Deviance for binary regression models for expected number of successes smaller than five?

I have a single response, which denote the number of faulty products/total products (success, or failure ratio) over quite a number of factors (~15). Since these 15 factors are not...

30 January 2015 2,270 11 View