A Bayesian Meta-Analysis of NCLEX-RN Success Predictors

Park, Kyuwon

Introduction:

Bayesian statistical analysis is seldom used in nursing research despite its widespread use in other fields. Meta-analysis is particularly suitable for the Bayesian inference because the data collected from the previous studies may contribute well to establish a prior distribution and it often deals with missing effect sizes that can be estimated well from Bayesian inference methods. This study utilizes Bayesian meta-analysis methods on the previous NCLEX-RN predictor studies.

Method(s):

Any Bayesian statistical analysis needs a prior distribution of the particular parameter of investigation. It enables a representation of the past knowledge regarding the parameter. Since this study is the first attempt of Bayesian meta-analysis on NCLEX-RN predictor studies, non-informative prior is used so that a posterior distribution of the interested parameter, mean of the effect sizes, relies more heavily on the previous studies themselves. The effect sizes are considered as a part of a random variable and the heterogeneity of the mean of the effect sizes are incorporated into the meta-analysis model. The parameter and the credible intervals are estimated by MCMC (Markov Chain Monte Carlo) simulation.

Results:

Various predictors tested in the previous studies are investigated inclduing SAT, ACT, GPA, and statndardized test scores. Mean and credible intervals of logged odds ratio of NCLEX-RN pass vs fail for each predictor are estimated.

Discussion & Conclusions:

Despite its complex solution and a consequential computation requirement, Bayesian meta-analysis has been utilized a lot more lately in physical and social science studies. Technological advances in the personal computers and increased availability of Bayesian analysis tools enable the widespread use of Bayesian statistical methods and challenge the dominant frequentist paradigm in statistics. Bayesian way of thinking may also contribute to the advancement of nursing science as it has been to the other fields.