Figure 1
Graphical illustration of the example application, at the start of the first, second, fourth, eighth and sixteenth trial. Left panels show the 30 ‘hypotheses’ (i.e., possible values of µ, given the model in Equation 2). The darkness of the curves is proportional to the relative likelihood of the estimate being true. Right panels show the posterior density function, with a blue vertical line denoting the current estimate (here based on the mean value). The stimuli, x, and observer’s responses, r, aren’t shown. However, it can be inferred that in the first few trials the observer successfully discriminated a number of mid-magnitude stimuli, making high values of µ (poor sensitivity) unlikely.
