CODEWITHSUNDEEP
saslogistic-regressionquartilecategorical
miczan
miczan

Reputation: 1

How do I categorize a non-linear continuous variable in logistic regression

I am performing a logistic regression using a continuous IV that represents a number of days. There are 4 values within the continuous IV - 89, 90, 91, 92. This IV is not linear with the log odds of the dependent variable, which means I need to treat it as a categorical variable.

The quartiles of the continuous IV are min = 89, p25 = 89, p50 = 92, p75 = 92, max = 92 because the distribution of this IV is skewed towards the extremes. I am not sure how to recategorize this continuous IV as a categorical variable because of this. Can someone please help?

I tried just making each value a category and using "89" as the reference category, but this feels wrong...

proc logistic data = data; class independent_variable (ref = '89') / param = ref; model outcome(event='1') = independent_variable; run;

Upvotes: 0

Views: 111

Answers (0)

Related Questions