Serial dependence in the estimation of a car's arrival time

Serial dependence of a car's arrival time. We have studied the sequential structure of data in the arrival- time estimations. Forty participants estimated the arrival-time of a vehicle under two experimental conditions: real car and video-image. Various time series regression models were fit to our data, and residuals autocorrelations were computed. For each serial-dependence model, data were fit to three functions, namely, power, logarithmic, and linear. In both experimental conditions, the response magnitude (R) on a given trial in t was a function of the stimulus intensity (S) in such a trial (t) and of the S and R on t-1 and t-2. Assimilation effect to the previous responses and contrast effect to the previous stimuli has been found.