ANOVA result No p-value.

4 visualizaciones (últimos 30 días)
Kaminosono Shougo
Kaminosono Shougo el 24 de Nov. de 2022
Comentada: Kaminosono Shougo el 17 de En. de 2023
I was using ANOVA, but when I changed the display of results from "interaction" to "full", the p-value was not displayed.
I would like to know if anyone knows what the problem is.
dataFileName1 = 'AllParameter_Exp2';
dataFileID1 = fopen([dataFileName1,'.txt'],'r');
if (dataFileID1 == -1);
error('data file not exist');
end
formatSpec = '%f';
y = fscanf(dataFileID1,formatSpec)
y = 90×1
58.6870 65.6794 44.2342 36.2030 0.0007 45.4349 0.0277 44.6501 54.4564 0.0011
fclose(dataFileID1);
g1 = [1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10,1,2,3,4,5,6,7,8,9,10];
g2 = [1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,3,3,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,3,3,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,3,3];
g3 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3];
%{
par = num2cell(g1);
sti = cellstr(g2);
sub = cellstr(g3);
%}
p = anovan(y,{g1,g2,g3},'Model','full','Varnames',{'parameter','stimuli','subjects'});

Respuesta aceptada

Jeff Miller
Jeff Miller el 24 de Nov. de 2022
The problem is that the full model has 90 parameters and predicts a separate mean for each of the 3*3*10 conditions. But you only have 90 observations, so there are no left-over degrees of freedom to estimate an error term. No error term => no F's or p's.
Your only options are to (a) fit a simpler model (e.g., with just 'interaction') or (b) get multiple data values within at least some of the 90 cells so that you have more data than cells and thus some d.f.'s to estimate an error term.
  1 comentario
Kaminosono Shougo
Kaminosono Shougo el 17 de En. de 2023
Thank you! This problem have been solved.

Iniciar sesión para comentar.

Más respuestas (0)

Etiquetas

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by