Garry Anderson <g.anderson@unimelb.edu.au> asks about the binomial confidence
interval reported by -roctab- when the estimated area under the receiver
operating characteristics curve is one:
>
> ROC -- Binomial Exact --
> Obs Area Std. Err. [95% Conf. Interval]
> --------------------------------------------------------
> 109 1.0000 0.0000 0.00023 0.05006
>
>
> How does one interpret the 95% CI of 0.00023 to 0.05006 when the ROC area is
> 1.00? I have seen a dataset (n=47, 15 +ve) where the ROC area was 1.00 and
> I wish to determine the lower 95%CI.
Garry discovered a bug in -roctab- with the -binomial- option when the ROC
area estimate is one. We will fix this in the next update. Meanwhile, Garry
can obtain the lower bound for the ROC area estimate by using -cii-. For the
above example the lower bound is 0.9667233:
. cii 109 109
Variable | Obs Mean Std. Err. [95% Conf. Interval]
-------------+---------------------------------------------------------------
| 109 1 0 .9667233 1*
(*) one-sided, 97.5% confidence interval
For a dataset with 47 observations the lower bound is 0.9245153:
. cii 47 47
-- Binomial Exact --
Variable | Obs Mean Std. Err. [95% Conf. Interval]
-------------+---------------------------------------------------------------
| 47 1 0 .9245143 1*
(*) one-sided, 97.5% confidence interval
-- Yulia
ymarchenko@stata.com
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/