Mapping discursive dynamics of the financial crisis: a structural perspective of concept roles in semantic networks

Abstract

Background/purpose: Convenient access to vast and untapped collections of documents generated by organizations is a highly valuable resource for research. These documents (e.g., press releases) are a window into organizational strategies, communication patterns, and organizational behavior. However, the analysis of large document corpora requires appropriate automated methods for text mining and analysis that are able to take into account the redundant and predictable nature of formalized discourse. Methods: We use a combination of semantic network analysis and network centrality measures to overcome these particular challenges and to explore the dynamic structural space of concepts in formalized documents pertaining to the recent financial crisis. Data: For our analyses, we collect the press releases of the European Central Bank (ECB) and the United States' Federal Reserve System (Fed) issued between 2006 and 2013 in order to examine their semantic networks before, during, and after the recent financial crisis. Their press releases are notably impactful in their influence on other financial institutions and society at large, especially during times of financial volatility. Results: The structural space created from joint centrality metrics reveals salient shifts in the discursive practices of the ECB and Fed. In particular, the Fed exhibits greater attentiveness to the financial crisis especially during the crisis itself, while the ECB's attention is delayed and increasing steadily. Furthermore, we show both the Fed's and the ECB's discourse transitioning into a new "hybrid" state, rather than returning to the pre-crisis status quo. Conclusions: Examining the semantic networks of organizational text documents, we find that our analytic approach reveals important discursive shifts, which would not have been discovered under traditional text-analytic approaches. We demonstrate the utility of this approach in investigating large text corpora of organizational discourse, and we anticipate our methods to be comparably valuable in the analysis of a large spectrum of formal and informal discourse.

Local meaning structures: mixed-method sociosemantic network analysis

“…We focus on how individuals associate (collocate) words in their verbal statements, e.g., within a sentence (Sinclair, 1991). This approach aligns with co-occurrence-based semantic network analysis (Carley, 1994;Diesner, 2013;Lee & Martin, 2015;Nerghes et al, 2015), a technique starting with a corpus of texts rather than with a set of predefined categories and producing maps of associations between word stems-traditionally named 'concepts'-as they appear next to each other within a certain interval in these texts. Because such co-occurrences are rarely controlled by the speakers, this technique is capable of capturing the associations put forward unconsciously, among others.…”

“…Secondly, we applied Porter's stemming procedure for the Russian language (all textual data was collected in Russian) to aggregate variants of the same word, e.g., singular and plural forms, to their stems by removing suffixes (Porter, 1980). 8 These generalized forms of aggregated words are the nodes in our semantic network, referred to as 'concepts' (see Carley, 1986;Diesner, 2013;Lee & Martin, 2015;Nerghes et al, 2015). Table 1 presents the number of concepts per member per text type.…”

“…To understand (Verstehen) these focal meaning structures in the practical context of the group, we use ethnographic data and qualitatively examine how the corresponding associations are used in practice. Simple and straightforward as this method may be, such an approach has, to the best of our knowledge, not yet been proposed (compare, e.g., with Erickson, 1988;Schultz & Breiger, 2010;Vaisey & Lizardo, 2010;Nerghes et al, 2015;Basov & Brennecke, 2017;Godart & Galunic, 2019).…”

Social Networks and Construction of Culture: A Socio-Semantic Analysis of Art Groups

“…Using semantic network analysis based on word collocation [34,6,32,16], we trace cultural constructs as patterns of associations between concepts expressed by individuals' and relate the properties of those cultural constructs to positions in net-works of social ties occupied by individuals. Hence, we apply the growingly popular socio-semantic framework [32,31,29,27].…”

Predicting New TV Series Ratings from their Pilot Episode Scripts

“…Several distinct approaches exist for creating networks from texts (Nerghes, Lee, Groenewegen, & Hellsten, 2015). They differ along a number of dimensions including the level of automation, whether and how words are abstracted to higher-order conceptual categories, and the nature of the underlying relationship used to connect the words or concepts.…”