Despite the general recognition that incumbency has influenced voters' deci
sions to split their ballots for president and the House, past research has
not focused on the specific magnitude of this effect and its responsibilit
y for growing ticket-splitting in the United States. In this study, I find
that incumbency was a powerful determinant of the step jump in ticket-split
ting that occurred from the 1956-68 to 1972-92 periods. This is in contrast
to the weak expansive force exerted by declining partisan intensity in the
electorate. Incumbency's impact, however, was confined to districts where
members of the losing presidential party run for reelection; in districts w
ith campaigning incumbents of the winning presidential party, it made for o
nly about the level of ticket-splitting that could be expected in open seat
s.