Second Order Generalized Estimating Equations 
    for Correlated Binary Data (v0.5-pc)
      (C) Bahjat Qaqish, 1990

--  Title1: test of GEE2 for the PC --
--  Title2:  --
Data file: test3.dat
dim(x):            2
dim(b):            3
# classes          2
max #iter:        10
tolerance:  0.0010000000000
# clusters:       50
cluster size          #clusters
     5                     50
The model:
1 : B1 X1  B2 X2  
2 : B1 X1  B2 X2  
1.1 : B3 X1  
1.2 : B3 X1  
2.2 : B3 X1  

Method of computing 3rd and 4th order moments: The exact method.
-- GEE2 --

monitorb: iter =    0  b = 
       1   1.00000
       2   1.00000
       3   0.10000
monitorb: iter =    1  b = 
       1   0.83765
       2   1.06489
       3   0.74968
monitorb: iter =    2  b = 
       1   0.86078
       2   1.02693
       3   0.62425
monitorb: iter =    3  b = 
       1   0.85791
       2   1.03209
       3   0.62273
monitorb: iter =    4  b = 
       1   0.85789
       2   1.03217
       3   0.62245
Converged after     4 iterations.
Method of computing 3rd and 4th order moments: The exact method.
-- GEE2 --

Estimates:
Beta:                    Estimate    Robust s.e.     Naive s.e.     Robust z
1 Yi:1                  0.85789        0.22747        0.21845        3.77138  
2 Yi:x                  1.03218        0.29357        0.32668        3.51598  
3 logOR:1               0.62244        0.32293        0.33791        1.92745  
(Lower = Correlation \ Upper = Variance)
Robust Correlation \ Variance Matrix Estimate:
( 1)  0.05174  -0.03931  -0.00983  
( 2) -0.58868   0.08618   0.01385  
( 3) -0.13376   0.14607   0.10429  

Naive  Correlation \ Variance Matrix Estimate:
( 1)  0.04772  -0.03638  -0.00102  
( 2) -0.50980   0.10672  -0.00210  
( 3) -0.01387  -0.01902   0.11418