Likelihood Ratio Test

From Q
Jump to: navigation, search

The test statistic is:

X^2 = 2 (LL_u - LL_r)

where:

LL_u is the log-likelihood of the unrestricted model estimated using the Calibrated Weight,
LL_r is the log-likelihood of the restricted model estimated using the Calibrated Weight,
p \approx \Pr(\chi^2_{k_u - k_r} \ge X^2) , and
k_u and k_r are the number of parameters of the unrestricted and restricted models respectively.

See also