Exploring the Sensitivity of Horn's Parallel Analysis to the Distributional Form of Random Data

Abstract

Horn’s parallel analysis (PA) is the method of consensus in the literature on empirical methods for deciding how many components/factors to retain. Different authors have proposed various implementations of PA. Horn’s seminal 1965 article, a 1996 article by Thompson and Daniel, and a 2004 article by Hayton et al., all make assertions about the requisite distributional forms of the random data generated for use in PA. Readily available software is used to test whether the results of PA are sensitive to several distributional prescriptions in the literature regarding the rank, normality, mean, variance, and range of simulated data on a portion of the National Comorbidity Survey Replication (Pennell et al., 2004) by varying the distributions in each PA. The results of PA were found not to vary by distributional assumption. The conclusion is that PA may be reliably performed with the computationally simplest distributional assumptions about the simulated data.

Propiedades psicométricas del FACESIII - Réplica

“…El comentario en cuestión se resume en el cuestionamiento del uso del Scree Test como técnica para la elección del número de factores en el análisis factorial exploratorio (AFE) y, en su lugar, se recomienda el uso de un análisis paralelo (AP). Al respecto, debemos precisar que conocemos la técnica del AP, sabemos de sus bondades (1) , de sus limitaciones cuando se trabaja con matrices policóricas y el modelo de factor común (2) , y que en la actualidad aún se están buscando mejores opciones (3) . No obstante estas limitaciones -y porque en general es un estándar-, también aplicamos la técnica del AP antes de elaborar el manuscrito del presente estudio, utilizando el programa Stata (4) ; y sus resultados fueron similares a los del Scree Test (Tabla 1).…”

“…Estudiar los aspectos asociados al sistema familiar está tomando mayor relevancia en las últimas décadas, debido a su influencia en diversas facetas de la vida del individuo. Por ello, el trabajo de Bazo-Álvarez et al (1) representa un aporte en cuanto a la evaluación de la funcionalidad familiar en el contexto peruano.…”

Constructs of social and emotional effectiveness: Different labels, same content?

“…It was beyond the scope of this analysis to propose theoretically motivated latent variables that influenced participants' responses on the observed variables, for which exploratory factor analysis would have been the more appropriate method. The number of components to extract was determined with parallel analysis implemented in the ''paran'' package in R (Dinno, 2009). Parallel analysis adjusts the number of components to extract by the number of components that would be derived from random data.…”

Using Horn’s Parallel Analysis Method in Exploratory Factor Analysis for Determining the Number of Factors

“…Humphreys and Montanelli (1975) compared the parallel analysis method to the maximum likelihood method, and parallel analysis was found to give results that were consistent (almost 100%) with the number of factors obtained from the actual data set. Dinno (2009;2010) examined the consistency of the parallel analysis method with the number of factors obtained from the actual data set for both factor analysis and principal components analysis, by changing distribution properties of the simulative data in the parallel analysis method and concluded that the methods were independent of the distribution (distribution-free) properties of data, and that for determining number of factors, parallel analysis was found to give results that were consistent with the number of factors obtained from the actual data set. Crawford et al (2010) compared the parallel analysis method, the principal components and principal factor methods and the criteria for the mean Eigen value to those of an eigenvalue of 95%.…”