Introduction The combined use of quantitative and qualitative methods is almost as old as sociological research, but this combination has only recently come to the forefront of the methodological debate. This is shown by the large number of reference works advocating a high level of methodological integration (Tashakkori and Teddlie 1998, 2003; Creswell 2003; Brewer and Hunter 2006; Creswell and Plano Clark 2007; Bergman 2008) that have been published since the late 1990s. However, works of this type have often limited their focus to the stage of gathering data. The methodological literature is fairly lacking in presenting and discussing strategies of analysis in which the data analysis is neither strictly quantitative (mathematical) nor strictly qualitative (interpretive). This chapter presents and discusses one example of this kind of analysis as applied to narrative interviews. More precisely, this chapter presents an analysis procedure in which, from information obtained through qualitative techniques (narrative-biographical interviews) matrices of relations between actors are drawn up and analyzed using standard (quantitative) procedures of social network analysis. It is important to note that this transformation of narrative information from the interviews into a matrix of quantified data is preceded by a preliminary stage in which an interpretively generated code takes into account the syntactic and semantic nature of the text of the interviews. This strategy prevents loss of information about the content of the texts and respects the articulation of the textual units. The second stage, of transforming the already interpreted qualitative data to matrix form and submitting them to the corresponding algebra, can be described, in mixed methods terminology, as a “quantitizing strategy” (Tashakkori and Teddlie 1998:126; Onwuegbuzie and Teddlie 2003:355) or a “quantitative translation” (Boyatzis 1998:129). The particularity of our analysis lies in the fact that the quantitative data obtained are of a relational rather than a purely statistical-attributive nature.
Stored only when both papers are in the corpus — 9 citations exist upstream.