A "cognote" system was developed for coding electronic discussion groups and promoting critical thinking. Previous literature has provided an account of the strategy as applied to several academic settings. This paper addresses the research around establishing the inter-rater reliability of the cognote system. The findings suggest three indicators of reliability: (1) raters assign similar grades to student's discussion group contributions; (2) raters predominantly assign the same cognotes to student's discussion group contributions; and (3) raters are selecting in excess of 50% of the same text in assigning the same cognotes. (Contains 18 references.) (AEF)