Large Language Model Text Analysis and Attribution Theory Explain the Return Paradox of E‑Commerce Consumers
DOI: https://doi.org/10.62381/E264724
Author(s)
Lin Liu
Affiliation(s)
Wuhan City Polytechnic, Wuhan, Hubei, China
Abstract
In the digital e‑commerce landscape, the return paradox—whereby high post‑purchase satisfaction coexists with high return rates—has become increasingly prominent. Traditional consumer behaviour theories fall short of fully explaining the underlying mechanisms of this contradictory phenomenon. Drawing on attribution theory and leveraging large language model (LLM)‑based text analysis to mine real user review data from e‑commerce platforms, this paper deconstructs consumers’ return attribution logic from two dimensions—internal personal perception and external contextual environment—and clarifies the formation mechanism of the satisfaction‑driven return paradox. The findings reveal that consumer satisfaction is not the sole core factor inhibiting return behaviour; rather, non‑negative evaluative factors such as implicit cognitive biases and contextual fit gaps are key triggers for returns following satisfaction. This study addresses the limitations of traditional questionnaire‑based research, which suffers from fragmented data and ambiguous attribution, enriches the theoretical framework of consumer return behaviour in the digital era, and provides empirical evidence and practical references for e‑commerce platforms and merchants to optimise operational strategies and reduce irrational return costs.
Keywords
Large Language Model Text Analysis; Attribution Theory; E‑Commerce Consumers; Return Paradox; Consumer Behaviour
References
[1] Cheng Fan. A study on the influence of negative emotions on residents’ consumption behaviour under sudden events. Commercial Exhibition Economy, 2022, (6): 79‑81. DOI:10.19995/j.cnki.CN10‑1617/F7.2022.06.079.
[2] Xiao Qiang, Zhang Yuanyuan, Shi Wei. Research on cross‑channel return strategies under consumer impulsive consumption. Journal of Lanzhou Jiaotong University, 2025, 44(4): 58‑67.
[3] Wu Yang. Research on influencing factors of consumers’ return intention in live‑streaming context. Master’s thesis, Jiangxi University of Finance and Economics, 2024. DOI:10.27175/d.cnki.gjxcu.2024.000779.
[4] Li Jie. A review of influencing factors of consumers’ impulsive buying behaviour in live‑streaming context. Northern Economy and Trade, 2022, (9): 49‑53.
[5] Wan Fang. Legal application of the business operator’s duty of disclosure under China’s Consumer Rights Protection Law. Politics and Law, 2017, (5): 151‑160. DOI:10.15984/j.cnki.1005‑9512.2017.05.013.
[6] Zheng Ziqiang. A study on the improvement of the right of withdrawal system for online consumers. Master’s thesis, Tiangong University, 2025. DOI:10.27357/d.cnki.gtgyu.2025.000725.
[7] Xu Yinbo. Regulation on the validity of induced consumption behaviour. Politics and Law, 2026, (1): 163‑176. DOI:10.15984/j.cnki.1005‑9512.2026.01.001.
[8] Huang Shijing, Xu Man. A study on the influence mechanism of mobile e‑commerce live streaming on users’ online purchase intention based on SOR theory. Statistics and Management, 2021, 36(7): 122‑128. DOI:10.16722/j.issn.1674‑537x.2021.07.022.
[9] Liu Qingqing. The impact of product reviews on consumer behaviour on cross‑border e‑commerce platforms: A case study of Amazon. Master’s thesis, Zhejiang University, 2023. DOI:10.27461/d.cnki.gzjdx.2023.001086.
[10] Mumuni Alhassan G.; O’Reilly Kelley; MacMillan Amy; Cowley Scott; Kelley Brett. Online product review impact: The relative effects of review credibility and review relevance. Journal of Internet Commerce, 2020.