The first task of the study was to develop valid, short-cut methods of indexing the narrative content of evaluation reports that would extract the differentiating information contained in evaluative comments simply, reliably, and with as good classification accuracy as the longer initial procedure. In the second task, the original inter-indexer reliability study was extended to clarify the issue of reliability of the complex, lengthy indexing procedure. Section 2 of the report reexamines the pilot study sample, the cross validation sample, and the generalization sample. Section 3 reviews the original content analysis methodology and includes a description of the two shortcut indexing methods that were devised. In Section 4 the performance of the two shortcut indexing methods in classifying the three experimental samples into correct criterion groups is compared with that of the original lengthy indexing procedure. Section 5 presents the results of the extension of the original inter-indexer reliability study. In Section 6 future areas of investigation are delineated. (Author/MW)