Choice Reaction Time and Conjunctions of Form Size Color and Orientation

Reaction time (RT) is frequently used as a dependent measure in studies that examine the organization of perceptual and cognitive processes. Hick's law states that RT in a choice-reaction-time (CRT) task is a linear function of the amount of information transmitted per stimulus-response event. RT has been used to try to determine whether serial or parallel processing is being employed in CRT tasks (Sternberg, 1966). There is research evidence suggesting a serial, exhaustive search process (Sternberg, 1969a; Wingfield, 1973), as well as findings which suggest parallel processing (Neisser, 1963; Neisser, Novick, and Lazar, 1963) in CRT tasks. All the experiments used to study CRT and serial and parallel processing required subjects to make decisions based on a union or "or" rule. Although Hick's law is based on union searches, most reality-based decisions are made using an intersection or "and" rule. The present study investigates intersection decisions involving form, size, color, and orientation, and their relation to RT. The study used a 2 (union or intersection) x 2(positive or negative trial) x 2 (male or female) x 3 (2, 3, or 4 dimensions) x 5 (blocks of trials) mixed model factorial design with subjects nested in decision type, number of dimensions, and gender. Ninety-six male and female subjects were randomly assigned to one of six conditions and viewed stimuli on a computer screen consisting of large or small shapes (triangle, square, diamond) of different colors (red, blue, green) with black arrows (pointing left, right, up, or down) in them. Members of the positive set were displayed at the bottom left of the screen throughout the entire experiment. Subjects completed 240 stimulus trials. In the union condition, subjects responded positively if any member of the positive set was present, and negatively if no member of the positive set was present. In the intersection condition, subjects responded positively if all members of the positive set were present, and negatively if no members of the positive set were present. A 2 x 2 x 2 x 3 x 5 ANOVA was used to analyze the RT data. A significant main effect was found for Type of Trial g F( 1 g 84 ) 15 87 g such that Negative trials had a significantly higher RT than Positive trials. There was a significant main effect for Number of Dimensions, F(2,84) = 9.90. A SNK post hoc test showed the RT for the Two-dimension group to be significantly lower than the RT for the Three- and Four-dimension groups. A significant Block effect, F(4,336) = 111.89, revealed a learning effect occurring in Blocks 1 and 2. The significant Type of Decision x Type of Trial x Number of Dimensions interaction provides evidence for serial processing in the union condition and parallel processing in the intersection condition. Future studies should include higher dimension levels to explore further the exact relationship between number of dimensions and RT. Since most of our everyday choices involve intersection-decision processes, research in this area will provide invaluable information.