DelosReyes John Mart V, Padilla Miguel A
Old Dominion University, Norfolk, VA, USA.
Educ Psychol Meas. 2025 Jan 31:00131644241311877. doi: 10.1177/00131644241311877.
A new alternative to obtain a Bayesian estimate of coefficient alpha through a posterior normal distribution is proposed and assessed through percentile, normal-theory-based, and highest probability density credible intervals in a simulation study. The results indicate that the proposed Bayesian method to estimate coefficient alpha has acceptable coverage probability performance across the majority of investigated simulation conditions.
本文提出了一种通过后验正态分布获得系数α的贝叶斯估计的新方法,并在模拟研究中通过百分位数、基于正态理论的和最高概率密度可信区间对其进行了评估。结果表明,所提出的估计系数α的贝叶斯方法在大多数调查的模拟条件下具有可接受的覆盖概率性能。