Closed jorpppp closed 5 months ago
In https://github.com/JMSLab/xtevent/commit/5b042a8383c9165c01d6df0044e9381465374948 I changed the sup-t code because of discrepancies with the Matlab implementation. The critical value coming out from the original Stata implementation was slightly off (by about 0.1 in my tests) so the sup-t CIs in previous versions before these bug fix were conservative.
Summary: In this issue we fixed a bug with the sup-t CIs where they were slightly wide. Thread continues in #172
There is a bug with the sup-t confidence intervals as they may be too wide.