SearcharxivSearch

arXiv subjects

Poompak Kusawat

Publications and source records attributed to Poompak Kusawat.

5 recordsLinked to original sources

Opportunities and Challenges of Integrating ChatGPT in Education: Sentiment Analysis and Topic Modeling

Since its recent debut, ChatGPT has become a global sensation and significantly impacted the field of education. Both educational researchers and practitioners have identified opportunities as well as risks associated with the use of this novel tool in educational settings. Despite the ongoing debate, there is still no research exploring occupational differences in the perception of ChatGPT in education. In this paper, we analyzed Twitter data using topic modeling and sentiment analysis to investigate how ChatGPT is perceived and discussed differently in different occupations. Our study found diverse topics discussed including its use in schools, impact on exams, academic integrity concerns, and response accuracy evaluations. While most tweets were positive or neutral, concerns about integrity and response accuracy were evident. Analysis revealed sentiment and topic variations among users' occupations. These findings emphasize the opportunities and challenges of integrating ChatGPT in education, necessitating continued monitoring and informed policy-making for responsible utilization.

cs.HC

The Roles of Culture in Online User Reviews: An Empirical Investigation

Electronic word-of-mouth (eWOM) is a prominent source of information that significantly influences consumer purchase decisions. Recent literature has extensively explored the impact of eWOM on consumers-generated reviews and purchase decisions. However, few studies have analyzed the role of culture on eWOM. We use a novel dataset of Airbnb eWOM messages in order to empirically extend the findings by Banerjee and Chai (2019). We find that the sentiment of individualistic customers is worse than that of their collectivistic counterparts when both groups experience the same level of negative disconfirmations. Furthermore, guests from a relatively more distant culture rely less on heuristics. In particular, quality signals, such as the "superhost" status, are more influential to consumers from a less distant cultural background.

cs.HC

Cross-cultural electronic word-of-mouth: a systematic literature review

Purpose: Global adoption of the internet and mobile usage results in a huge variation in the cultural backgrounds of consumers who generate and consume electronic word-of-mouth (eWOM). Unsurprisingly, a research trend on cross-cultural eWOM has emerged. However, there has not been an attempt to synthesize this research topic. This paper aims to bridge this gap. Methodology: This research paper conducts a systematic literature review of the current research findings on cross-cultural eWOM. Journal articles published from 2006 to 2021 are included. This study then presents the key issues in the extant literature and suggests potential future research. Findings: The findings show that there has been an upward trend in the number of publications on cross-cultural eWOM since the early 2010s, with a relatively steeper increase toward 2020. The findings also synthesize cross-cultural eWOM research into four elements and suggest potential future research avenues. Value: To the best of the authors' knowledge, there is currently no exhaustive/integrated review of cross-cultural eWOM research. This research fills the need to summarize the current state of cross-cultural eWOM literature and identifies research questions to be addressed in the future.

cs.HC

Sponsorship Disclosure in Native Advertising: A Theoretical Framework

Native advertising is one of the fastest growing areas of online promotion. After reviewing extant literature via EBSCOhost database, this study draws on Persuasion Knowledge Model and develops a theoretical framework which facilitates a clearer understanding of the relationship between sponsorship disclosure in native advertising and consumer outcome. The framework suggests that sponsorship disclosure has a negative effect on electronic word of mouth (eWOM), and further proposes the interplay between the main effect with brand prominence and the type of device. This is highly relevant to marketer as regulators have been pressuring for the disclosure of native advertising. As this is likely to have detrimental effect to the eWOM, marketer may employ the boundary conditions proposed by this framework to attenuate that negative effect.

cs.SI

Impact of Investing Characteristics on Financial Performance of Individual Investors: An Exploratory Study

This exploratory study examines which investing characteristics determine success in an equity market. Based on data from 403 respondents, exploratory factor analysis results in 13 factors: middle/long time horizon, qualitative analyst, open-minded/disciplined, organized, emotional stability, naïve, growth stock, concentrated portfolio, contrarian, value stock, globalized, intrinsic value, and price-independent. Multiple linear regression of individual investors' excess return on these factors show statistically significant relationship. These results deepen our knowledge on what sort of investing characteristics are required to survive in equity markets.

q-fin.GN