Estimación de la magnitud del efecto en invarianza de medición

Abstract

RESUMENEste trabajo presenta un programa en MS Excel ® para evaluar la magnitud del efecto (ES, por las siglas en inglés) en invarianza de medición de diferentes parámetros como las cargas factoriales, interceptos/thresholds y residuales, con base en estadísticos estandarizados ya conocidos. El funcionamiento del programa y la interpretación de las medidas de ES se ejemplificaron con datos empíricos.Palabras clave: invarianza de medición; magnitud del efecto; programa. RESUMO -Avaliação do tamanho do efeito na invariância de medidaEste artigo apresenta um programa no MS Excel ® para avaliar o tamanho do efeito (ES, pela sigla em inglês) na invariância de medida de diferentes parâmetros dos itens, como cargas fatoriais, interceptos/thresholds e resíduos, com base em estatísticas padronizadas bem conhecidas. O funcionamento do programa e interpretação das medidas de ES é exemplificado com dados empíricos. Palavras-chave: invariância de medida, tamanho do efeito, programa. ABSTRACT -Estimation of effect size in measurement invarianceThis paper presents a MS Excel ® program for estimating effect size (ES) in the measurement invariance of different item parameters, such as the factor loadings, intercepts/thresholds and residuals, based in well-known standardized statistics. Examples with real data are provided for the functioning of the program and the interpretation of the ES measures.

Psychometric Validation of the Adult Prosocialness Behavior Scale in a Professional Teaching Context

“…Third, the factorial invariance of the APBS was examined based on gender. The results showed that the APBS measures prosociality in an equivalent way between men and women, with both the variation of indices of fit provided by the general approach [ 104 , 118 , 119 ] and the specific one that measured the effect size of the individual parameter differences [ 120 , 121 , 122 ]. This study considered factor loadings (weak invariance), thresholds (strong invariance), and residuals (strict invariance), which supports hypothesis 3 (H3).…”

“…This way, there is unfavorable evidence for the measurement invariance if ΔCFI < −0.01 and ΔRMSEA ≥ 0.01 [ 104 ] or if ΔCFI < −0.002 and ΔRMSEA ≥ 0.007 [ 119 ]. The other approach is a specific one based on effect size using the comparison of specific parameters (factor loadings, thresholds, and residuals) associated with the invariance between the groups [ 120 , 121 ], taking Cohen’s indices as the reference [ 122 ]. Then, to compare the factor loadings of men and women, the coefficient q was used; magnitudes below |0.10| were expected to determine that these were equivalent between the groups.…”

Adaptation and Validation of the Perception of Anomie Scale in Chilean University Students

“…In that sense, implementing ESEM on theoretically complex constructs could provide a more realistic point of view, and thus, an understanding of the dimensional nature of the construct [ 55 ]. Figure 1 shows the fitted models.…”

Revisión psicométrica de la Escala de autorregulación de hábitos alimentarios en adultos mexicanos: ¿Es factible una versión breve?

“…Adicionalmente, se implementó una rotación target oblicua (ε = .05) (18) ; la cual estima libremente aquellas cargas factoriales que pertenecen a los ítems de cada factor teórico (cargas principales), y especifica como cercanas a cero (~0) las cargas factoriales que pertenecen a los factores secundarios. Finalmente, se usaron módulos específicos para el cálculo de los índices complementarios para el análisis bifactor y el análisis de la invarianza de medición desde la perspectiva de la magnitud del efecto (27) .…”