BackHome
Refresh: OffView Personal Schedule
Individual Submission Summary
Share...

Direct link:

Using Latent Class Analysis to Develop Typologies of Online Sellers of Potentially Illegal Animal Leather Products

Thu, September 4, 5:30 to 6:45pm, Communications Building (CN), CN 3103

Abstract

After COVID-19, there has been a major shift in illegal wildlife trade (IWT) from physical stores to online marketplaces, leading to a variety of challenges for e-commerce regulators and law enforcement. While research in this realm has identified typologies of sellers involved in this activity, there is a significant gap in the literature related to online seller characteristics, which is what this study aims to fill. Using machine learning (ML) and large language models (LLMs), we extracted data on potential illegal sales of small leather products on Ebay as the case-study marketplace. We used latent class analysis (LCA) to identify offender typologies based on the combination of 15 different indicators related to the sellers and their online interactions, among which are the type, description, and price of the products sold, as well as frequency of sales, duration of the seller’s presence on Ebay, and their cross-platform engagement. Analyses revealed four distinct types of offenders, including “Storefront and Visibility” sellers who have strong storefront presence and branding, but less exclusivity; “Small Independent Sellers” who have low storefront presence, but sell a diverse number of products derived from wildlife; “Luxury and eBay Sellers” who sell luxury items at lower exclusivity and are likely to sell products made of diverse species; and, lastly, “Exclusive and Specialized Sellers” who cater to a niche group of buyers looking for highly exclusive luxury products. The proposed framework can inform targeted and effective responses to IWT carried out on online marketplaces.

Authors