Anova

[C Norton]

What p-value is the one I should use? I would like to know the difference in fisk_skaldjur for variable var394 adjusted for age. Is it p=0.1038 or p=0.0068 that I should use?


anova sqrt_fisk_skaldjur age var394

Number of obs = 268 R-squared = 0.2789
Root MSE = 3.57509 Adj R-squared = 0.0608

Source | Partial SS df MS F Prob > F
-----------+----------------------------------------------------
Model | 1013.51467 62 16.3470108 1.28 0.1038
|
age | 912.159713 61 14.9534379 1.17 0.2098
var394 | 95.7092491 1 95.7092491 7.49 0.0068
|
Residual | 2620.16543 205 12.7812948
-----------+----------------------------------------------------
Total | 3633.6801 267 13.6092888

[Roman Mostazir]

Your table is very difficult to read. Looks like your model is not significant as it has an F = 1.28, p=0.1038. One reason perhaps for your model being insignificant is to have age in your model. Age looks insignificant with F = 1.17, p=0.2098. var394 is highly significant with F=7.49, p=0.0068. Run the model again without age since it is not significant which will provide you a significant model . After running the model type "reg" command to obtain the parameter estimate of var394 if you need that. Consult the Anova help manual in Stata which containsall you need.

[Killian Mullan]

Is age continuous? If so, you want to estimate an ANCOVA. To do this, simply preface age with 'c.' (as in write c.age) so that Stata does not proceed 'thinking' that is it a categorical variable. You may also wish to mean-centre age.

[Richard Williams]

As already noted, the output is very hard to read. When writing your message, click on the A next to the smiley, and choose # (which stands for code). Enter the output between the two code tags.

I think Killian is right. It appears that age has 61 degrees of freedom, meaning it is being treated as categorical rather than continuous. Treating it as continuous will probably help a lot.