# Today, we performed a likelihood ratio test comparing the
# saturated salary model with the unsaturated model that we
# estimated on Tuesday.

# Here's the log likelihood from the saturated model:
> ll1 <- -1303.602

# This is for the unsaturated model:
> ll2 <- -1303.806

# Here's the difference:
> ll1-ll2
[1] 0.204

# To get the chi-square statistic, we multiply the difference by two:
> 2*.204
[1] 0.408

# There are two constrained parameters in the unsaturated model, so the
# chi-square has two degrees of freedom.  The fit is not significantly worse:
> 1-pchisq(.408,2)
[1] 0.8154624