writing anova results for a paper

#1
Hi all,

It is first time that I am trying to write a scientific paper based on statistical analysis and i am a bit confused.

The results from my anova analysis are

The GLM Procedure

Dependent Variable: ratio

Sum of
Source DF Squares Mean Square F Value Pr > F

Model 6 1.10360473 0.18393412 3.83 0.0135

Error 17 0.81724419 0.04807319

Corrected Total 23 1.92084892


R-Square Coeff Var Root MSE ratio Mean

0.574540 25.13621 0.219256 0.872271


Source DF Type III SS Mean Square F Value Pr > F

crop 1 0.24024558 0.24024558 5.00 0.0391
ni 1 0.59001854 0.59001854 12.27 0.0027
su 1 0.02786590 0.02786590 0.58 0.4569
ni*su 1 0.06913113 0.06913113 1.44 0.2469
replicate 2 0.17634358 0.08817179 1.83 0.1900


the model significance is higher than the significance in the type III error.

When I am writing the paper and refering to the individual factor effect should i use the signicance level of the Type III error (which lower than the model) or the p-value of the model?


thanks
 

bugman

Super Moderator
#2
depends on the jounral, but usually its the p-value:


for example, the GLM model showed a significant effect of.. (F1,34 = Fvalue; P=0.01).


I'd advise you to look at examples of the Jounral you are submitting to.
 
#3
well with significant difference I was refering to the p-value (sorry but I suck in statistics),
but I don't know if it should be the p value of the model or the p value of the seperate factor effects
 

bugman

Super Moderator
#4
Model if you are saying overall there was a difference.

If you are reporting on the specific factors then use the factor effects