自己顶

你可以先定义部门、性别、工龄的变量（如男女为1 2），然后再点数据-拆分文件-比较分组-按。。分组，之后再进行分析就可以了

谢谢，那如何比较各组的回归系数是否显著呢？

SORT CASES BY SEX RACE.
SPLIT FILE BY SEX.
FREQUENCIES VARS=INCOME /STATISTICS=MEDIAN.
SPLIT FILE BY SEX RACE.
FREQUENCIES VARS=INCOME /STATISTICS=MEDIAN.
• The first SPLIT FILE command applies to the first FREQUENCIES procedure. The second
SPLIT FILE command overrides the first and splits the file by sex and race. This split is in
effect for the second FREQUENCIES procedure.