Tagged: , ,

Viewing 4 posts - 1 through 4 (of 4 total)
• Author
Posts
• Marina Selini Katsaiti
Participant
Post count: 4

Hello,

I am new to this and my questions might seem stupid, but I was wondering whether you can any advice regarding the following

1. How to nest sum output in a by command, suppose I want the mean expenditures (edu food housing etc.) by level of disability. How can I nest the output in a table similar to nesting the regression results?

2. Is there a way I can tell stata to put the categories of the by command in a different order?

Thank you
Selini

Dr. Attaullah Shah
Keymaster
Post count: 106

Welcome to the Forum Selini

Yes you create summary statistics using a grouping variable. See the following example. I use `industry` as a grouping variable and find descriptive statistics for the variable `wage`.

```sysuse nlsw88, clear
bys industry : asdocx sum wage, replace```

I think you should post the second question as a new topic so that it is answered there.

Marina Selini Katsaiti
Participant
Post count: 4

Yes, exactly this is how I do it. However, I was wondering in the case you want to summarize many variables (like wage) by industry (and suppose you only need their means) whether its possible to nest the results?

Dr. Attaullah Shah
Keymaster
Post count: 106

If you just need the mean for each industry, then you use the option `stat()`. See for example:

```sysuse nlsw88, clear
bys industry : asdocx sum wage, stat(mean) replace

----+---------------------------------
1 |                            wage
----+---------------------------------
2 |Ag/Forestry/Fisheries      5.621
3 |Business/Repair Svc        7.516
4 |Construction               7.565
5 |Entertainment/Rec ~c       6.724
6 |Finance/Ins/Real E~e       9.843
7 |Manufacturing              7.502
8 |Mining                     15.35
9 |Personal Services          4.401
10 |Professional Servi~s       7.871
11 |Public Administrat~n       9.148
12 |Transport/Comm/Uti~y      11.443
13 |Wholesale/Retail T~e       6.126
--------------------------------------
```
Viewing 4 posts - 1 through 4 (of 4 total)
• You must be logged in to reply to this topic.