Development, validation, and applications to marketing research of a comprehensive Italian dictionary for automated text.
Progetto From sponsored articles to print advertising, from tweets to Facebook posts, text is an ever-present feature of communication. Text is produced, in enottnous amount, by professionals within ovations, andby individuals who write reviews, posts, chats. Despite being mostly unstnictured, texts can be automascally coded and analyzed thanks to text mining and natural language processing tools (Berger et at., 2020). Such algoñthms allow researchers to gather and to analyze textual data from the internet and social media, thus enabling the investigation of several research questions by social scientists. Italian is the fourth most studied language (1) and tefers to the seventh largest economy in the world, but a validated dictionary for automated analysis of Italian texts is not available. The LIWC software (Boyd et at, 2022), which is a standard for comprehensive dictionaries foe text analysis, proposes an Italian adaptation of LIWC 2007. This version, however, appears incomplete (it contains only 5156 wold-stems) and shows some pitfalls that limit its validity.
Given these premises, and following directions proposed by the text mining
literature (e.g., Humphreys & Wang, 2018), this project aims: (i) to develop and to validate a comprehensive Italian dictionary for automated text analysis, which we call ITA.LI.AN. (ITAlian Llnguistic ANalysis), and (ii) to use the dictionary in a set of rriarketing applications. The ITA.LI.AN. dictionary will be able to track mole than 90’ Zo of words in texts and to quansfy texts upon a wide range of linguistic (e.g., verb time-frame, pronouns) and psychological (e.g., emotions, COgnitioa, motivation, perception) constructs. The quandfied scores for such constructs can be used to analyze textual data in specific areas of interest for academic ieseatchexs (e.g., test of specific hypotheses), companies (e.g., sentiment analysis of their brands), and institutions (e.g., identifying the most effecsve commurticas on approaches). The project will involve two research units. The University of Calabcia unit will work on the development and validation of ITA•LI.AN. and on two applications: the analysis of the inteiacfion effect between content of messages and distinctiveness of texts on the evaluafion of sponsored articles in an Italian website; and the study of multi-modal COHMTliinirsfion in Italian creative industries (music, podcastinQ, by integrating the study of texts with that of images, videos, and sounds. The University of Chieti-Pescara unit will aim at applying ITA.LUN. to analyze tweets of established and new entrant Italian fashion companies’ related to environmental and social sustainability, investigating the effects of different
types of tweets (in teans of content and style) on consumer reactions.