The method of Mattox et al. (1996ApJ...461..396M) for identifying EGRET sources with 5GHz radio sources has been applied to the 3EG EGRET catalog of Hartman et al. (1999, Cat. <J/ApJS/123/79>). Complete results are tabulated. We tabulate separately 46 EGRET sources with radio identifications which we expect to have a high probability of being correct. We suggest that these sources are appropriate for most studies of the properties of {gamma}-ray blazars as a class. All but one of these 46 sources were classified by Hartman et al. (1999, Cat. <J/ApJS/123/79>) as high-confidence identifications; and the additional source was classified by Hartman et al. (1999, Cat. <J/ApJS/123/79>) as a plausible identification. We also tabulate separately 37 additional "plausible identifications of EGRET sources with radio sources." These less secure possibilities include the remaining 21 "high-confidence identifications" of Hartman et al. (1999, Cat. <J/ApJS/123/79>), three of the 27 "lower confidence potential blazar identifications" of Hartman et al. (1999, Cat. <J/ApJS/123/79>), and an additional 15 plausible identifications which have not been suggested previously. Many of these sources require new radio and optical observations to establish them as blazars. We suggest that the 23 "lower-confidence potential blazar identifications" of Hartman et al. (1999, Cat. <J/ApJS/123/79>) which we do not find plausible should not be used in studies of the properties of the EGRET blazars. For this analysis, we have made elliptical fits to the 95% confidence contours of the position uncertainty regions of the 3EG sources.