bourn 發(fā)表于 2012-4-20 11:20
如圖。
this test is to compare the fitted model to the saturated model(full model). since the p is highly sig, it implies that the current model may NOT be a good fit.
since you mentioned it is a logit model,so i guess the number of response levels =4 (3+1) like the upstairs said. but i have trouble to see how many populations in your data because the variable names look like being manipulated already.
anyway, since so many dfs left, you may still have room to fit more variables in, such as crossing ones.
jingju