| In daily life as well as in the psychological laboratory, people continuously make decisions. These decisions pertain to widely different activities, such as buying new sun-glasses, driving your car to work, or writing grant proposals. All of these decisions, however, fall prey to the same dilemma. This dilemma concerns the meta-decision of when to stop information processing and commit to a decision. This is particularly evident in tasks where one can choose to respond faster at the cost of making more errors. Clearly then, task performance is a function of both response accuracy and response speed. A pervasive problem in cognitive psychology is how to combine speed and accuracy so as to obtain separate indices for task performance and response conservativeness. Perhaps the only way to make progress is to use a mathematical model that explicitly addresses the tradeoff between speed and accuracy. The current proposal focuses on Ratcliff s diffusion model, which is arguably the most popular model of how people process information. The diffusion model allows one to estimate unobserved psychological processes such as perception, speed of information accumulation, response conservativeness, and response bias. The proposed projects seek to theoretically extend and empirically test the diffusion model account of the speed-accuracy tradeoff. This account currently leaves open several important questions. The first project shows that the Fuzzy Logical Model of Perception (FLMP) can be unified with the diffusion model in a way that allows the FLMP to simultaneously account for response speed and response accuracy. The second project studies what happens under conditions in which there is almost no value in accurate responding. The third project considers variability in response conservativeness as an explanation for fast errors, and the fourth project concerns the changes in information processing that occur after an error. |
