Previous investigators (Hick, 1952;Hyman, 1953;Crossman, 1956;Doherty, 1965) have found a linear relationship between choice reaction time (RT) and stimulus information when all stimulus alternatives are equally likely. However, when stimuli are not equally likely, the results are less clear-cut. Hyman (1953) and Stone and Calloway (1964) found that for low information stimuli (those with a high probability of occurrence) RT was greater than would be predicted using the regression line for equally likely alternatives, and conversely for high information (low probability) stimuli. Lamb and Kaufman (1965) and Kaufman and Levy (1966) both found results contradictory to the Hyman results; RT for low information stimuli was markedly less and RT for high information stimuli was greater than would be predicted from the regression line for equally likely alternatives, giving a generally quadratic function. Although a variety of different procedures were used in the above studies, only Doherty for equally likely alternatives and none for the unequally likely case have used the method of absolute judgments with unfamiliar stimuli. The present experiment was designed to determine the shape of the function relating RT to stimulus and transmitted information for unequally likely stimuli in an absolute [udgment situation and to explain previous discrepancies.
METHOD SubjectsThe Ss were 60 male and female undergraduate students from introductory psychology courses at the University of Connecticut. The 60 Ss were assigned randomly to six groups of 10 each.
ApparatusThe apparatus consisted of a Gerbrands tachistoscope, a voice key and a Hunter millisecond timer. A button held by S initiated the trials and actuated the timer; S's response activated the voice key, turning off the stimulus and stopping the timer.The stimuli were eight lines (Black on White) 1/8 in. wide and varying in length from 1 in. to 3.72 in. by successive ratios of 1.22 to 1. These stimuli have been previously shown (Doherty, 1965) to be relatively easy to discriminate. The responses used were bun, boo, bee, bore, bive, bix, bev, bate for stimuli 1 to 8.
Experimental conditions and procedureThere were six experimental conditions, three with equally likely alternatives (ELA) and three with unequally likely alternatives (ULA). The ELA conditions had 2, 4, and 8 alternatives; for 2 alternatives, stimuli 1 and 2 were used; for 4 alternatives, stimuli 1 to 4; and for 8 alternatives the entire set was used. In the ULA conditions, two stimuli were always used, the pairs being drawn randomly from the entire set with the restriction that they be consecutive members of the set, in order to keep discriminability approximately constant, The relative frequency ofthe members of each pair was varied, the values being 90/10, 75/25, and 60/40.Each group of 10 Ss was assigned to an experimental condition and served in one experimental session lasting approximately 1 hr. The E informed the Ss of the relative frequency of occurrence of the stimuli for all conditions. Pretraining...