Where and are two probability % and , are their effective sample sizes and and are their standard errors, the test statistic is:

,

where: