Quantitative analysis of interviews in cooperation contexts: a stylometric profiling of relevant psychological processes

This contribution is rooted within the area of sustainable development, es- pecially for what concerns social development and reduction of inequalities, and within the capacity building approach, over the premise that development can be attained only when a common goal is pursued by relevant stakeholders that have the capabilities to work on it. Within this field, a key communicative com- ponent is the capability to involve multiple stakeholders and to define common local goals that can be pursued. Therefore, the focus of this contribution is to identify linguistic statistical markers functioning as indicators of psychosocial processes that come into play in the definition of such common goals. This research relies on Construal Level Theory – a psychosocial theory that concep- tualizes the level of abstraction of an object in terms of perceived distance from the self - to categorize several conceptions of social development as high-level or low-level construal, following the assumption that the closer an “object” is in term of social or psychological distance (low level), the more concretely it is mentally construed; the further away it is (high level), the more abstractly it is construed. Therefore, having the possibility to guide the identification of personal ideas of development goals as low-level vs high-level construal using quantitative measures and statistical markers could be a key tool within the cooperation contexts, where the possibility of having several visions needs to be considered - possibly combining them. The research aim is to identify specific stylometric features associated with each construal level within text produced by people working in cooperation contexts, using statistical measures like the Part-of-Speech frequencies, word length, and sentence length. The research corpus has been composed of 20 semi-structured interviews with employees of organizations working in social development and local businesses, using ques- tions aimed at generating high-level vs low-construal level material on topics such as goals in development, ideal and actual strategies on how they pursue these goals as citizens vs as member of these organization, their perception of criticalities. The interviews have been transcribed and the resulting text has been divided into two categories, high vs low construal levels, according to the sections of the interviews. All text has then been processed with Python to identify the following measures: Part-of-Speech frequencies, word length and sentence length. These measures underwent a MANOVA to analyze the effect of each stylistic specificity between the two construal levels. The results will focus on the presentation of the previously indicated measures and the outcome of the MANOVA, expecting a relevance of POS tags adjectives within the high construal text, and nouns and verb tags within the low construal text. This could act as a first preliminary step in the construction of a text-based cate- gorization system of these processes from a linguistic perspective, as well as an enrichment of the relevant psychological theory.