Vehicle chassis dynamometer tests were performed to compare predicted and measured total feedgas emissions and fuel economy for dynamic operation of an engine. In general, these tests showed that predictions based on steady-state mapping data agreed well with measured values except for NOx emissions. Subsequent engine-dynamometer tests indicated that the discrepancy between predicted and measured NOx emissions was due to competing effects of combustion chamber wall temperature and dynamic EGR time response. A technique was developed which utilized the results of a simple transient test to improve the accuracy of predicting NOx emissions when EGR time response was not a factor. The effect of degraded EGR time response on both instantaneous and total NOx emissions was also explored.