Mitotic recombination is the predominant mechanism for repairing double-strand breaks in Saccharomyces cerevisiae. Current recombination models are largely based on studies utilizing the enzyme I-SceI or HO to create a site-specific break, each of which generates broken ends with 3’ overhangs. In this study sequence-diverged ectopic substrates were used to assess whether the frequent Pol δ-mediated removal of a mismatch 8 nucleotides from a 3’ end affects recombination outcomes and whether the presence of a 3’ versus 5’ overhang at the break site alters outcomes. Recombination outcomes monitored were the distributions of recombination products into crossovers versus noncrossovers, and the position/length of transferred sequence (heteroduplex DNA) in noncrossover products. A terminal mismatch that was 22 nucleotides from the 3’ end was rarely removed and the greater distance from the end did not affect recombination outcomes. To determine whether the recombinational repair of breaks with 3’ versus 5’ overhangs differs, we compared the well-studied 3’ overhang created by I-SceI to a 5’ overhang created by a ZFN (Zinc Finger Nuclease). Initiation with the ZFN yielded more recombinants, consistent with more efficient cleavage and potentially faster repair rate relative to I-SceI. While there were proportionally more COs among ZFN-than I-SceI-initiated events, NCOs in the two systems were indistinguishable in terms of the extent of strand transfer. These data demonstrate that the method of DSB induction and the resulting differences in end polarity have little effect on mitotic recombination outcomes despite potential differences in repair rate.