A simple deterministic macroeconomic computer model was tested in a junior-senior level economic statistics course to demonstrate how and why some common errors arise when statistical estimation techniques are applied to economic relationships in empirical problematic situations. The computer model was treated as the true universe or real-world system, and each student was required to use it to generate his own sample data. The class experiment was successful in that few of the students were able to obtain by computer perfect statistical estimates, since most results possessed estimation errors similar to those found in empirical analysis. Weaknesses of the experiment are discussed, and the macroeconomic model and sample data generated by students are presented in figures. (CMV)