本文主要探討縱貫研究之計量結果若以潛在成長模式（latent growth model, LGM）做統計分析時，所需之樣本數與統計檢力（power）之間的關係。潛在成長模式是近年來廣泛被利用於縱貫研究分析的模式之一，潛在成長模式是從結構化方程模式（structural equation models, SEM）的理論架構推導而來。本文利用Satorra與Saris（1985）所提的結構化方程模式的參數檢定力研究方法來對潛在成長模式的參數檢定正確性與樣本數大小之間的關係進行探討。針對線性潛在成長模式中的平均成長率（average growth rate），本文列舉出可供實徵研究參考的檢定力與樣本數大小的檢定力表。

This paper shows practical guidelines of sample size requirements when results of longitudinal study are to be analyzed by using latent growth models (LGM). The latent growth models are much popularized in recent years and very worth investigated. Latent growth model was derived from theories of the structural equation models (SEM). We use the method proposed by Satorra and Saris (1985) who originated the method to test SEM models and investigate relationships between parameter testing accuracy and statistical power of SEM models. The parameters of latent growth model under investigation are the parameters for modeling linear latent growth rates. We summarize the guidelines in several tables.