1

How could that be possible? The model is significant too. The null model already predicted 70% correctly. After adding predictors, the model predicts 100%. But no single predictors are significant. I used SPSS. Here are the results captures enter image description hereenter image description hereenter image description hereenter image description hereenter image description hereenter image description here

Gemini
  • 51
  • 1
  • 1
  • 5
  • 1
    Welcome, Gemini! You should probably fill us all in with some more details: what software are you using? What is the exact syntax of the command for logistic regression you are using? Can you copy in the results (this will help in questions of interpretation). You can edit your question by clicking the "edit" link in the lower left. – Alexis Dec 11 '14 at 20:04
  • 1
    Depending on details requested from the OP, http://stats.stackexchange.com/questions/11109/how-to-deal-with-perfect-separation-in-logistic-regression might contain the answers. – whuber Dec 11 '14 at 20:24
  • 1
    Thank you for the clarification. As the final contingency table shows, you have achieved *perfect separation.* The duplicate (and other threads it links to) addresses this situation. – whuber Dec 11 '14 at 20:37
  • What does perfect separation suggest? What should I do about it? I tried to eliminate some of the variables. But it did not work well. – Gemini Dec 11 '14 at 21:03

0 Answers0