Efficiency considerations have led to increased use of multiple-choice questions to assess economics understanding at the secondary and tertiary levels throughout Australia. A multiple-choice test would suffice if multiple-choice and essay questions measure the same dimensions of knowledge, as suggested by least squares estimation of the relationship between these two forms of testing. We show a simultaneous equation bias inherent in least squares estimation of the relationship between these two forms. A two-stage least squares estimation reveals no relationship, implying that these testing forms measure different dimensions of knowledge. Thus, a single form of testing economics knowledge must be avoided.