Zara E. Sadler, PhD, MS¹, Barbara Tilley, PhD¹, Philip F. Rust, PhD¹, Peng Huang, PhD¹, and Linda M. Kaste, DDS, PhD². (1) Department of Biometry and Epidemiology, Medical University of South Carolina, 135 Cannon Street, Suite 303, P.O. Box 250835, Charleston, SC 29425, 843-876-1100, sadlerze@musc.edu, (2) Department of Pediatric Dentistry, University of Illinois at Chicago, College of Dentistry, 801 S. Paulina Street, Chicago, IL 60612

Structural equation modeling often is evaluated using a related class of goodness-of-fit (GOF) measures and with the chi-square distribution typically used for acceptance or rejection of the null hypothesis. These GOF measures have limitations, including little a priori knowledge of empirical and theoretical distributions in applied settings with categorical data. As a result, hypothesis testing, sampling distribution, confidence intervals construction, or significance comparisons in these GOF indices are eased by employment of re-sampling techniques, such as bootstraping. Via four widely used fit statistics, this research conducted simulation studies to estimate means, standard deviations, skewness, and coefficients of variation, as well as estimated potential bias for the estimation procedures. Using re-sampling techniques, the estimates of mean values for these four indices appear consistent and precise, with the bias associated with estimating techniques small. Hence, by evaluating the distribution of these fit statistics in these special settings, bootstrap confidence intervals can be derived for making multiple groups comparisons.

Learning Objectives: At the conclusion of the session, the participants in this session will be able to

• develop an approach for comparing independent fit statistics when data are binomially distributed;
• articulate the procedure for assessing the empirical distribution of four widely used fit statistics obtained from confirmatory factor analysis.

Presenting author's disclosure statement:
I do not have any significant financial interest/arrangement or affiliation with any organization/institution whose products or services are being discussed in this session.