Abstract：There are lots of evidences show that participant’s performance on bayesian inference、syllogistic reasoning and probability reasoning could been promoted by cumulative frequency tree, but very few study focus on the promotion effect of frequency tree on causal reasoning. The study conducted four experiments to explore the impact of frequency tree on causal strength inference, research hypothesis include (a)The boundary of frequency facilitating lies in the supply of focus set or not: frequency tree which supplies focus set can improve rationality of participant reasoning, frequency tree which conceals focus set can’t improve rationality of reasoning; (b)participant estimate causal strength of different contingency by different mode in experimental treatment that adopted frequency tree which conceal focus set, but by the same mode in experimental treatment that adopted frequency tree which supply focus set, and (c)counter-factual question bring more rational reasoning than ability question.
2(frequency tree, level 1: supply focus set, level 2: conceal focus set) x3 (contingency, level 1: Δp=0.33 and Power-PC=0.5; level 2: Δp=0.33 and Power-PC=0.83; level 3: Δp=0.67 and Power-PC=0.83) completely random design were used in all experiments. 231 undergraduate students participated in Experiment 1 which adopted counter-factual question and production contingency, 238 participated in Experiment 2 which adopted counter-factual question and prevention contingency, 238 participated in Experiment 3 which adopted ability question and production contingency, 225 participated in Experiment 4 which adopted ability question and prevention contingency. Contingency offered by a booklet which contain 30 page, each page present one sample of causal. Participant completed a frequency tree base on contingency, and made causal strength estimate individual.
Results show that (a)there are three common models of causal reasoning：Δp, Power-PC and P(e/c)(P(-e/c) for prevent contingency), the most popular model changes with the change of experiment treatment; (b)Most subjects(70.06% on average) estimate causal strength by Power-PC model when they reason by the aid of frequency tree that supplying focus set, and only a few participants(mean 21.28%) estimate causal strength by Power-PC model when they reason by the aid of frequency tree that concealing focus set; (c)Counter-factual question can lead to subjects focus on the information of focus set, but it can’t guarantee subjects to estimate causal strength by Power-PC model; (d)there were no causal direction effect on participant’s causal strength estimate. Experiment results fully support research hypothesis (a) and (b), and partial support research hypothesis (c).
The frequency facilitating depend on supply focus set or not, whether in counter-factual question treatment or in ability question treatment. It seem that participant apt to make rational causal inference when they reason by the aid of frequency tree that supply focus set and when they are questioned by counter-factual format.