Fig 1.
Trust interaction with the escrow system.
Fig 2.
Plot descriptive analysis.
Table 1.
Descriptive statistics.
Table 2.
GAM Models M1 and M2.
Fig 3.
Results of generalised additive models for the estimated impact of escrow (a) and percentage of positive reviews (b) on fraud.
Note: In this Fig (a) represents the estimated impact of escrow on fraud, while (b) represents the estimated impact of the percentage of positive reviews on fraud, as analysed through GAMs. The plot displays the predicted values and their corresponding confidence intervals, offering insights into the relationship between escrow, percentage of positive reviews and fraud cases.
Fig 4.
Results of generalised additive models for the estimated impact of (a) length of text of item offers, (b) linguistic diversity of text of item offers, and (c) sentiment score of text of item offers on fraud cases.
Note: In this Fig 4, (a) represents the estimated impact of the length of text of item offers, (b) represents the estimated impact of the linguistic diversity of the text of item offers on fraud cases, while (c) represents the estimated impact of sentiment score of the text of item offers on fraud cases as analysed through GAMs.
Fig 5.
Results of generalized additive models for the estimated impact of day selling on fraud cases.
Note: The plot illustrates the nonlinear relationship between the number of days the product has been on the market and frauds.