Wan T T
Health Serv Res. 1976 Winter;11(4):464-77.
Two-stage multivariate analysis was used to examine factors affecting personal perception of health status. In the first stage, sociodemographic variables were used as independent variables in Automatic Interaction Detector (AID) analysis in order to partition the study sample (11,153 civilian noninstitutionalized adults aged 58-63) into subgroups. In the second stage, binary multiple regression analysis was performed on each AID subgroup and on the total sample. Predictors used were indicators of psychological, socioeconomic, and sociomedical well-being. Finally the applicability of these indicators in classifying persons in one of the two categories of perceived health status was examined by discriminant function analysis. Sociomedical health indicators were better explanatory variables of self-assessed health status than socioeconomic or psychological indicators of well-being.
采用两阶段多变量分析来检验影响个人健康状况认知的因素。在第一阶段，社会人口统计学变量被用作自动交互检测（AID）分析中的自变量，以便将研究样本（11153名年龄在58 - 63岁的非机构化成年平民）划分为不同亚组。在第二阶段，对每个AID亚组和整个样本进行二元多元回归分析。所使用的预测变量是心理、社会经济和社会医学幸福感指标。最后，通过判别函数分析检验这些指标在将个体分类到两种感知健康状况类别之一中的适用性。社会医学健康指标比社会经济或心理幸福感指标更能解释自我评估的健康状况。
