A Problem of Testing Homogeneity Against Order Constraints on Risk Differences

Consider I pairs of independent binomial variates x0i and x1i with corresponding parameters P0i and p1i and sample sizes n0i and n1i for i=1, …,I. Let △i = P1i-P0i be the difference of the two binomial parameters, where △i's are to be of interest and P0i's are nuisance parameters. The null hypothesis of homogeneity on the risk difference can be written as     

一个关于风险差齐一性对序约束的检验问题

Consider I pairs of independent binomial variates X0i and X1i, with corresponding parameters P0i and P1i and sample sizes n0i and n1i for i = 1,…… , I. Let △i = P1i -P0i be the difference of the two binomial parameters, where △i‘s are to be of interest and Poi‘S are nuisance