Fitting a generalized mixed effects model

Hi there,

I am trying to fit a generalized mixed effects model for data with a binomial distribution (binomial size = 1) yet, although I do get some model fit statistics, the "fitglme" function does not return any coefficients. The code that I am using is copied below and the data table is attached:

Lme = fitglme(Table,'Error ~ 1 + G0 + G1 + G2 + (1|Trials) + (1|Subjs)',...
    'Distribution','Binomial','BinomialSize',1,...
    'Link','logit',...
    'FitMethod','MPL');

Please help, Thanks very much.

Answer 1

[Sruthi Ayloo]

Hi Chris,

From your question, I understand that you want to fit a generalized effects model for your data with a binomial distribution of size 1. Using “table.mat” and the command you mentioned in the question, I see the following statistics when I display your model:

To access the co-efficients in a matrix format, you can use the following command:

Lme.Coefficients
ans = 

    Fixed effect coefficients: DFMethod = 'residual', Alpha = 0.05

    Name                 Estimate     SE          tStat       DF      pValue        Lower       Upper   
    '(Intercept)'          -1.9861     0.27292     -7.2772    7300    3.7621e-13     -2.5211     -1.4511
    'G0'                   0.23288    0.060126      3.8733    7300    0.00010833     0.11502     0.35075
    'G1'                 -0.031703    0.053529    -0.59227    7300       0.55369    -0.13663    0.073228
    'G2'                 -0.055686    0.042758     -1.3024    7300       0.19284    -0.13951    0.028132

Can you let the user community know what do you mean by “fitglme” does not return any co-efficients, as this will help us answer your question in a better way.

Refer to the following link for more information on how use "fitglme" with an example: fitglme

[dh34]

I'm sorry I'm very confused. I didn't ask this question and I don't recognise it at all. I find it odd, if someone had coopted my user details you would think they would do something more nefarious than ask complicated statistics questions with it.

I'm sorry your time was wasted, I will delete this question tomorrow.

Many thanks, Chris