Do uncontrolled processes contribute to evaluative learning? Insights from a new two-US process dissociation procedure and ambivalence measures.

The contribution of uncontrolled processes to evaluative learning has been examined in evaluative conditioning procedures by comparing evaluations of conditioned stimuli between tasks or within tasks but between learning instruction conditions. In the present research, we introduced a new procedure that keeps both tasks and instructions constant. In addition, we introduced ambivalence measures to address this uncontrollability question. The new procedure involves forming an impression of conditioned stimuli based on their pairing with one unconditioned stimulus while attending but discarding the influence of another unconditioned stimulus holding the same (congruent trials) versus a different (incongruent trials) valence. When the to-be-used and to-be-discarded unconditioned stimuli share the same (vs. a different) valence, controlled and uncontrolled processes should support the same (vs. opposite) responses. We used this approach in two preregistered experiments (Ntotal = 467) using dichotomous evaluative classifications (Experiments 1 and 2), evaluative ratings, and two measures of attitudinal ambivalence: mouse trajectories and felt ambivalence (Experiment 2). While we failed to find evidence for uncontrolled processes in evaluative classification frequencies separately in Experiments 1 and 2, analyses of aggregated classification frequencies across Experiments 1 and 2 suggested a small contribution of uncontrolled processes. In addition, we found larger felt ambivalence for incongruent than congruent trials. Overall, the present findings are mixed but support the possibility of a contribution of uncontrolled processes to evaluative learning, even when control is applied to a focal stimulus and additional influences come from a to-be-disregarded stimulus.

Is Earth a perfect square? Repetition increases the perceived truth of highly implausible statements.

A single exposure to statements is typically enough to increase their perceived truth. This Truth-by-Repetition (TBR) effect has long been assumed to occur only with statements whose truth value is unknown to participants. Contrary to this hypothesis, recent research has found that statements contradicting participants' prior knowledge (as established from a first sample of participants) show a TBR effect following their repetition (in a second, independent sample of participants). As for now, however, attempts at finding a TBR effect for blatantly false (i.e., highly implausible) statements have failed. Here, we reasoned that highly implausible statements such as Elephants run faster than cheetahs may show repetition effects, provided a sensitive truth measure is used and statements are repeated more than just once. In a preregistered experiment, participants judged on a 100-point scale the truth of highly implausible statements that were either new to them or had been presented five times before judgment. We observed an effect of repetition: repeated statements were judged more true than new ones, although all judgments were judged below the scale midpoint. Exploratory analyses additionally show that about half the participants showed no or even a reversed effect of repetition. The results provide the first empirical evidence that repetition can increase perceived truth even for highly implausible statements, although not equally so for all participants and not to the point of making the statements look true.