Presentation of stimuli without reg

Presentation of stimuli without regard to the positioning of responses relative to the anchors is likely to distort the observed data from the straight-line psychophysical function by, for example, ceiling or floor (end) effects, centering bias at the start and range–frequency bias throughout (Booth et al., 1983; Conner et al., 1986, 1988a, 1988b). It is therefore important not to present any sample that is likely to give a response close to or beyond “never choose”. Such an intolerable sample can be identified with increasing accuracy by extrapolation of each emerging function from the second sample onward