Key points are not available for this paper at this time.
This paper argues that the text processing and interpreting capabilities of computers can and should be applied to the problems of managing and analyzing nonnumeric social data. It identifies behavioral records such as observation records or open-ended interview responses as good candidates for such an application because they are text based and at least semistructured. It describes one specific implementation of a behavioral record data analysis program. In contrast to most analyses of verbal records that begin by first transforming them into numeric form, the method described here preserves the data in verbal form. Notable features of computer-based processing of verbal data are reducing researcher clerical tasks, keeping attribute descriptors embedded in their context, and preserving the information content of descriptors.
Sproull et al. (Wed,) studied this question.