Our analyses thus far have established that when the justices formally alter precedent they vote markedly more compatibly with the attitudinal model than with the legal model. But we have not determined what attitudinal variables cause them to behave this way. We suspect that it may concern the ideology of the individual justices, and by extension, that of the Court as well. If ideology explains the justices' attitudinal voting, that ideology is most likely to be the one that describes the behavior of other American political elites: liberal and conservative. Thus, we hypothesize that conservative Courts and justices will vote to overrule liberal decisions and vice-versa.

We do not expect a perfect fit, of course. Conservative and liberal Courts will occasionally overturn decisions of the same ideological stripe. And of course over time the definition of “liberal” and “conservative” will change. As a result, we expect that overrulings of nineteenth or early twentieth century decisions may pertain to issues that either do not lend themselves to today's ideological definitions or do so only tangentially. But overall we anticipate an inverse relation between a Court's ideological direction and that of the decisions it overrules. We also expect the same inverse relation for the individual justices (e.g., conservative justices vote to overrule liberal decisions, and vice-versa). We define “liberal” and “conservative” compatibly with conventional usage.