[EN]People believe others are telling the truth more often than they actually are, called the
truth bias. Surprisingly, when a speaker is judged at multiple points across their statement
the truth bias declines. Previous claims argue this is evidence of a shift from (biased)
heuristic processing to (reasoned) analytical processing. In four experiments we contrast the
heuristic-analytic model (HAM) with alternative accounts. In Experiment 1, the decrease in
truth responding was not the result of speakers appearing more deceptive, but was instead
attributable to the rater’s processing style. Yet contrary to HAMs, across three experiments
we found the decline in bias was not related to the amount of processing time available
(Experiment 1-3) or the communication channel (Experiment 2). In Experiment 4 we find
support for a new account: that the bias reflects whether raters perceive the statement to be
internally consistent.
