scholarly journals Comparison of confirmatory factor analysis estimation methods on mixed-format data

2021 ◽  
pp. 21-37
Author(s):  
Abdullah Faruk KILIÇ ◽  
Nuri DOĞAN
Keyword(s):  
Factor Analysis ◽  
Confirmatory Factor Analysis ◽  
Estimation Methods ◽  
Confirmatory Factor ◽  
Mixed Format

Estimation methods for multivariate Tobit confirmatory factor analysis

2014 ◽  
Vol 79 ◽  
pp. 248-260 ◽  
Author(s):  
D.R. Costa ◽  
V.H. Lachos ◽  
J.L. Bazan ◽  
C.L.N. Azevedo
Keyword(s):  
Factor Analysis ◽  
Confirmatory Factor Analysis ◽  
Estimation Methods ◽  
Confirmatory Factor

scholarly journals Confirmatory Factor Analysis of Ordinal Variables: A Simulation Study Comparing the Main Estimation Methods

2018 ◽  
Vol 36 (3) ◽  
pp. 601 ◽  
Author(s):  
Francisco Pablo Holgado-Tello ◽  
Mª Ángeles Morata-Ramirez ◽  
Maria Isabel Barbero García
Keyword(s):  
Factor Analysis ◽  
Confirmatory Factor Analysis ◽  
Simulation Study ◽  
Estimation Methods ◽  
Ordinal Variables ◽  
Confirmatory Factor

Para obtener evidencias sobre la validez de constructo a través de Análisis Factorial Confirmatorio, ha sido habitual tratar las escalas tipo Likert como si fueran variables continuas medidas según una escala de intervalo. Por tanto, el método de estimación de Máxima Verosimilitud ha sido  ampliamente aplicado, pero a su vez esto implica problemas en torno a las correlaciones de Pearson y la asimetría de la distribución de respuestas a los ítems. En este estudio de simulación analizamos —a través de χ2, del error tipo I y de la potencia— modelos bien y mal especificados comparando cinco métodos de estimación (Máxima Verosimilitud —ml—, Máxima Verosimilitud Robusta —rml—, Mínimos Cuadrados Ponderados —mls—, Mínimos Cuadrados no Ponderados —uls— y Mínimos Cuadrados no Ponderados Robustos —ruls—) en relación con las características de los modelos: número de factores, número de categorías de respuesta, asimetría de los ítems y tamaño muestral. Aconsejamos usar el método ruls de estimación, en el cual están implicadas las correlaciones policóricas.Palabras clave: análisis Factorial Confirmatorio, escalas tipo Likert, métodos de estimación, error tipo I, potencia.

Evidence-based subtypes in functional dyspepsia with confirmatory factor analysis: Psychosocial and physiopathological correlates

2001 ◽  
Vol 120 (5) ◽  
pp. A51-A52 ◽  
Author(s):  
B FISCHLER ◽  
J VANDENBERGHE ◽  
P PERSOONS ◽  
V GUCHT ◽  
D BROEKAERT ◽  
...  
Keyword(s):  
Factor Analysis ◽  
Confirmatory Factor Analysis ◽  
Functional Dyspepsia ◽  
Evidence Based ◽  
Confirmatory Factor

The French Version of the Revised Child Anxiety and Depression Scale (RCADS) in a Nonclinical Sample

2015 ◽  
Vol 74 (3) ◽  
pp. 119-127 ◽  
Author(s):  
Martine Bouvard ◽  
Anne Denis ◽  
Jean-Luc Roulin
Keyword(s):  
Factor Analysis ◽  
Anxiety Disorder ◽  
Confirmatory Factor Analysis ◽  
Generalized Anxiety Disorder ◽  
Child Anxiety ◽  
Depression Scale ◽  
Anxiety And Depression ◽  
Generalized Anxiety ◽  
French Version ◽  
Confirmatory Factor

This article investigates the psychometric properties of the Revised Child Anxiety and Depression Scale (RCADS). A group of 704 adolescents completed the questionnaires in their classrooms. This study examines potential confirmatory factor analysis factor models of the RCADS as well as the relationships between the RCADS and the Screen for Child Anxiety Related Emotional Disorders-Revised (SCARED-R). A subsample of 595 adolescents also completed an anxiety questionnaire (Fear Survey Schedule for Children-Revised, FSSC-R) and a depression questionnaire (Center for Epidemiological Studies Depression Scale, CES-D). Confirmatory factor analysis of the RCADS suggests that the 6-factor model reasonably fits the data. All subscales were positively intercorrelated, with rs varying between .48 (generalized anxiety disorder-major depression disorder) and .65 (generalized anxiety disorder-social phobia/obsessive-compulsive disorder). The RCADS total score and all the RCADS scales were found to have good internal consistency (> .70). The correlations between the RCADS subscales and their SCARED-R counterparts are generally substantial. Convergent validity was found with the FSSC-R and the CES-D. The study included normal adolescents aged 10 to 19. Therefore, the findings cannot be extended to children under 10, nor to a clinical population. Altogether, the French version of the RCADS showed reasonable psychometric properties.

Sign in / Sign up

Export Citation Format

Share Document