还有个小问题， 由于总体数剧中有某一元素的离散度过高， 导致到解释力度的降低，或者是说REGRESSION结果不满意， 能否对数据组中的某一元素进行数据的分布， 比如只取起 2*Standard Deviation 间的值？ 比如说， 我的数据中包括了以下元素

ROE, ROA, MBV, TobinQ, DA, LnAsset, etc.

其中ROE的离散度较高， 希望能对ROE进行处理，把数据中ROE超过 MEAN ROE+/- 2*STD的数据排除，然后再进行 regression analysis, 可以这样做么？ 如果可以的话，请问如何在SPSS中实现。。。谢谢！！！

Question: 我的数据中包含了几个DUMMY VARIABLE， 不知道在SPSS中如何进行DUMMY的定义。

The following program places dummy variables in an hourly format in SPSS.

DATA LIST LIST /caseid(F8) start(F8.2) end(F8.2). BEGIN DATA. 1 0710 0730 1 0805 0900 2 0650 0759 3 0925 1010 3 0605 0805 4 0730 0735 5 0724 0800 6 0552 0910 7 0759 0859 8 0800 0859 9 0800 0900 10 0759 0859 END DATA. LIST.

COMPUTE s=TRUNC(start/100)-.1. COMPUTE e=TRUNC((end+1.1)/100)+.1. VECTOR v(24). LOOP #cnt=1 TO 24. COMPUTE v(#cnt)=RANGE(#cnt,s,e). END LOOP. EXECUTE.

Question: 如何进行分类对比， 或者是按某一数据的上下25%进行对总体 数据进行分类对比

It is often useful to examine means or averages of a variable according to categories of other variables, such as education. When we do this, we should have some statistical model or hypothesis in mind.

In this example, using the data set IMMUNR.SAV, we wish to confirm the expected result that on average older children will have older mothers than younger children. We do this by examining whether the mean age of mothers will be older for older children than for younger children. The variable for which the mean is calculated must be an interval variable (or possibly a dichotomous variable with values of 0 and 1) if the calculation is to be meaningful.

Exercise E.1

Open the data set IMMUNR.SAV. Request the mean age of mothers for each age category of youngest child:

Exercise E.2

Calculate means for age of mother according to education of the mother:

Question: 如何进行分类对比， 或者是按某一数据的上下25%进行对总体 数据进行分类对比

Answer:

http://www.bioon.com/biology/Class45/spss/200407/55159.html