Analyzing Chatbot-Mediated Communication Effects on Language Use and User Engagement

Thesis Abstract

This study investigates the impact of chatbot-mediated communication on language use and user engagement within digital interaction contexts, addressing the growing prominence of artificial intelligence (AI) chatbots as primary interactive agents in customer service, education, and social networking. The proliferation of chatbots has raised concerns regarding their influence on users’ language behavior and engagement levels, yet comprehensive empirical evidence remains limited. The primary aim of this research is to analyze how interaction with chatbots influences linguistic patterns and user engagement metrics, with specific objectives to (1) identify changes in language complexity, formality, and lexical diversity in chatbot interactions; (2) assess variations in user engagement measures such as interaction duration, frequency, and satisfaction levels; and (3) explore the theoretical relationships between chatbot conversational features and linguistic as well as engagement outcomes. The study adopts a mixed-methods research design, integrating quantitative and qualitative approaches to provide a multidimensional analysis. The population under investigation comprises 300 university students from tertiary institutions actively engaging with customer service or educational chatbots. A stratified random sampling technique ensures representation across different demographic groups, with 150 participants selected for quantitative analysis through structured questionnaires and chat transcripts, and 150 participating in semi-structured interviews for qualitative insights. Data collection instruments include a standardized linguistic analysis tool that assesses language features in chatbot interactions, Likert-scale surveys measuring engagement and user satisfaction, and interview protocols exploring subjective experiences. Validity and reliability of the survey instruments are established through pilot testing with a subsample of 30 students, ensuring Cronbach’s alpha coefficients exceeding 0.85 for engagement scales. Chat transcript analysis employs computational linguistics techniques such as lexical diversity indices and syntactic complexity measures, while engagement variables are analyzed via descriptive statistics, multiple regression analysis, and ANOVA to identify statistically significant differences and relationships. Thematic analysis is employed for interview data, guided by relevant communication theories including the Media Richness Theory and the Computer-Mediated Communication (CMC) framework, to interpret user perceptions and contextual factors influencing language and engagement. Expected findings suggest that chatbot interaction can significantly alter users’ language use, notably increasing lexical diversity but potentially decreasing language formality over sustained interactions. Additionally, high levels of chatbot responsiveness and conversational fluidity are anticipated to correlate positively with user engagement, reflected in longer interaction durations and higher satisfaction ratings. Conversely, limitations such as repetitive language or technical glitches may diminish engagement and discourage complex language use. These findings are expected to contribute new empirical evidence to the fields of computational linguistics, human-computer interaction, and communication studies by delineating the influence of chatbot features on linguistic and engagement outcomes. The study’s principal contribution lies in providing a nuanced understanding of how artificial conversational agents shape language behaviors and user experiences, generating theoretical and practical insights for designing more effective chatbot systems that foster richer language use and sustained engagement. Based on the findings, the study recommends integrating adaptive linguistic scripts, enhancing emotional responsiveness, and incorporating user feedback mechanisms to optimize chatbot interactions. Future research directions include longitudinal studies to examine long-term language and engagement trajectories and experimental designs testing specific chatbot features hypothesized to influence linguistic and behavioral variables. Ultimately, this research aims to inform developers, communicators, and educators on constructing chatbots that promote positive linguistic and engagement outcomes in diverse digital contexts.