DATA Step, Macro, Functions and more

Proc ttest

Accepted Solution Solved
Reply
Occasional Contributor
Posts: 10
Accepted Solution

Proc ttest

I don't use proc ttest often nor do I do basic ttests often so I just wanted to check if this code looks right. It runs the way I would expect. I am just comparing the ratio of means between 2 populations. Thanks!


Accepted Solutions
Solution
‎04-16-2018 11:47 AM
Super User
Posts: 24,010

Re: Proc ttest

Posted in reply to waxenbaum

You can't do that without the N's at minimum. A mean of 200 with an N=2 is very different from a mean of 200 with a N=2000. 


@waxenbaum wrote:

I don't use proc ttest often nor do I do basic ttests often so I just wanted to check if this code looks right. It runs the way I would expect. I am just comparing the ratio of means between 2 populations. Thanks!


 

View solution in original post


All Replies
Highlighted
Respected Advisor
Posts: 3,277

Re: Proc ttest

[ Edited ]
Posted in reply to waxenbaum

Next time, include the code directly in your message (not as an attachment) by clicking on the running man icon and pasting your code into the box that opens.

 

TTEST does not compare ratios of means. It tests to see if the two means are equal (actually performing a subtraction of the two means, among other things)

 

I don't think your code is correct, it needs to have a CLASS statement to tell the PROC what data is in group 1 and what data is in group 2.

 

If I am understanding your code and data properly, you cannot perform t-tests on this data at all, as you have only one data point in each gropu for Nuchal and one data point in each group for Mastoid and so on. This is not a limitation of SAS; there is no way to perform any t-test (pencil and paper, or otherwise) when you have one data point in each group.

--
Paige Miller
Solution
‎04-16-2018 11:47 AM
Super User
Posts: 24,010

Re: Proc ttest

Posted in reply to waxenbaum

You can't do that without the N's at minimum. A mean of 200 with an N=2 is very different from a mean of 200 with a N=2000. 


@waxenbaum wrote:

I don't use proc ttest often nor do I do basic ttests often so I just wanted to check if this code looks right. It runs the way I would expect. I am just comparing the ratio of means between 2 populations. Thanks!


 

☑ This topic is solved.

Need further help from the community? Please ask a new question.

Discussion stats
  • 2 replies
  • 163 views
  • 4 likes
  • 3 in conversation