phson

I support more condition for this issue a,b,c >=1 ; d >=2 a+b = D1; c+d = D2 (D1 and D2 are constant)

Dear all, I have a problem with comparing two lower confidence intervals of odds ratio. The detail is follow: I have: lci1 = e^[ln(a*d/b*c)-1.96*square root(1/a+1/b+1/c+1/d)] lci2 = e^[ln(a*(d-1)/b*(c+1) - 1.96*square root(1/a+1/b+1/(c+1)+1/(d-1)] I want to approve lci1 > lci2 with all 0 < a,b,c,d < N I have progammed this problem in computer. As the result, it is true ( lci1 > lci2 ). We want demonstrate this inequation in mathematisc style. Could you please help me? Thank you in advance! Best,