Perceptual processing and speed-accuracy trade-off.

Three experiments are described on speed-accuracy trade-off during a visual fixation. In a typical trial subjects started by fixating a fixed position at their left side at which a degraded digit (SL) was presented. At about the same time an intact digit (SR) was briefly presented at a fixed right side position. SL and SR were separated by a binocular visual angle of 45 degrees. The task was to verbally report both digits, which implied that the eyes should be rapidly shifted from SL to SR, so as not to miss SR. Under these conditions the distribution of the fixation latencies of SL appeared to consist of two parts, suggesting the occurrence of either anticipations or reactions. This argues against a strategy of optional stopping of information accrual during the fixation of SL and is in line with a strategy of either fully neglecting or fully encoding SL. However, this interpretation meets the difficulty that, at trials where SL is reacted to, subjects showed a shorter fixation latency than in a control condition, in which they had ample time to view both SL and SR. The results of the second and third study supported the view that, when under time stress, subjects can end a fixation as soon as a preliminary code of SL is obtained. This code (a) is insensitive to the complexity of SL, (b) cannot be elaborated during the saccade to SR, and (c) can be used as a starting point for further analysis, as soon as the eyes reach SR.