# Ordinal categorical predictor in multiple regression

I have an ordinal predictor (family income), which has 4 levels:

1- Below $$2000 2-$$2000 - $$3999 3-$$4000 - $$5999 4-$$6000 - $7999  So each level is an increase of$2000.

How do I include this variable in a multiple regression so that the output for family income only contains 1 estimate, which is the amount of change in the DV per \$2000 increase in family income. (In r)

I have ordered family income in r such that:

Levels: Below $$2,000 per month <$$2,000 - $$3,999 <$$4,000 - $$5,999 <$$6,000 - $$7,999 <$$8,000 - $$9,999 <$$10,000 & above


But if I put this family income variable into lm(), it produces an output with estimates for each level.

Another way to do that is using dummy or effect coding, which also produces an estimate for each level, and this is not what I am looking for.

Any help is appreciated!

• Recode the variable as numeric and you'll get what you're looking for. Commented May 17, 2019 at 4:58
• So do I just put them as 1,2,3,4? Or 2000, 4000,6000,8000?
– TYL
Commented May 17, 2019 at 6:17
• 1-2-3-4 is simpler. Try both, see the "difference". Commented May 17, 2019 at 6:24