Procedures Employed in the Study
This study used a computer program for qualitative research to facilitate analysis of the debates: The Ethnograph. This program requires computer text files for the discourse being studied, and these files must be formatted into a particular format. Computer files of the transcripts of the three 1992 presidential debates were edited to prepare them for this analysis (one file for each debate.). First, all questions and comments by the moderators were eliminated: only the utterances by the candidates were coded (of course, transcripts of both questions and answers were kept to help interpret and code the candidates' responses where necessary). Ethnograph does not require this procedure, but the computer files for each of the three debates were fairly long. Because we were coding only the candidates' utterances, and not the questions, we decided to delete all other comments to shorten the three files.
Second, in order for the computer program to be able to process the transcript files, margins were changed. Each candidate's turn at talk was formatted into a hanging indent, with the beginning of each utterance flush left, and the remainder of the turn indented. A right margin of forty left room for Ethnograph to number each line of the transcript. This margin also provided space for Ethnograph to insert coding symbols into the file on each numbered line.
Furthermore, symbols were inserted to identify the beginning of discrete segments of the transcripts (i.e., the beginning of a candidate's turn). Ethnograph uses a plus sign (+) to identify the beginning of a turn at talk. We provided two bits of information for each turn at talk. We labeled each turn as D1, D2, or D3, depending upon whether the transcript belonged to the first, second, or third debate. Also, we identified the candidate who was