SPSS Statistics

Expand all | Collapse all

Poisson output- tests of model effects

  • 1.  Poisson output- tests of model effects

    Posted Tue January 19, 2021 12:58 PM
    Dear colleagues,
    I'm a bit confused about which p value considering in a GLM poisson regression.
    I have some ordinal independent variables and a bivariate nominal outcome.
    When looking the "tests of model effects" I get p=0.156, for example, but in the "parameter estimates" one of the categories has a p less than 0.05. In this case, should I consider this variable not statistically significant (tests of model effects) or forget that and consider the parameter estimates (report the prevalence ratio of categorie is higher/lower than the reference group)? 
    Sometimes same happens in the opposite way (tests of model effects is statistically significant p<0.05 but when checking the parameter estimates the categories are not statistically significant p>0.05). Any help?
    Thanks in advance!

    ------------------------------
    Rafael Mello
    ------------------------------


  • 2.  RE: Poisson output- tests of model effects

    Posted Tue January 19, 2021 04:50 PM
    Hi everyone,
    I already got the answer from Jon Peck!

    Thank you Jon.

    ------------------------------
    Rafael Mello
    Federal University of Technology - Parana, Brazil.
    Post-Graduate Program in Physical Education.
    ------------------------------