The main objective of this research is to find out the effect of deviation in the aggregate gradients of asphalt mixtures from the Job Mix Formula (JMF) on the general mixture performance. Three road layers were worked on (wearing layer, binder layer, and base layer) and statistical analysis was performed for the data of completed projects in Baghdad city, and the sieve that carried the largest number of deviations for each layer was identified. No.8 sieve (2.36mm), No.50 sieve (0.3mm), and 3/8'' sieve (9.5mm) had the largest number of deviations in the wearing layer, the binder layer, and the base layer respectively. After that, a mixture called Mix 1, was made. This mixture was selected from a number of completed mixtures, and it represents the worst mixture. Mix 1 was compared with two other mixtures, Mix 2 and Mix 3, Mix 2 represents the mean of JMF for the gradients of aggregates, and Mix 3 is the same as Mix 1 except for the sieve that contains the largest number of deviations, so the gradient of aggregates for it is the mean of JMF. Fifteen Marshall specimens were made for each mixture and for each layer in order to know the differences in Marshall properties between the mixtures. Also, 6 specimens were made for each mixture (the total is 18 specimens for each layer) to check the indirect tensile strength, for the purpose of knowing the mixtures' susceptibility to moisture. Finally, 1 specimen was made for each mixture for repeated load test for the purpose of knowing the performance of the mixtures with respect to permanent deformation. The tests showed that the performance of Mix 2 and Mix 3 was improved in comparison with Mix 1. The deviation of the aggregate gradients in specific sieves may be higher than the limits of the JMF or it may be less, and in both cases, the implementation of a mixture like Mix 1 for the streets is bad for the performance of the road and failures occur due to the wrong implementation of the JMF. On the other hand, there are much better mixtures in all respects such as Mix 2 and Mix 3, and if they are implemented on the streets, they will certainly have much better.
