topic Re: Multinomial regression and likelihood ratios in New SAS User

Multinomial regression and likelihood ratios

Sakshi13 — Sun, 07 Apr 2019 02:54:54 GMT

Hello, I need to recreate the following table, could anyone please guide me for proper code. It says likelihood ratios calculated using multinomial regression.

My outcome variable is binomial : 1 indicating return and 0 means no return

Age and height are continuous variable

while others are categorical with 0 and 1 values.

When I searched , I get rather results of likelihood ratio tests, I will really appreciate all the help here.

Thank you

Table 5. Likelihood ratios for all cause return

Parameter

Likelihood ratio

Age

Gender

Height

Renal disease

Liver disease

0.06

0.03

1.09

0.70

0.07

* Via multinominal regression analysis, significance p<0.05.

Re: Multinomial regression and likelihood ratios

Reeza — Sun, 07 Apr 2019 03:06:38 GMT

Likelihood ratio or odds ratio?

Re: Multinomial regression and likelihood ratios

Sakshi13 — Sun, 07 Apr 2019 03:17:26 GMT

Hi Reeza, likelihood ratio as mentioned in Table. I got this as an assignment.

Re: Multinomial regression and likelihood ratios

Reeza — Sun, 07 Apr 2019 04:19:30 GMT

Do you have a reference for that then? Is it for when you're comparing models against each other? So fit base model and then fit a model with each covariate and test against base? Or something else? Sorry, my stats are rusty but I can definitely help if I understand the formula.

Re: Multinomial regression and likelihood ratios

Sakshi13 — Sun, 07 Apr 2019 22:11:30 GMT

Hello, I did email to my advisor what he actually meant. I am afraid I have no idea, but just have this Table. Would you be able to guide me towards the code for Multinomial regression?

Re: Multinomial regression and likelihood ratios

Reeza — Mon, 08 Apr 2019 04:02:05 GMT

PROC LOGISTIC and see the example section, nominal regression, there's a fully worked example in the docs.