Meta-analysis on conditional reasoning
- by Sunny Khemlani
- in News
- posted July 22, 2019
Marco Ragni, Hannah Dames, and Phil Johnson-Laird recently published a new meta-analysis on conditional reasoning at the International Conference on Cognitive Modeling. Here’s the abstract:
Conditional premises are assertions with “if”, e.g., If I have measles, then I have fever. They provide a connection between different propositions and can express causal relations. Conditional inferences often comprise conditional and categorical assertions, e.g., such as modus tollens: If I have measles, then I have fever; I don’t have fever; So, I don’ t have measles. Most research has concerned four sorts of conditional inference, examining them separately. Only a few studies have focused on the patterns over the four sorts of inference (e.g., Oberauer, 2006). Our meta-analysis was of 39 experiments (with 2378 participants) that reported these patterns. It showed that a version of the mental model theory best fits the results when participants produced their own conclusions or evaluated a given conclusion, whereas the suppositional theory provided the best fit when participants chose a conclusion from a list of options.