Developing a taxonomy for strategising in industrial networks

The purposes of the project described here were (1) to develop a taxonomy of terms relating to strategy used in industrial networks research studies, and (2) to compare manual qualitative content analysis with a computer-assisted text mining approach to taxonomy creation in a social science context. The unit of analysis was abstracts from the IMP research database (publicly available at www.impgroup.org). The main sample used in the analysis comprised 107 abstracts that contained 'strategy' as a keyword. There were marked similarities between the lists of key terms generated by the manual content analysis and by the text mining approach. Where there were differences between the lists of key terms, it was not possible to say whether these were because of unconscious biases in the manual analysis (analysts finding what they expected to find), or because of inadequacies in the text mining approach (which can only identify terms that exist within the data and cannot 'understand' meanings that are implied, but not explicitly stated, by authors).