Transcript Reliability Cleaning Percentage: An Alternative Interrater Reliability Measure of Message Transcripts in Online Learning Peter K. Oriogun and John Cook London Metropolitan University In this article, we extend previous work with respect to interrater reli- ability measure of computer-mediated conferencing and suggest cod- ing categories relevant to problem-based learning. Calculating interrater reliability agreement by using a Transcript Reliability Cleaning Percentage (TRCP) approach is simple for academics with limited mathematical background and can provide insights into the na- ture of the learning process from the prospective of categorization of online discourse. TRCP enhances interrater reliability (percentage agreement between coders) of the rating of online transcripts. Research on dialogue analysis has explored the relation between online di- alogue features (e.g., roles, strategies, form, and content) and learning (Pilkington 2001). Such an analysis can provide useful insights into the na- ture of the learning processes from the perspective of, for example, what a speaker’s intention is in a transmitted message and what the receiver per- ceives has been communicated by the message. However, problems can arise if one attempts to investigate specific categories or variables of the learning process—for example, participation, interaction, social elements, cognitive elements, and metacognitive elements (Henri 1992). In the case of coding protocols that include several categories, coders may not agree on interpretations. For this reason, some researchers (e.g., Potter and Le- 221 THE AMERICAN JOURNAL OF DISTANCE EDUCATION, 17(4), 221–234 Copyright © 2003, Lawrence Erlbaum Associates, Inc. Requests for reprints should be sent to Peter K. Oriogun, London Metropolitan Univer- sity, Department of Computing, Communications Technology and Mathematics, London North Campus, 2-16 Eden Grove, England. E-mail: p.oriogun@londonmet.ac.uk