Table 2. Random effects probit model results.
| Main effecta (βj) | s.e. | 95% Cl | P>∣z∣ | |
|---|---|---|---|---|
| Information | 0.997 | 0.113 | 0.776–1.218 | 0.000 |
| Surveillance | 0.463 | 0.036 | 0.393–0.533 | 0.000 |
| Preparation | 0.267 | 0.045 | 0.182–0.353 | 0.000 |
| Direction | 0.114 | 0.047 | 0.023–0.206 | 0.014 |
CI=confidence interval.
Individuals=193; observations=965.
Log-likelihood=−510.209; Wald χ2(4)=250.15; Prob.>χ2=0.000.
a
An alternative approach to estimating the main effects model is to use dummy variables for categorical attribute levels in the independent variable set. However, this approach becomes problematic owing to colinearity in the set of dummy variables.