当我通过多个变量collapse
数据集时,默认情况下,Stata仅保留有观察值的组合,因此可以计算统计数据。我可以做些什么来保留所有组合,并且无法计算无法计算的统计数据?在一个例子中:
sysuse nlsw88, clear
collapse (mean) grade, by(race occupation)
table occupation race
--------------------------------------------
| race
occupation | white black other
-----------------------+--------------------
Professional/technical | 1 1 1
Managers/admin | 1 1 1
Sales | 1 1 1
Clerical/unskilled | 1 1
Craftsmen | 1 1
Operatives | 1 1 1
Transport | 1 1
Laborers | 1 1 1
Farmers | 1
Farm laborers | 1 1
Service | 1 1
Household workers | 1
Other | 1 1
--------------------------------------------
. list if race == 2 & occupation == 9
你可以看到没有对黑人农民的观察。我想拥有的是:
. list if race == 2 & occupation == 9
+--------------------------+
| race occupa~n grade |
|--------------------------|
32. | black Farmers . |
+--------------------------+
答案 0 :(得分:2)
您可以使用fillin
:
sysuse nlsw88, clear
collapse (mean) grade, by(race occupation)
fillin occupation race
table occupation race
| race
occupation | white black other
-----------------------+--------------------
Professional/technical | 1 1 1
Managers/admin | 1 1 1
Sales | 1 1 1
Clerical/unskilled | 1 1 1
Craftsmen | 1 1 1
Operatives | 1 1 1
Transport | 1 1 1
Laborers | 1 1 1
Farmers | 1 1 1
Farm laborers | 1 1 1
Service | 1 1 1
Household workers | 1 1 1
Other | 1 1 1
--------------------------------------------
list if race == 2 & occupation == 9
+------------------------------------+
| race occupa~n grade _fillin |
|------------------------------------|
26. | black Farmers . 1 |
+------------------------------------+
根据fillin
的帮助文件:
fillin 添加缺少数据的观察结果,以便 varlist 的所有互动都存在