- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I'm trying to output population standard deviation , not just standard deviation. On SQL, it is simple as using the std_pop. Is there anyway on SAS that I can turn the keys on proc univariate and proc means?
Accepted Solutions
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hi @lydiawawa,
You can also use the VARDEF=N option of PROC MEANS (same in PROC UNIVARIATE) to switch from sample to population standard deviation.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
I think you you have to run PROC MEANS and create an output data set, then in a DATA step you take the outputs from PROC MEANS and then multiply by the proper factor to convert it to the population standard deviation.
Paige Miller
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
@lydiawawa wrote:
How do you convert the sum function with multiple layers of groups?
Is this related to your first question? Can you provide more explanation of what you are asking?
Paige Miller
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
If I want to compute pop standard deviation of supposedly college admission rate among counties by state. How should I convert mean to population standard deviation ?
State County College Admission Rate
MA 01 0.7
MA 03 0.5
VA 02 0.3
VA 05 0.4
VA 07 0.6
Mean:
State College Admission Rate
MA 0.6
VA 0.43
How do I convert the sum function and derive the pop std manually?
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
I'm afraid I don't understand. You keep jumping from standard deviation to mean and back. Please explain in more detail.
Paige Miller
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
When I spoke about using PROC MEANS, it was to get a sample standard deviation. Then you can multiply the sample standard deviation by the proper factor to get the population standard deviation. You don't need to do anything manually here (other than write the proper code) and you don't specifically need to compute the mean and then use it, because PROC MEANS does that internally.
Paige Miller
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hi @lydiawawa,
You can also use the VARDEF=N option of PROC MEANS (same in PROC UNIVARIATE) to switch from sample to population standard deviation.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
@FreelanceReinh wrote:
You can also use the VARDEF=N option of PROC MEANS (same in PROC UNIVARIATE) to switch from sample to population standard deviation.
I forgot about that. Definitely a better solution than what I was suggesting.
Paige Miller