Programming the statistical procedures from SAS

Regression Analysis

Accepted Solution Solved
Reply
Contributor
Posts: 67
Accepted Solution

Regression Analysis

Hello

 

Please help me with below questions

 

1) Difference between estimates and standardized estimes in regression analysis.

 

2) Why do we get t stats in Linear regression and z stats in Logistic regression.

 

3) Please provde some references to Time series Analysis.

 

Thanks


Accepted Solutions
Solution
‎10-11-2016 04:54 AM
SAS Employee
Posts: 187

Re: Regression Analysis

[ Edited ]

The Wald chi-square statistic presented in the LOGISTIC and GENMOD procedures for a logistic model are just the square of a Z statistic with identical p-value. 

 

...and standardized estimates are detailed in the description of the STB option in the "Syntax: MODEL statement" section of the LOGISTIC documentation.

View solution in original post


All Replies
Contributor
Posts: 62

Re: Regression Analysis

This site will have the answers to those questions:

 

http://www.ats.ucla.edu/stat/sas/

 

Norman.
SAS 9.4 (TS1M0) X64_7PRO WIN 6.1.7601 Service Pack 1 Workstation

Grand Advisor
Posts: 16,933

Re: Regression Analysis

I don't think Q2 is correct. 

Contributor
Posts: 67

Re: Regression Analysis

Let me frame the question again

 

In the linear regression analysis output table we are having a column on t value.

 

In the logistic regression anlaysis output table we are having a column on z value.

 

Is there any specific reason for it?.

 

Thanks

Contributor
Posts: 62

Re: Regression Analysis

Yes, there is a specific reason for this. Have a look here:

 

http://zencaroline.blogspot.co.uk/2011/04/difference-between-z-test-and-t-test_10.html

Norman.
SAS 9.4 (TS1M0) X64_7PRO WIN 6.1.7601 Service Pack 1 Workstation

Grand Advisor
Posts: 16,933

Re: Regression Analysis

Post a screenshot of what you're referring to please. 

 

Grand Advisor
Posts: 9,466

Re: Regression Analysis

t value  is ususally for small size of sample.

z value is for the large size of sample.

 

When the size of sample is big enough, t almost equal z.

Here REG use t value , I guess REG want get more accurate parameter estimate.

Contributor
Posts: 62

Re: Regression Analysis

With simple linear regression, by default the "Parameter Estimates" table contains the estimates of $\beta _0$ and $\beta _1$together with the t statistics and the corresponding p-values for testing whether each parameter is significantly different from zero. 

 

http://support.sas.com/documentation/cdl//en/statug/68162/HTML/default/viewer.htm#statug_reg_getting...

Norman.
SAS 9.4 (TS1M0) X64_7PRO WIN 6.1.7601 Service Pack 1 Workstation

Grand Advisor
Posts: 16,933

Re: Regression Analysis

[ Edited ]

And logistic regression using chi square AFAIK, not Z. I'm curious as to where OP is seeimg Z or under what circumstances that's generated. 

 

http://support.sas.com/documentation/cdl//en/statug/68162/HTML/default/viewer.htm#statug_logistic_ex...

Grand Advisor
Posts: 16,933

Re: Regression Analysis


@KafeelBasha wrote:

 

 

2) Why do we get t stats in Linear regression and z stats in Logistic regression.

 


I tried a variety of regressions this morning and couldn't get any z-statistics from Logistic Regression using PROC LOGISTIC. Are you using a different procedure? Please post code and data to replicate your issue. 

Contributor
Posts: 67

Re: Regression Analysis

Hello

 

I tried with R function glm().

 

Please refer the below link

 

http://www.ats.ucla.edu/stat/r/dae/logit.htm

 

Thanks

Solution
‎10-11-2016 04:54 AM
SAS Employee
Posts: 187

Re: Regression Analysis

[ Edited ]

The Wald chi-square statistic presented in the LOGISTIC and GENMOD procedures for a logistic model are just the square of a Z statistic with identical p-value. 

 

...and standardized estimates are detailed in the description of the STB option in the "Syntax: MODEL statement" section of the LOGISTIC documentation.

Grand Advisor
Posts: 16,933

Re: Regression Analysis

@StatDave_sas thanks for clarification! 

 

☑ This topic is SOLVED.

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

Discussion stats
  • 12 replies
  • 410 views
  • 4 likes
  • 5 in conversation