I have fitted a glm Poisson to my frequency data and obtained the result:

Null deviance: 657.49 on 583 degrees of freedom

Residual deviance: 575.00 on 571 degrees of freedom

AIC: 1534.4

Is the high AIC value and relatively high residual deviance a serious matter to be taken into consideration? Thanks

  • $\begingroup$ Thanks but still it is mentioned there the lower the better AIC that what i wanted to confirm here an also on the residual deviance which has not been discussed there @ Andy $\endgroup$
    – user40494
    Mar 1, 2014 at 13:33
  • 2
    $\begingroup$ What makes you assert that the AIC is high? I already explained to you on a previous question that that residual deviance is not high. Do you think the answer will be different on different days? $\endgroup$
    – Glen_b
    Mar 1, 2014 at 15:05

1 Answer 1


AIC is used in comparing models; a single AIC has, as far as I know, little utility.


Not the answer you're looking for? Browse other questions tagged or ask your own question.