Test statistic
The test statistic relies on dividing the observation window into \(m\) subregions.
For each subregion \(i\) we compare:
- the observed number of events \(n_{i,obs}\)
- the – under CSR – expected number of events \(n_{i,exp}\).
For subregions of equal size, this is: \(n_{i,exp} = n / m\)