engleski

The effect of believability and validity of categorical syllogisms on metacognitive assessments

There are two types of processes involved in thinking Type 1 and Type 2 (also called System 1 and System 2 thinking). Type 1 process produces rapid, intuitive reactions and instantaneous decisions and is based on heuristics. Type 2 is slower and deliberative type of thinking, based in mental skills such as arithmetic or logic rules. Every Type 2 process requires mental effort and it interferes with other Type 2 processes when performed in the same time. Research of metacognitive assessments in reasoning tasks shows that there is an occasional discrepancy between the objective accuracy and the evaluated accuracy. The often proposed interpretation for this is that participants lack insight into Type 1 thought processes. Due to this, participants rely on indirect signs such as e.g. fluency when concluding. The aim was to estimate how metacognition depends on the believability and the validity of the syllogism, as well as on the accuracy and the fluency of the participants’ response. Believability was defined by whether premise matches common sense, and validity was defined by logical rules. The metacognitive assessment was operationalized as judgement of confidence for given answer. Thirty participants were included in this experiment. The experimented had 2x2 design, with two levels of believability and conclusion validity. Syllogistic conclusions appeared on a screen and participant’s task was to verify the conclusion. After this, participants had to indicate level of confidence in his/her answer on a 1-5 scale (from random guessing to complete confidence). The results showed the validity of a conclusion has a significant effect on both accuracy (F(1, 29) = 14.344 ; p < 0.01) and on response time (F(1, 29) = 7.85 ; p < 0.01). Valid conclusions were verified more quickly and more accurately. Participants showed higher level of confidence for responses to valid conclusions (F(1, 29) = 45.28 ; p < 0.01). The influence of the believability was demonstrated only for accuracy, and not for the response time (F(1, 29) = 7.31, p < 0.05). The correlation between the response times and confidence judgments were significant only in certain situations such as in believable non-valid syllogisms (rs = -0.37 ; p < 0.05) which in part confirms the starting hypothesis.

metacognition ; confidence judgments ; deductive reasoning ; categorical syllogisms

nije evidentirano