Are there any psychophysical applications of single-observation confidence intervals?