Two factors that might affect the novelty value of a test stimulus (the frequency of appearance of features common to the test stimulus and the set of preceding stimuli, and the serial position of these features) were systematically manipulated, and their effects on the electrodermal component of the orienting response (OR) were examined. We presented 256 participants with both verbal and pictorial stimulus sequences. Following 12 presentations of control stimuli, a test stimulus, which shared two common components with some of the control stimuli, was presented, followed by two additional presentations of control stimuli. The results revealed that recent presentations of the common components significantly reduced OR magnitude to the test stimulus, whereas the presentation frequency of common components had no significant effect. The implications of these findings for the feature–matching theory are discussed and a modification of the theory is proposed.