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
- /
- Analytics
- /
- Stat Procs
- /
- Why there is not report of Gamma value to indicate...

Topic Options

- Subscribe to 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
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

08-10-2016 07:59 AM

Dear All,

I am using PROC FREQ to get the proportion of discordant pairs and its 90% Confidence interval. However, I found one special case when the Binary response all equal to 1, then there is no report on the ODDS ratio, Gamma, etc.

Anybody can help me to figure out this? To my understanding, if all binary Response equal to 1, then the Gamma value should be 1.

Many thanks for your help to me on this.

```
proc freq data=dar_s order=data;
tables drgdsc1a* SWAEVL2C/measures CL;
run;
```

Accepted Solutions

Solution

08-11-2016
02:31 AM

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

08-10-2016 08:35 AM

These measures of association are defined for ordinal variables in two-way tables. From the doc: "A pair is concordant if the observation with the larger value of X also has the larger value of Y." (X=row variable; Y=column variable)

In your example, the Y variable has only one value. There is no "larger value of Y" to use to define pairs as condordant or discordant. Thus any formula that uses "number of condordant pairs" is undefined.

All Replies

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

08-10-2016 08:18 AM

Check the log. Is there a NOTE that says:

NOTE: No statistics are computed for drgdsla * SWAEVL2C because SWAEVL2C has less than 2 nonmissing levels.

Solution

08-11-2016
02:31 AM

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

08-10-2016 08:35 AM

These measures of association are defined for ordinal variables in two-way tables. From the doc: "A pair is concordant if the observation with the larger value of X also has the larger value of Y." (X=row variable; Y=column variable)

In your example, the Y variable has only one value. There is no "larger value of Y" to use to define pairs as condordant or discordant. Thus any formula that uses "number of condordant pairs" is undefined.

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

08-11-2016 02:21 AM - edited 08-11-2016 03:40 AM

Dear Rick,

You ansswered my question perfectly. I understand the "concordant and discordant" more clearly with you post.

Best,

Jack

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

08-11-2016 02:46 AM

As Rick pointed out, you can't get Gamma due to there is only one level(value 1) of SWAVEL2C. Gamma measure the correlation between row variable and col variable, so you can't get it if col variable only have one level.

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

08-11-2016 03:41 AM

Great thanks, got now.