0
$\begingroup$

When I performed logistic regression and generated a confidence interval of the odds ratio with R using confint, it gave a different confidence interval than the one generated in SPSS. However, when I use confint.default in R instead, the confidence intervals were the same as those generated in SPSS. Can someone explain why I am getting different confidence intervals from SPSS vs R with confint and confint.default?

$\endgroup$
2
  • $\begingroup$ confint.default uses "asymptotic normality" while confint is a generic method which uses more appropriate calculation for your particular model, if it exists. Can you show a small toy example with inputs, code and outputs, which we can reproduce? $\endgroup$
    – Henry
    Jul 27, 2022 at 14:25
  • $\begingroup$ You might want to look at stats.stackexchange.com/questions/5304/… and the questions linked to it $\endgroup$
    – Henry
    Jul 27, 2022 at 14:27

0