bootstrap confidence intervals for ordinary least squares…：普通最小二乘法的自举置信区间….pdf

Multivariate Behavioral Research, 45:104–134, 2010 Copyright ? Taylor & Francis Group, LLC ISSN: 0027-3171 print/1532-7906 online DOI: Bootstrap Con?dence Intervals for Ordinary Least Squares Factor Loadings and Correlations in Exploratory Factor Analysis Guangjian Zhang University of Notre Dame Kristopher J. Preacher University of Kansas Shanhong Luo University of North Carolina at Wilmington This article is concerned with using the bootstrap to assigncon?dence intervals for rotated factor loadings and factor correlations in ordinary least squares exploratory factor analysis. Coverage performances ofSE-based intervals, percentile intervals, bias-corrected percentile intervals, bias-corrected accelerated percentile intervals, and hybrid intervals are explored using simulation studiesinvolving different sample sizes, perfect and imperfect models, and normal and elliptical data. The bootstrap con?dence intervals are also illustrated using apersonality data set of 537 Chinese men. The results suggest that the bootstrap is aneffective method for assigning con?dence intervals at moderately large sample sizes. Exploratory factor analysis (EFA) is widely used in the social sciences (Cud- eck & allum, 2007; Fabrigar, Wegener, allum, & Strahan, 1999; Gorsuch, 1983; Woods & Edwards, 2008; Yanai & Ichikawa, 2007). It allows the researcher to study unobservable constructs (., personality traits) through observable measures (., questionnaire items). The unobservable constructs are Correspondence concerning this article should be addressed to Guangjian Zhang, Psychology De- partment, Haggar Hall, University of Notre Dame, Notre Dame, IN 46556. E-mail: ******@ 104 Downloaded By: [University of Notre Dame] At: 16:49 12 March 2010 THE BOOTSTRAP OF EFA 105 referred to as latent factors; the directly measured variables are referred to as manifest variables. Factor loadings of manifest variables on latent factors reveal the nature of latent facto