Dear Statistics Professionals: I am recently working some data analysis and have great difficuly to find the approriate statistics method. The analysis is: I have one binary variably (y=1 or y=0) and one continuous variable, say 'X'. 'X' range from 0-100, with abou25% of them equal to 100 and 95% of them above 85. Now I need to find the 'best' region of 'X' (somthing like from a to b) in which that the event rate (y=1) is the lowest. I can always plot x vs. y and find the 'best' region visually and then get some p value out of it. But then it doesn't sound like statistics to me. There are about 500 observation, by the way. Could somebody kindly provide some suggestion? Many Thanks. Gang