A multinomial modeling approach to dissociate different components of the truth effect

The subjective impression that statements are true increases when statements are presented repeatedly. There are two sources for this truth effect: An increase in validity based on recollection (a controlled process) and increase in processing fluency due to repeated exposure (an automatic process). Using multinomial processing trees (MPT), we present a comprehensive model of the truth effect. Furthermore, we show that whilst the increase in processing fluency is indeed automatic, the interpretation and use of that experience is not. Experiment 1 demonstrates the standard use of the fluency experience and Experiment 2 demonstrates that people can change the interpretation of the experience according to its ecological validity. By implication, the truth effect represents the adaptive usage of feedback received from internal processes.