topic Re: linearity assumption for logistic regression model in SAS Programming

linearity assumption for logistic regression model

stellapersis7 — Fri, 16 Feb 2024 04:50:36 GMT

Hi all,

I am student and doing a project, so I am confused about a linearity assumption of a continuous independent variable which is included in the model.

My aim is to check the graphs between gestage and estimated probability for genders male and female.

I need to check when the variable is included as 1) categorical and 2) continuous, if the linearity assumption is met.

My code is

proc genmod data = SAS.z_final descending;

class gender under_weight QNO;

model under_weight = gender gestage /dist=b link=logit type1 type3;

repeated subject = QNO/ corr=cs corrw;

output out = pain pred=p resraw=r;

run;

My question is how to get the linearity assumptions for the independent variable gestage as continuous and categorical variables for gender(independent categorical variable) male and female?

Re: linearity assumption for logistic regression model

Ksharp — Fri, 16 Feb 2024 05:01:31 GMT

Check @Rick_SAS blogs:
https://blogs.sas.com/content/iml/2016/06/22/sas-effectplot-statement.html
https://blogs.sas.com/content/iml/2019/05/30/visualize-interaction-effects-regression.html
https://blogs.sas.com/content/iml/2019/02/11/proc-plm-regression-models-sas.html

Re: linearity assumption for logistic regression model

PaigeMiller — Fri, 16 Feb 2024 11:08:33 GMT

For continuous X in a logistic regression, this is suggested in the SAS FASTats:

Linearity in the logit (or link), testingTo test for nonlinearity in the logit (for logistic models) and more generally for finding a suitable functional form for an independent variable in a generalized linear model (GLM), use the spline smoother on the variable in SAS/STAT PROC GAMPL or SAS Viya PROC GAMMOD. See the example titled "Nonparametric Logistic Regression" in the GAMPL documentation. Also see the ASSESS statement in SAS/STAT PROC GENMOD for assessing the covariate or link function in a GLM or Generalized Estimating Equations (GEE) model.

Also https://www.lexjansen.com/sesug/2019/SESUG2019_Paper-233_Final_PDF.pdf

https://stats.stackexchange.com/questions/169348/how-should-i-check-the-assumption-of-linearity-to-the-logit-for-the-continuous-i/211044#211044 points to using the Box-Tidwell test.

Checking linearity for categorical predictor makes no sense.

Re: linearity assumption for logistic regression model

Ksharp — Fri, 16 Feb 2024 11:19:04 GMT

Paige,
Sure. I forgot the "ASSESS statement in PROC GENMOD " .