Lipfert F W
J Environ Econ Manage. 1984 Sep;11(3):208-43. doi: 10.1016/0095-0696(84)90019-6.
"A series of cross-sectional multiple regressions of 1969 and 1970 U.S. mortality rates is presented for up to 112 SMSAs, against various demographic, environmental, and lifestyle variables. The basic data set is the same as that used by L. B. Lave and E. P. Seskin...for 1969, except that many more independent explanatory variables have been added. Since not all of these variables were available for all of the SMSAs, there were several data sets analyzed, depending on the selection of independent variables. The regression coefficients for air pollution tended to be quite sensitive to both the inclusion of the new independent variables and the selection of data sets."
针对多达112个标准都市统计区,给出了1969年和1970年美国死亡率的一系列横断面多元回归分析结果,这些回归分析涉及各种人口统计学、环境和生活方式变量。基本数据集与L.B.拉夫和E.P.塞斯金用于1969年的数据集相同,只是增加了更多的独立解释变量。由于并非所有这些变量在所有标准都市统计区都可获取,所以根据自变量的选择分析了几个数据集。空气污染的回归系数往往对新独立变量的纳入和数据集的选择都相当敏感。