Intertrial interval and sequential effects in magnitude scaling

The intertrial interval (ITI) was varied within subjects in magnitude estimation and cross-modality matching experiments. Fits of a recently proposed time series regression model show that the influence of the previous stimulus intensity on the current response decreases when the ITI is increased. The results can be interpreted as showing that an assimilative or additive perceptual or memory effect decreases with an increase in ITI. Fits of an earlier model, on the other hand, suggest that the influence of the previous stimulus intensity increases with an increase in ITI, which is counter to expectations. The new regression model (a) provides a simple explanation for the counterintuitive results obtained with the earlier model, (b) shows that assimilation in perception or memory can appear as contrast, and (c) reduces to a simpler model for longer ITIs.