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
- /
- which statistical tool is appropriate for my regre...

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

01-21-2013 03:24 PM

Hello,

I have a question and I will be grateful if you can help me,

I am running a regression which I have one dependent variable (varies from 1 to 6) and several independent variable, which some of them are dummy variables(like gender) and some of them are metric variables like age, and I have some other variables like education which I ranked them .I am using SAS for my research.would you please help me which statistical tool is appropriate for my regression?

Thank you in advance.

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

01-21-2013 03:31 PM

There are probably several techniques that might work here, depending on what assumptions you want to make about errors and objective function, etc.

You might want to try ordinal logistic regression, since your dependent variables are really 6 different categories. See PROC LOGISTIC

I might also go with PROC PLS, as your independent variables are most likely correlated with one another. There's also a Logistic version of PLS in the literature, although I think you'd have to code that yourself, as far as I know it doesn't exist in SAS.

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

01-21-2013 03:40 PM

Thank you so much for your answer,

my dependent ranked from 1 to 6 ( I mean it is not 6 different categories). my dependent variable is actually the likelihood of one action, it means that when is 6 is more likely to happen than when is 1.do you think still I can use PROC LOGISTIC?

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

01-21-2013 03:49 PM

Based upon my understanding of your description, and my understanding of ordinal logistic regression, I would the answer is YES! PROC LOGISTIC still applies here.

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

01-21-2013 04:06 PM

Thank you so much,

I will check the link and will come back again if I have question.

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

01-24-2013 05:02 AM

Hello,

I used proc logistic for my regression, but it is just consider 3 levels of my dependent variable,(as I mentioned before it has 6 level).

would you please help me?

Thanks a lot.

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

01-21-2013 03:57 PM

Look over the example here, but yours is more complex. This is assuming you've done the appropriate univariate modeling and summary already.

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

01-21-2013 04:04 PM

Yes, that is a very good introductory example, for a simple ordinal logistic regression.

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

01-24-2013 06:31 PM

Hello,

I used proc logistic for my regression, but it is just consider 3 levels of my dependent variable,(as I mentioned before it has 6 level).

would you please help me?

Thanks a lot.

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

01-25-2013 07:58 AM

Show us your SASLOG for what you tried. You might also want to show us a portion of your data.