Identification of manipulative features in social media advertisements

Abstract

In today's world, consumers' social interactions have largely shifted to social media, significantly boosting the use of marketing tools tailored for these platforms. Consequently, social media marketing (SMM) management as well as its primary tool—social media advertising, particularly targeted advertising— have become increasingly relevant. The aim of this study is to develop a mechanism for detecting signs of manipulation in advertising texts on social media platforms. The proposed interdisciplinary research has been carried out at the crossroads of SMM management, corpus linguistics and natural language processing. The dual nature of advertising as both a socio-economic and linguistic phenomenon creates its fundamental contradiction: it is originally intended to provide objective information about the consumer value of a product, yet simultaneously serves as a tool for manipulative influence aimed at encouraging consumers to make a purchase. Targeted advertising is particularly manipulative, as it is shown to a specific audience selected based on socio-demographic characteristics. This approach does not ensure equal access to information about products for all consumers, thereby violating one of the core principles of a market economy. Advertising texts often make extensive use of linguistic and stylistic tools for verbal manipulation, such as comparisons, epithets, euphemisms, and metaphors, each containing an emotional-evaluative component. Evaluative vocabulary is an integral part of the manipulative strategy of advertising discourse; as such, predicates persuade consumers and reinforce the idea of a product’s superiority over its competitors. To test the hypothesis that evaluative adjectives dominate over those describing objective characteristics of products in advertising messages, a corpus of 120 advertisements from Facebook and Instagram was compiled. The results showed that 86% of the adjectives were evaluative (e.g., perfect, incredible, stylish, beneficial), while only 14% described objective features (e.g., ceramic, winter, portable). This confirms the hypothesis about the dominance of evaluative vocabulary, which manipulates consumer perception, creates positive associations, and stimulates purchasing behavior.