useful test statistic
in the likelihood ratio for rejection of overly simple factor models. A reliability coefficient is proposed to indicate quality
of representation of interrelations among attributes in a battery by a maximum likelihood factor analysis. Usually, for a
large sample of individuals or objects, the likelihood ratio statistic could indicate that an otherwise acceptable factor
model does not exactly represent the interrelations among the attributes for a population. The reliability coefficient could
indicate a very close representation in this case and be a better indication as to whether to accept or reject the factor
solution.
Akaike's Information Criterion is systematically dependent on sample size, and therefore cannot be used in practice as a basis for model selection. An alternative measure of goodness-of-fit, based like Akaike's on the noncentrality parameter, appears to be consistent over variations in sample size.
Factor analysis, path analysis, structural equation modeling, and related multivariate statistical methods are based on maximum likelihood or generalized least squares estimation developed for covariance structure models. Large-sample theory provides a chi-square goodness-of-fit test for comparing a model against a general alternative model based on correlated variables. This model comparison is insufficient for model evaluation: In large samples virtually any model tends to be rejected as inadequate, and in small samples various competing models, if evaluated, might be equally acceptable. A general null model based on modified independence among variables is proposed to provide an additional reference point for the statistical and scientific evaluation of covariance structure models. Use of the null model in the context of a procedure that sequentially evaluates the statistical necessity of various sets of parameters places statistical methods in covariance structure analysis into a more complete framework. The concepts of ideal models and pseudo chi-square tests are introduced, and their roles in hypothesis testing are developed. The importance of supplementing statistical evaluation with incremental fit indices associated with the comparison of hierarchical models is also emphasized. Normed and nonnormed fit indices are developed and illustrated.
Covariance structure analysis uses chi 2 goodness-of-fit test statistics whose adequacy is not known. Scientific conclusions based on models may be distorted when researchers violate sample size, variate independence, and distributional assumptions. The behavior of 6 test statistics is evaluated with a Monte Carlo confirmatory factor analysis study. The tests performed dramatically differently under 7 distributional conditions at 6 sample sizes. Two normal-theory tests worked well under some conditions but completely broke down under other conditions. A test that permits homogeneous nonzero kurtoses performed variably. A test that permits heterogeneous marginal kurtoses performed better. A distribution-free test performed spectacularly badly in all conditions at all but the largest sample sizes. The Satorra-Bentler scaled test statistic performed best overall.
Although covariance structure analysis is used increasingly to analyze nonexperimental data, important statistical requirements for its proper use are frequently ignored. Valid conclusions about the adequacy of a model as an acceptable representation of data, which are based on goodness-of-fit test statistics and standard errors of parameter estimates, rely on the model estimation procedure being appropriate for the data. Using analogies to linear regression and anova, this review examines conditions under which conclusions drawn from various estimation methods will be correct and the consequences of ignoring these conditions. A distinction is made between estimation methods that are either correctly or incorrectly specified for the distribution of data being analyzed, and it is shown that valid conclusions are possible even under misspecification. A brief example illustrates the ideas. Internet access is given to a computer code for several methods that are not available in programs such as EQS or LISREL.
Statistically based tests for the number of common factors. Pa-per presented at the annual meeting of the Psychometric Society Effect of estimation method on incremental fit indexes for covariance structure models
- J H Steiger
- J C Lind
Steiger, J. H., & Lind, J. C. (1980, May). Statistically based tests for the number of common factors. Pa-per presented at the annual meeting of the Psychometric Society, Iowa City, IA. Sugawara H. M., & MacCallum, R. C. (1993). Effect of estimation method on incremental fit indexes for covariance structure models. Applied Psychological Measurement, 17, 365-377.
An evaluation of incremental fit indices: A clarification of mathematical and empirical properties Ad-vanced structural equation modeling: Issues and techniques Goodness-of-fit indices in confirmatory factor analysis: Effects of sample size
- H W Marsh
- J R Balla
- K.-T H W Hau
- J R Balla
- R P Mcdonald
Marsh, H. W., Balla, J. R., & Hau, K.-T. (1996). An evaluation of incremental fit indices: A clarification of mathematical and empirical properties. In G. A. Marcoulides & R. E. Schumacker (Eds.), Ad-vanced structural equation modeling: Issues and techniques (pp. 315-353). Mahwah, NI: Lawrence Erlbaum Associates, Inc. Marsh, H. W., Balla, J. R., & McDonald, R. P. (1988). Goodness-of-fit indices in confirmatory factor analysis: Effects of sample size. Psychological Bulletin, 103, 391-411.
EQS for Windows user's guide. Encino, CA: Multivariate Software Sample size and Bentler and Bonett's nonnormed fit index
- P M Bentler
- E J C Wu
Bentler, P. M, & Wu, E. J. C. (1995). EQS for Windows user's guide. Encino, CA: Multivariate Software. Bollen, K. A. (1986). Sample size and Bentler and Bonett's nonnormed fit index. Psychometrika, 51, 375-377.
USREL V: Analysis of linear structural relationships by the method of maximum likelihood
- K G Jöreskog
- D Sörbom
Jöreskog, K. G., & Sörbom, D. (1981). USREL V: Analysis of linear structural relationships by the method of maximum likelihood. Chicago: National Educational Resources.