论文部分内容阅读
线性回归模型被广泛应用于精神卫生和卫生服务相关研究。然而,经典线性回归分析是假设该数据为正态分布的,但是很多研究所获得的数据并不符合这种假设。解决该问题的方法之一是采用不要求数据为正态分布的半参数模型。但是,半参数模型对离散数据相当敏感,因此在处理包含离散值的数据时产生的估计值是不可靠的。在这种情况下,一些研究者在删减这些极端值后再进行分析,但是,删减数据的事先法则(ad-hoc rules)是基于主观标准的,所以不同的调整方法就会产生不同的结果。等级回归为处理包括离散值的非正态分布数据提供了更为客观的方法。本文采用虚拟和实际数据来阐述这个非常有用的处理离散值的回归方法,并与采用经典回归模型和半参数回归模型所得出的结果进行比较。
Linear regression models are widely used in mental health and health services related research. However, the classical linear regression analysis assumes that the data is normally distributed, but the data obtained in many studies do not fit this assumption. One way to solve this problem is to use a semi-parametric model that does not require the data to be normally distributed. However, the semiparametric model is quite sensitive to discrete data and so the estimates produced when processing data containing discrete values are not reliable. Under these circumstances, some researchers cut back on these extremes before analyzing them. However, the ad-hoc rules of data deletion are based on subjective criteria, so different adjustment methods will produce different result. Level regression provides a more objective approach to dealing with non-normal distributions of data that include discrete values. This article uses both virtual and real data to illustrate this very useful regression method for dealing with discrete values and compares them with results obtained using classical regression models and semi-parametric regression models.