Figural Effects And Belief Bias In Category Syllogistic Reasoning: Evidence From Eye-Movement

Figure effects refers to the phenomenon that different figures will influence scores and process time when people are reasoning and make people drew preferred conclusions. Belief bias refers to the phenomenon that people tend to accept or draw believable conclusions which coincide with their faiths when they are reasoning. Belief bias will also effect reasoning score and process time. A lot of researches have been on this two factors at home and abroad. Researchers have put forward some theories to explain, some of most influential theories are standard mental model theory and the variant. But throughout the researches at home and abroad, researchers studied figure bias almost by use of generated paradigm, studied belief bias almost b use of the evaluation paradigm, seldom simultaneously in two paradigms and rarely studied under two kinds of factors combined with research. On methods, there is no research on testing processing time using eye movement technology. Based on the previous studies, this study aims to use the real-time eye movement technique to study figure effect and belief bias in two paradigms and to further explore the essences of effect figure and belief to examine the different theories.The first experiment is a 2(figure:figure AB-BC, figure BA-CB) x 2(conclusion believability:believable conclusion, unbelievable conclusion) x2 (conclusion validity: valid conclusions, invalid conclusion)×2 (conclusion partial temper:A-C, C-A) four-factor mixed experimental design.Figure is between variables designed, other factors are internal variables. The dependent variable are test scores and eye movement index (duration on premise, duration on conclusion, first duration on premise, regression times on premise), This experiment aims to examine the effects of figure faith, validity, and preference on reasoning result and eye movement model in evaluation paradigm The second experiment is a 2(figure:figure AB-BC, figure BA-CB)×2(belief: believable conclusion, unbelievable conclusion) x 2 conclusions partial preference (A-C, C-A) three-factor mixed experimental design. Figure is between variables, other factors are internal variables. The dependent variables are scores and duration on premises. This experiment aims at examining the effects of figure, beliefs, conclusion partial preference on reasoning result influence and eye movement model in generating paradigm. Two experiments use Eye-Link CL and eye tracking system which are made in SR RESEARCH company.Two experiments are controlled by the Experiment-Builder software. Research shows that:(1) In evaluation paradigm, on reasoning scores, validity has significant effect; Partial preference has significant effect; the interaction between figure and faith is significant; the interaction between validity and believability is significant. On duration on premise, the interactions among partial preference, validity, partial preference are significant. On duration on conclusion, the interaction between validity and partial preference is significant, the interaction between validity and partial preference is significant, the interaction between partial preference and believability is significant. (2) In evaluation paradigm, on difficulty on early processing, believability has significant main effect; the interaction between figure and validity is significant; the interaction between validity and partial preference is significant. In late processing, validity has significant main effect; the interaction between figure and partial preference is significant; the interaction between believability and validity is significant;other effects are not significant. (3) In generating paradigm, on reasoning results, the interaction between figure and partial preference interaction is significant; believability has significant main effect; Other effects are not significant. On duration on premise, partial preference has significant main effects; figure has significant main effects markedly; Other effects are not significant. (4)The Mental Model theory cannot explain some results in this study, but compared with the standard Mental Model theory and Selectivity Processing Model experiment, more results support Meta-cognitive Uncertainty Model. On the mode of reasoning processing, research results advocate leading backward reasoning in evaluation paradigm. In generating paradigm forward reasoning exists in generating, but backward reasoning that people use the faith in their brains to guide to premise processing.