Extended Unbiased Distribution Free Estimator With Mean Structures

Abstract

Listen

Translate article

To handle the nonnormal data issue, Browne proposed an unbiased distribution free (DF) estimator () and an asymptotically distribution free estimator () of the covariance matrix of sample variances/covariances to calculate robust test statistics and robust standard errors. However, is ignored in methodological and substantive research, and has not been extended to models with mean structures. To improve robust standard errors and the model fit statistic for nonnormal data with mean structures (e.g., growth curve models), we propose an unbiased distribution free estimator with mean structures considered. In growth curve models, we apply to four robust statistics that have relatively simple forms and denote them as and We compare their performance with 7 robust test statistics that employ We find that with the same model fit statistic, generally leads to smaller Anderson-Darling distances from the theoretical distribution than except in some skewed cases. Additionally, the p-values from are distributed closest to the theoretical distribution among the 12 examined statistics. In terms of Type I error rates, and are the most stable statistics. Additionally, provides smaller relative biases of the robust SE estimates than Hence, we suggest using in both model fit statistics and robust SE calculation. Among the model fit statistics using we suggest