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
- /
- What is null hypothesis for PROC GLM? what is esti...

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
- Permalink
- Email to a Friend
- Report Inappropriate Content

06-13-2017 05:43 AM - edited 06-13-2017 05:45 AM

After running proc GLM with multiple regressors.

I got the regressors and their Estimate, P, and T values.

I want to find out the best regressors for my model.

how should I use P, T, or Estimate column to determine which regressors are best for my model.

Also, what is the meaning of Estimate, P, and T values in model output.

Please help

this is how the output looks like

Dependent | Parameter | Estimate | Biased | StdErr | tValue | Probt |

ln_dep1 | ln_rw_base*region_nm PACIFIC NORTHWEST | 0.9604395696 | 0 | 0.00481642 | 199.41 | 0.0000 |

ln_dep1 | ln_rw_base*region_nm PACIFIC CENTRAL | 0.9601801559 | 0 | 0.00476460 | 201.52 | 0.0000 |

ln_dep1 | ln_rw_base*region_nm NEW ENGLAND | 0.9600740196 | 0 | 0.00479192 | 200.35 | 0.0000 |

ln_dep1 | ln_rw_base*region_nm PACIFIC NORTH | 0.9600171774 | 0 | 0.00475982 | 201.69 | 0.0000 |

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

Posted in reply to captainprice0

06-13-2017 05:47 AM

This sounds like multiple questions regarding understanding the GLM output.

I recommend that you take a look at some of the vell described PROC GLM examples from the documentation here:

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

Posted in reply to captainprice0

06-13-2017 08:30 AM - edited 06-13-2017 08:32 AM

P-Value is more small , the variable is more significant .

Also Check PROC GLMSELECT + SELECTION=LASSO

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

Posted in reply to Ksharp

06-13-2017 03:56 PM

Ksharp wrote:

P-Value is more small , the variable is more significant .

Also Check PROC GLMSELECT + SELECTION=LASSO

My $0.02 Philosophical difference on p-value interpretaton: The p-value is used to reject the Null hypothesis or not. The rejection level should be specified before the test is done.

With that caveat small values such as 0.05 or 0.025 are common values to compare and a p-value smaller than those will reject the null hypothesis.

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

Posted in reply to captainprice0

06-13-2017 08:30 AM

Agreeing with @draycut, your questions are all basic statistics, that can be found in documentation provided by SAS and in many statistics textbooks, and you need to do some reading there to learn these fundamental concepts.

--

Paige Miller

Paige Miller