As the previous chapters indicate, our study concerns the cases that altered precedent and the precedents that were altered. Researchers studying stare decisis would not necessarily examine these cases. They might, for example, investigate the cases that followed precedent, or those in which one of the parties asked the Court to overturn precedent. The first enterprise would be exceedingly difficult to undertake because it is not easy to distinguish between the citation of a precedent and adherence to it. Even Shepard's Citations, the most authoritative source of this information, does not set forth the guidelines it uses to assemble its list. The second venture would entail an examination of written briefs over a period of time. This would depend on access to the briefs themselves and is also difficult and time consuming. How, for example, should the consequences of a litigant's request to overturn precedent be determined? What about cases overturned even when no litigant so requests?

Both kinds of research might be worth doing. We, however, chose a different path: cases that formally alter precedent. In choosing this course, our initial problem is to select a list of overruling and overruled cases.

As indicated in Chapter 2, Banks and Epstein turned to the Congressional Research Service (CRS) to obtain a list of overturned and overturning cases. This list is unreliable, however, for one fundamental and four ad hoc reasons. Fundamentally, those who compiled it failed to indicate the guidelines which they used and which others could follow to determine the accuracy of the list.