Flipbook

6 APPROACH BEHAVIOR AS INFORMATION 151 your shopping bag, the kind of spider you have seen on television. You think: A tropical spider!”; or (2) objective safety information, e.g., “At home you see a big house spider in your shopping bag”. The objective information was followed by (3) approach behavior, e.g., “You grab a jar to catch the spider”; or (4) no approach behavior, in which case nothing followed the objective information. Participants were asked to evaluate the events as if they were happening to them at this moment and to identify themselves with the description as much as possible. Each scenario was presented on a new page in a randomized order, with the restriction that scenarios about the same situation were never presented in succession. A validation study prior to the experiment showed that the approach manipulation was unsuccessful in the public speaking scenarios. They were therefore used as filler scenarios in the experiment and not included in the analyses. 2 Data analysis Our hypotheses consisted of specific expectations that could be formulated in terms of equality and inequality constraints (i.e., informative hypotheses; Mulder et al., 2009), and were therefore evaluated using Bayesian model selection (see Klugkist, van Wesel, & Bullens, 2011). Results of Bayesian model selection are expressed in term of Bayes factors (BF; Kass & Raftery, 1995), which represent to what extent the data support the hypothesis (i.e., the constrained model) compared to the unconstrained model (Mulder, Hoijtink, & de Leeuw, 2012). BF < 1 indicates that there is no support for the hypothesis. BF > 1 means that the hypothesis is supported by the data, with a higher BF indicating more support. BFs of competing hypotheses can be compared to obtain the relative support for each hypothesis. The data were analyzed with BIEMS (see Mulder et al., 2012) software. We expected that scenarios with objective danger information would on average (µ) receive higher danger ratings than scenarios with objective safety 2 Public speaking scenarios, general scenarios, and information about the validation study are available on request from the first author.