Sites were selected according to several criteria: absolute number of habeas petitions filed in Federal court; the proportion of habeas petitions to State prisoner population; geographic diversity among sites; States with a death penalty and States without a death penalty; and States having the court records necessary to complete the study. Samples of habeas petitions were drawn from each State's highest criminal courts to ensure a sample that was representative of petitions disposed in the State as a whole. The Advisory Committee of this research effort helped staff identify the data elements required to test key propositions. A coding sheet of minimum essential data elements was prepared, reviewed, refined, and pretested in Arizona. Implications of the pretest for the research design are discussed here. 2 tables and 11 notes