The conventional wisdom in the partisan change literature predicts that increasing party conflict on one issue agenda leads to a decline in party conflict on another agenda—a process called “conflict displacement.” We have argued that recent party politics in the United States has experienced “conflict extension,” with the Democratic and Republican parties in the electorate growing more polarized on cultural, racial, and social welfare issues, rather than conflict displacement. Here, we suggest that the failure of the literature to account for conflict extension results from incomplete assumptions about individual-level partisan change. The partisan change literature typically considers only issue-based change in party identification, which necessarily leads to the aggregate prediction of conflict displacement. This ignores the possibility of party-based change in issue attitudes. If party-based issue conversion does occur, the aggregate result can be conflict extension rather than conflict displacement. Our analysis uses data from the three-wave panel studies conducted by the National Election Studies in 1956, 1958, and 1960; in 1972, 1974, and 1976; and in 1992, 1994, and 1996 to assess our alternative account of individual-level partisan change. We show that when Democratic and Republican elites are polarized on an issue, and party identifiers are aware of those differences, some individuals respond by adjusting their party ties to conform to their issue positions, but others respond by adjusting their issue positions to conform to their party identification.