I used stata to run logistic regression, dependent variable is 被访谈的村民所选择的答案（1，或 0），在方程式的右边，控制了村dummy,但stata给出了这样的信息：                                                            
note: _Ivillageid_1522 != 0 predicts failure perfectly
         _Ivillageid_1522 dropped and 17 obs not used

note: _Ivillageid_3511 != 0 predicts failure perfectly
      _Ivillageid_3511 dropped and 21 obs not used

note: _Ivillageid_4121 dropped because of collinearity

高手帮我看看怎么解释啊？怎么把这些删掉的observations保留住？