It occurs to me that the better example might be 1976. In that year, Gerald Ford and Ronald Reagan fought each other all the way until the convention in August. Not only that, but the contest was pretty bitter. Ford accused Reagan of wanting to gut Social Security and of being reckless when it came to foreign policy. Reagan, for his part, accused Ford of being soft on the Soviets and for wanting to give away the Panama Canal. Nonetheless, Ford managed to close the gap with Jimmy Carter and only narrowly lost the election. Had Ford not freed Poland in one of the debates, he may well have won the election.All things considered, Ford should have lost the 1976 election in a landslide. The fact that he didn't — that, in fact, he might actually have won if not for a single gaffe — suggests that primary divisiveness that year had no noticeable effect.
Over at the Monkey Cage, John Sides glosses the academic research on this issue:
Perhaps the most relevant study is by Lonna Rae Atkeson (here, gated). She examines presidential elections from 1936-1996 and finds that the relative divisiveness of the two parties primaries is not related to the general election outcome, once other factors, namely the state of the economy and the popularity of the incumbent president, are taken into account.Obviously none of this is conclusive, merely suggestive. But what it suggests is that primary divisiveness just isn't a major factor once the primary is over. I doubt that the Obama/Clinton nastiness will prove any different.