I am doing an ANCOVA model in order to explain the gap in a given distance. So I have a control group, and I have several quantitative variables, right now I am trying to evaluate the impact of the quantitative variables individually. But I think the results I get are incoherent, lest see:

                           Df Sum Sq Mean Sq F value   Pr(>F)    
as.factor(groupe)           1  738.8   738.8  21.931 1.03e-05 ***
BASE0008                    1   36.6    36.6   1.087  0.29992    
as.factor(groupe):BASE0008  1  270.0   270.0   8.015  0.00576 ** 
Residuals                  87 2930.9    33.7 

These are the results from the ANOVA table, we could say that the group has a significant effect so does the interaction, but when I look at the results of the regression model, I find this:

                              Estimate Std. Error t value Pr(>|t|)   
(Intercept)                    1.16666    3.38404   0.345  0.73111   
as.factor(groupe)NAV           4.48191    3.79515   1.181  0.24084   
BASE0008                       0.26027    0.08651   3.009  0.00343 **
as.factor(groupe)NAV:BASE0008 -0.26902    0.09503  -2.831  0.00576 **

Well, the interaction is still relevant, but it looks like there is not effect of the group and the quantitative variable is more important to determinate the output of the experiment. I want to know if my interpretation is accurate: What can I say about the group?

  • $\begingroup$ Are these any help? stats.stackexchange.com/questions/55571/… and also stats.stackexchange.com/questions/53432/… $\endgroup$ – Glen_b Apr 10 '13 at 7:42
  • $\begingroup$ thanks! I have read both threads, but I have still problems with the interpretation. So, the group influence is significant, but the effect between its categories is not different? I don't if that makes sense... and I think it's weird because I have only two categories $\endgroup$ – eyanquenb Apr 10 '13 at 8:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.