yes the constant is absorbed into other parameters in extended models
Since it's gone, shouldn't the degrees of freedom come back?yes the constant is absorbed into other parameters in extended models
But if there is no previous model, I may deduct 1df less, so I hope to find out why the constant of the later model is merged, but it needs to be calculated.no you are looking at nested models, each one building on a previous model, so you deduct the additional parameters from the degrees of freedom each time
You start with 2 for YO say
YO*FS*TC=YO+FS+TC
+YO.FS+FS.TC+YO.TC
+YO.FS.TC
=2+(2-1)+(2-1)
+(2-1)*(2-1)+(2-1)*(2-1)+(2-1)*(2-1)
+(2-1)*(2-1)*(2-1)=8
8 parameters
The starting point for a variable 2 for a factor the number of categories or for a constant 1
You only deduct the additional parameters each time for degrees of freedom
The constant is there but it has been combined into another parameter in the revised model
Hi all,
sorry to restart this thread. just wondering, how do we know the initial dof? i understand from your response that the inital parameter for variable is 2, cosntant is 1 and factor is n and so would know what to take away from the dof, but am unsure what this would be to start with?
Also, what do parameters and dof represent in this context?
thanks so much in advance
You would start with some n, where n is the number of data points, so in this question I assume it was 8, but it did not say instead it gave the starting model as 7 degrees of freedom
Hi ykai and andrea,
thank you so much for your help.
you have both cleared up my confusion on this!