This finding on its own is consistent with the conformity hypothesis: reviews in aggregate are deemed more helpful when they are close to the product average. However, a closer look at the data raises complications, as we now see. First, to assess the brilliantbut-cruel hypothesis, it is natural to look not at the absolute difference between a review’s star rating and its product average, but at the signed difference, which is positive or negative depending on whether the star rating is above or below the average. Here we find something a bit surprising (Figure 2). Not only does the median helpfulness as a function of signed difference fall away on both sides of 0; it does so asymmetrically: slightly negative reviews are punished more strongly, with respect to helpfulness evaluation, than slightly positive reviews. In addition to being at odds with the brilliant-but-cruel hypothesis for Amazon reviews, this observation poses problems for the conformity hypothesis in its pure form. It is not simply that closeness to the average is rewarded; among reviews that are slightly away from the mean, there is a bias toward overly positive ones