2

I am using the polr fucntion in R for ordered logistic regression with one covariate/predictor variable. Here is my formula below:

library(MASS)
telangiec_ordinal <- polr(telangiectasia_tumour_24_adj ~ prs, data = PRS_covar_ordinal_wo_na, method = "logistic")
summary(telangiec_ordinal)

Output:

Coefficients:
        Value Std. Error t value
prs -0.007486  0.0003646  -20.53

Intercepts:
    Value       Std. Error  t value    
0|1      0.6261      0.0000 124726.8769
1|2      3.3920      0.4336      7.8227

Residual Deviance: 651.8019 
AIC: 657.8019 

When I run confint function, I don't get an output or I get NAs. E.g. confint(telangiec_ordinal)

2.5 % 97.5 % 
NA     NA 

I can calculate confidence intervals manually, which is no problem using estimate +/- (1.96 x standard error) but I am not sure whether this way is applicable for ordered logistic regression?

HKJ3
  • 387
  • 1
  • 10

0 Answers0