Goldberg E L, Comstock G W
Am J Epidemiol. 1980 Jun;111(6):736-52. doi: 10.1093/oxfordjournals.aje.a112952.
The objective was to provide additional background information for a refinement of life events methodolgy. Data about life events for a one year period were gathered from a representative sample of the population in Kansas City, Missouri, and Washington County, Maryland, between 1971-1974. Using binary variable multiple regression, the relationships between individual events, overall scores and demographic variables were examined for 2780 subjects. Age, education, marital status, location and race were shown to be significantly related to scores of one or more on the life events scale. Individual items were also related to these and other demographic variables. Fifteen individual events were shown to be moderately related to one another, so that when one event occurred the other was likely to occur also. Because different subgroups of the population experience different frequencies of total life events and of particular individual events, life events scores can vary considerably from group to group, depending on demographic composition and the appropriateness of the life events list for each demographic subgroup. Such relationships, if not adjusted for, could lead to coincidental associations between life events and health-related outcomes.
目的是为完善生活事件方法提供更多背景信息。1971年至1974年间，从密苏里州堪萨斯城和马里兰州华盛顿县具有代表性的人群样本中收集了一年期间的生活事件数据。对2780名受试者，使用二元变量多元回归分析了个体事件、总体得分与人口统计学变量之间的关系。年龄、教育程度、婚姻状况、地点和种族与生活事件量表上一个或多个得分显著相关。各个项目也与这些以及其他人口统计学变量相关。15个个体事件显示彼此之间存在中度关联，因此当一个事件发生时，另一个事件也可能发生。由于不同人群亚组经历的总生活事件和特定个体事件的频率不同，生活事件得分可能因组而异，这取决于人口统计学构成以及生活事件列表对每个人口统计学亚组的适用性。如果不对此类关系进行调整，可能会导致生活事件与健康相关结果之间出现巧合关联。
