Suppose I have two confidence intervals that measure accuracy of two different classifiers on the same data. In the first, I have $0.9087 \pm 0.0073$ for one, and $0.9123 \pm 0.0075$ for the other. These intervals overlap by more than their margin of error.
Is there a nice measure or graphical representation that can elaborate on how these are different without resorting to barplots? For example, is it fair to compute the percentage of overlap between these two intervals?
