Help using Base SAS procedures

Odds ratio and logistic regression with continuous and categorical variables

Reply
New Contributor
Posts: 3

Odds ratio and logistic regression with continuous and categorical variables

Please help!

I have several categorical and continuous variables. When I entered a select few to a predictive model, continuous variables gave me an odds ratio and CL of 1, 1, 1 respectively. What am I doing or interpreting wrong?

proc logistic data=FA2013_7 descending;

class familygrosscat/param=ref;

model R1yr= firsttermGPA Total_Funds Budget familygrossCat;

format familygrossincome familygross.;

contrast 'familygrossincome 0 vs 1' familygrossCat 1 -1 0/ estimate=exp;

contrast 'familygrossincome 1 vs 3' familygrossCat  1 0 -1/estimate=exp;

contrast 'familygrossincome 3 vs 8' familygrossCat 0 1 -1/estimate=exp;

run;

only familygrossincome is categorical in this model, all else are continuous.

Analysis of maximum likelihood estimates

         

Intercept 1-2.75520.364257.2416<.0001
firsttermGPA 10.50230.074245.8382<.0001
Total_Funds 10.0000720.00001913.82910.0002
Budget 10.0000850.00001919.6326<.0001
familygrossCat20,000 to 39,99910.22540.20221.24200.2651
familygrossCat40,000 to 59,99910.62990.25815.95690.0147
familygrossCat60,000 to 79,99910.14330.34690.17080.6794
familygrossCatLess than 20,0001-0.22130.21961.01580.3135

Odds ratio estimates

firsttermGPA1.6521.4291.911
Total_Funds1.0001.0001.000
Budget1.0001.0001.000
familygrossCat 20,000 to 39,999 vs Unknown or not applicable1.2530.8431.862
familygrossCat 40,000 to 59,999 vs Unknown or not applicable1.8771.1323.113
familygrossCat 60,000 to 79,999 vs Unknown or not applicable1.1540.5852.278
familygrossCat Less than 20,000 vs Unknown or not applicable0.8010.5211.233

Thank you

Ask a Question
Discussion stats
  • 0 replies
  • 143 views
  • 0 likes
  • 1 in conversation