DescriptionNetwork text analysis proves useful in distinguishing organizational forms, particularly in capturing intra- and inter-organizational dependencies that define organizations. Through the ’Organizational Ecology’ (OE) literature, we infer that similarities among organizations’ dependencies coincide with the extent to which they are similar in form. Key in this definition of form is that it arises from a comparison of organizations, - the set of which is said to belong to an ‘organizational field’ - through which niches and competitive pressures may be detected. Given the presence of an ‘organizational field’ and a sufficiently large sample of observable members, the question remains as to which observable characteristics help us account for form. Language plays a pivotal role in OE, as the directly observable aspect in the theory from which most other concepts in the theory are derived. We learn that form can be defined by the activities and processes of organizations that are communicated by the fields’ audience (i.e. all stakeholders that have an interest in the field). Further, these activities instantiate dependencies (e.g., certain tasks require particular resources, linkages among certain actors, etc.). Within the communications about a field, we focus on those syntactical structures that reveal activities. In language, activities are syntactically represented by the combinations of verbs, subjects (sending activities), direct objects (undergoing activity), and indirect objects (receiving activity). Each verb now defines a relation in the corpus, in which the nominal subjects, direct objects, and indirect objects are nodes, connected, when they occur in the same sentence with the verb. Note that these connections are directional. Nominal subjects send ties to objects, and direct objects to indirect objects. Aggregating over verb relations (or relations within verb-classes) establishes what we will call ’activity networks’. We do this for each company in the organizational field that we observe. Subsequently, we compare the activity networks of each pair of organization on a number of measures (correlation, eigenvectors of Laplacian matrices, Jaccard distance). An inter-organizational similarity network now emerges that allows distinguishing niches of organizations and the competitive pressures to which they are subject. For this study, we use a corpus of electronic textual communications of a field’s audience as data source. We observe 61 companies in this field. For each company in the field, we obtained between 20 to 30 documents, noting distinct actor sources: the company itself and its broader audience. We assess the niche structure with multi-level permutation tests on ’activity network’ based similarity between all pairs of companies in the field. We present three versions of this test. First, one that generates random values of comparison statistics by permuting one of the ’activity networks’ in each pair of companies. Second, we permute company labels in the higher level, socio-semantic similarity network to randomize the comparison values, and, third the combination of these methods. We discuss the assumptions underlying each of these approaches, which allow us to detect competitive pressures within the fields niches.
|Period||17 Jul 2020|
|Degree of Recognition||International|