Fig. 11From: The role of luck in the success of social media influencersThis is an alternative version of Fig. 7 in which \(95\%\) confidence intervals are displayed. The plots show how different recommendation processes influence the chance of observing fair outcomes for different CCs. The x-axis is for CCs (from highest-quality, \(CC_1\), to lowest-quality, \(CC_{100}\)) while the y-axis is for the percentage of simulations which were \(CC_i\)-fair after 1000 timesteps. a focuses on the effect of increasing the visibility of low-popularity CCs when starting from the realistic recommendation process (i.e., PA with \(\alpha = 1\)). b and c look at the effects of further increasing or respectively decreasing the visibility of the most popular CCsBack to article page