Turn on suggestions

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

Showing results for

- Home
- /
- Analytics
- /
- Stat Procs
- /
- How to interpret the p-value of multi-level categorical variables in p...

Options

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

🔒 This topic is **solved** and **locked**.
Need further help from the community? Please
sign in and ask a **new** question.

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Posted 05-06-2019 03:26 PM
(3891 views)

Hi, I am a new user of SAS and I have some trouble with understanding of result of logistic regression.

For example, from type 3 analysis I know that race should have association with response variables as its p-value is small, but in the following estimation section, the different dummy variables of race have very large p-value. I don't know how to understand this part.

Can someone help me ?

Here is my code:

proc logistic data=exdata descending ;

class race

/ param=ref;

model Non_Statutory_Case_Filed= race;

run;

Thanks

1 ACCEPTED SOLUTION

Accepted Solutions

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

The Chi-squared test for RACE has p-value of 0.0050, indicating that the differences between the levels of your variable RACE are significantly different.

The p-values for the individual races (Asian, Black, *etc*.) indicate if that effect of that RACE is different from zero, and none of them are small, so the conclusion would be that the effects are not different from zero ... however, this is a different test than the one mentioned above which tests to see if the races are different from one another.

So which question do you really want the answers to?

I have never felt comfortable looking at the Type 3 effects. They can be confusing. They test things that maybe you aren't interested in. I feel that a better solution, and a much more interpretable solution, is to use the LSMEANS statement in PROC LOGISTIC to compare the different races.

--

Paige Miller

Paige Miller

5 REPLIES 5

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

You need to show the code or the design matrix so we can understand how you specified the categorical variables. For example, which parameterization method did you pick? If REF, which is the reference level? None is shown - did you filter the output somehow?

This example is a pretty detailed walk through of Logistic Regression:

I'll move your question to the Stats forum so someone there can answer the rest once you've responded.

@Ritter13 wrote:

Hi, I am a new user of SAS and I have some trouble with understanding of result of logistic regression.

For example, from type 3 analysis I know that race should have association with response variables as its p-value is small, but in the following estimation section, the different dummy variables of race have very large p-value. I don't know how to understand this part.

Can someone help me ?

Thanks

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Thanks for your response, I have updated my question

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

The Chi-squared test for RACE has p-value of 0.0050, indicating that the differences between the levels of your variable RACE are significantly different.

The p-values for the individual races (Asian, Black, *etc*.) indicate if that effect of that RACE is different from zero, and none of them are small, so the conclusion would be that the effects are not different from zero ... however, this is a different test than the one mentioned above which tests to see if the races are different from one another.

So which question do you really want the answers to?

I have never felt comfortable looking at the Type 3 effects. They can be confusing. They test things that maybe you aren't interested in. I feel that a better solution, and a much more interpretable solution, is to use the LSMEANS statement in PROC LOGISTIC to compare the different races.

--

Paige Miller

Paige Miller

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Thanks for your response!

Actually, I want to know that if the variable "race" have significant effect on the response variable, if the answer is yes, then what is the different effect of different races? For example, Asian have largest positive effect on response variable etc...

So I'm confusing about the result I got, Type 3 effect showed that race have some kind of relationship with response variable but p-value for individual races told me NO.

- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content

Adding: here's an example of using LSMEANS

but I like the LINES option or the LINESTABLE option (which are not shown in that example) and you can get some nice plots of this information as well.

--

Paige Miller

Paige Miller

Secure your spot at the must-attend AI and analytics event of 2024: SAS Innovate 2024! Get ready for a jam-packed agenda featuring workshops, super demos, breakout sessions, roundtables, inspiring keynotes and incredible networking events.

Register by March 1 to snag the Early Bird rate of just $695! Don't miss out on this exclusive offer.

** **

What is ANOVA?

ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.

Find more tutorials on the SAS Users YouTube channel.