Figure 6
From: Misinformation does not reduce trust in accurate search results, but warning banners may backfire
Higher rank (X-axis) is associated with an increase in an accurate result's click likelihood (Y-axis). This relationship (blue line) is stable across experiments (columns). The blue lines are fitted by the logistic regression model and show the predicted probabilities of each rank being clicked, while the black points show the observed probabilities. For example, the left-most point in the left-hand panel (“Experiment 1”) shows that an accurate result appearing at the highest rank in this experiment had a 22% likelihood to be clicked. Each datapoint represents probabilities generated from 598 to 3196 individuals' click decisions, depending on the experiment and rank displayed (since not all participants were exposed to accurate information in the first three ranks). Error bars represent 95% Confidence Intervals of the sample proportions.
