We can estimate the distribution of subject’s responses by multiplying the noise distribution with the prior