The study of a set of 20 asphalt cements reported in a recent publication is modified and extended to cover additional parameters and a precise statistical evaluation of correlations. For a given asphalt, the whole, original material is subjected to high performance gel permeation chromatography (HPGPC) using a gravimetric finish. The experimental conditions used are kept uniform for all molecular size distribution (MSD) analyses in this study. Another part of asphalt is fractionated by ASTM D 4124. Each Corbett fraction is subjected to HPGPC analysis. The asphaltene compatibility index (ACI) is determined as proposed by Branthaver et al. A part of original asphalt is subjected to 16-hr TFOT, and MSD of the aged sample is determined. Two asphalts from the same refinery showed MSD and ACI data that are outliers with respect to the other 18 asphalts. For many correlations, excluding these two asphalts causes the correlation coefficient, r, to change appreciably because r is not resistant (ie, one or more extreme entries can be influential). Several correlations are evaluated. Depending on the number of asphalts involved, n, the numerical value of r indicates whether a correlation is not significant, significant, or highly significant; the sign of r indicates the direction of slope of the regression line. Considering only the significant and highly significant correlations, the sign of associations tested indicate that a high LMS/MMS ratio [the molecular size index (MSI)] for original asphalt accompanies a low ACI, a minimum temperature susceptibility, a high resistance to aging (minimum change of MSI due to 16-hr TFOT), and a high viscosity, 135 deg C. From the MSD of original asphalt, eight with the highest MSI show common parameters that predict a promising field performance.