Evaluación del Análisis Factorial Confirmatorio con escalas de tipo Likert

• Luis González Fernández ^[1] ; Luis Arciniega Ruiz De Esparza ^[2]
  1. [1] Universidad de Salamanca

     Universidad de Salamanca

     Salamanca, España

     
  2. [2] Instituto Tecnológico Autónomo de México

     Instituto Tecnológico Autónomo de México

     México

     
• Localización: Actas del XI Congreso Nacional de Psicología Social: Tarragona, 1, 2 y 3 de octubre de 2009 / Jordi Tous Pallarès (ed. lit.), Joan Manel Fabra Sopeña (ed. lit.), 2009, ISBN 9788469256848
• Idioma: español
• Títulos paralelos:
  • Assessing Confirmatory Factor Analysis with Likert scales data
• Texto completo no disponible (Saber más ...)
• Resumen

  • Most of the instruments used in social psychology to assess latent constructs such as personality traits orattitudes, the Likert scales used to collect responses from the subjects. These scales have a number ofcategories to be arranged between the choice of subject. The data obtained from these scales are ordinal innature, but are often treated as continuous. In this regard, it is common practice in the confirmatory factoranalysis (AFC) the wider use of the procedures used to analyze continuous variables and ordinal data withmethods for estimating the maximum likelihood (ML) which represent a normal distribution multivalent.However, there are estimation procedures, such as the use of weighted least squares (WLS), less sensitiveto violation of assumptions of normality of the data. This paper aims to compare the confirmatory factoranalysis with the methods of ML and WLS estimation in the analysis of data from Likert scales. Morespecifically assessed the factorial structure of the resistance to change (Oreg, 2003; Arciniega andGonzalez, 2009) from data from a sample of 546 graduates in management and business administration.