A database of dynamic modulus values for typical Superpave mixes widely used in the state of Washington was developed and used to investigate the sensitivity of the dynamic modulus to aggregate gradation. Seven JMF mixes were selected for the study. Percent passing sieve #200 of these mixes were changed by ± 2% to prepare additional mixes referred to as ‘modified mixes’. Statistical analysis showed that using different JMF mixes significantly affected the dynamic modulus. This was not the case when modifying the JMF by changing the percent passing #200. A simple evaluation of the prediction accuracy of the 2002 Mechanistic-Empirical Pavement Design Guide (MEPDG) was conducted in which the measured dynamic modulus was used as a key input parameter to predict pavement distresses. Level 1 and Level 3 predictions of rutting, longitudinal cracking, alligator cracking, and IRI were compared with the field performance data. The Design Guide predicted IRI and alligator cracking reasonably well. The predicted rutting of the JMF mixes agreed well with the dynamic modulus trend. It was found that the ±2% change in the passing #200 aggregate gradation did not affect the predicted distress significantly.