turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- SAS Programming
- /
- SAS Procedures
- /
- Proc Means confidence limit gives "."

Topic Options

- RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

03-23-2016 03:57 PM - edited 03-23-2016 06:52 PM

Hi Everyone,

I am trying to get the 95% confidence interval for a disease severity progression data set in SAS 9.4. The data have been log transformed and are not normal (certain weeks have a lot of 0 values). Sometimes the confidence limit output contains "." rather than numbers. In the year 2 output it seems that this only happens when the disease severy mean in 0 (see year 2 below), but in the year 1 data it happens for weeks when the disease severity mean is 0.3010 or 0 (see year 1 below). My question is: what does the "." mean? Why is a "." returned even when the mean is not 0?

Thank you

Sherri

This is the coding for year 2:

Title2 "Confidence Interval";

proc means data= first maxdec=4 alpha=0.05 clm mean;

class trt;

var jul23 jul31 aug6 aug13 aug20 aug27 sep3 sep10 sep17;

run;

and this is part of the output:

Coding for year 1

Title2 "Confidence Interval";

proc means data= first maxdec=4 alpha=0.05 clm mean;

class trt;

var aug21 aug30 sep14 sep24 oct11;

run;

Output for year 1:

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

03-23-2016 04:12 PM

Is it possible you have missing values in your data? If only one observation has a nonmissing value, you can get a mean. But you can't get confidence limits. Perhaps you could check your data for a few of these cases.

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to Astounding

03-23-2016 04:24 PM - edited 03-23-2016 04:26 PM

There are 3 missing values in the year 1 data set (Oct 11 trt 2, and Sept 24 and Oct 11 trt 3), but they don't seem to correspond to the missing values in the confidence limit output. If it helps provide context, both data sets have 7 trt and four reps

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

03-23-2016 04:42 PM

Perhaps you could share some of the data. There are other conditions where there would be a mean but no confidence limit, such as all values being identical. Are you able to share all the data points for one set of values that has no confidence limits?

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to Astounding

03-23-2016 06:51 PM

Here is data for trt 7 of year 1

block trt aug21 aug30 sep14 sep24 oct11; 1 7 0 0 0.2 0.5 1 2 7 0 0 0 0.6 1 3 7 0 0.1 0 0.9 1 4 7 0 0 0 0.5 1 ;

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

03-23-2016 05:45 PM

Add Standard Deviation to the chart please and repost.

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

Posted in reply to Reeza

03-23-2016 06:53 PM

Please see orginal message for update chart.

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

03-23-2016 08:25 PM

Do you see the issue?

All results with missing CI have a standard deviation of 0 - implying that the values are all that single value.

- Mark as New
- Bookmark
- Subscribe
- RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content

03-24-2016 01:50 AM

As Reeza pointed out . You can not calculated CI when it is at the highest level (nway), there is only one obs for that.
Specify WAYS statement.

proc means data= first maxdec=4 alpha=0.05 clm mean; class trt; ways 1 ; var jul23 jul31 aug6 aug13 aug20 aug27 sep3 sep10 sep17; run;