The categories of reasoning described in each of the chapters were arrived at through an inductive, iterative procedure. The 160 subjects in the main sample produced a total of 480 protocols for the three topics. For each major skill examined (causal theories, supporting evidence, alternative theories, counterarguments, rebuttals), the procedure began with random selection of a set of approximately 20 protocols (representing all topics, as well as age, sex, and education groups). Copies of these protocols were distributed to the group undertaking category development for a major category (which included the author and one to three others). Each member of the group analyzed the protocols independently and devised a set of categories sufficient to account for the reasoning represented in these protocols. The group then met to compare category systems and to devise collaboratively a set of categories agreed on by the group. A second set of protocols was then randomly selected, copied, and distributed to the group for coding. As part of this second iteration, the group again met for reconciliation of the independently conducted coding and for further development and refinement of the category system. In addition, all protocols from earlier iterations were reviewed and recoded as necessary. This iterative procedure was continued until agreement was reached regarding the category system and the categories could be applied to a new set of protocols with high agreement (typically four to six iterations).