Feasible Sample Size Determination in Confirmatory Factor Analysis

• In Event: 46.079 - Appropriate Modeling of Latent Growth and Relationships

Abstract

Simple heuristics for estimating minimum sample size can be tremendously valuable in the early planning stages of SEM research to set parameters on what is realistic, long before the scope of the model and population are finalized. Existing heuristics are based on minimum sample size for estimation stability – not statistical power, usually of interest to researchers when planning a study. This study examined these heuristics for their ability to predict the sample size based on an a priori power analysis. Most of the sample sizes suggested by the power analysis were well above the suggested minimum requirements for estimation stability, and sample size was strongly tied to construct reliability, which is not usually considered in sample size heuristics.

Author

• Jennifer Koran, Southern Illinois University - Carbondale