Covratio

Fit Analyses

Covratio

Covratio measures the effect of observations on the covariance matrix of the parameter estimates. For linear models,

$C_{i} = \frac{| s^2_{(i)} ({X_{(i)}'} X_{(i)})^{-1} |}{| s^2 ({X'}X)^{-1} | }$

where X_(i) is the X matrix without the ith observation.

Values of C_i near 1 indicate that the observation has little effect on the precision of the estimates. Observations with ${|{C_{i}}-1| {\ge}3p/n}$ suggest a need for further investigation.

For generalized linear models,

$C_{i} = \frac{| \hat{ \phi}_{(i)} ({X_{(i)}'} W_{(i)} X_{(i)})^{-1} |}{| \hat{ \phi} ({X'}W{X})^{-1} | }$

where W_(i) is the W matrix without the ith observation, W = W_o when the full Hessian is used, and W = W_e when Fisher's scoring method is used.

The Covratio statistics are stored in variables named C_yname for each response variable, where yname is the response variable name.

Top of Page