0
$\begingroup$

Suppose in a logistic regression a one unit increase in a linear dependent variable corresponds to an odds ratio of 1.61. The 95 % confidence interval ranges from 1.19 to 2.54.

JMP calculates the confidence intervals for "range odds ratios" exponentially so that a 11 unit increase corresponds to an interval from 1.19^11 to 2.54^11. See http://www.jmp.com/support/help/Logistic_Fit_Platform_Options.shtml

Is this correct?

$\endgroup$
1
  • 1
    $\begingroup$ When you ask is it correct do you mean (a) have I read the documentation correctly, or (b) is what JMP does the right way to do this? $\endgroup$ – mdewey Feb 16 '17 at 13:51
1
$\begingroup$

Assuming the range of your variable is 11, then that is in principle right.

However, I would not do it that way: the 1.19 and 2.54 are rounded numbers, so they are slightly off. Those rounding errors tend to become surprisingly big if you raise it to a large number like 11.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.