engleski

Inferences with complex conditionals: The effect on accuracy and metacognitive assessments

The classic one-model Modus Ponens (MP) conclusion is more accurate, faster and easier than the multi-model Modus Tollens (MT) conclusion. Usual metacognitive assessment used in inference tasks is the judgment of confidence. The confidence judgment for provided answers can vary in accordance with the inference accuracy. And yet, judgment confidence also depends on the uniformity of conclusions among participants, otherwise known as consensuality. The aim of this research was to investigate what kind of effects conclusions with complex conditionals which include conjunctions and disjunctions (if A and B, then C ; if A or B, then C) would have on inference and metacognitive assessments. In pre-study participants (N = 19) provided conclusions to the premises. The analysis of their answers indicated a difference in the level of consensuality for these two types of conditionals. According to this, higher metacognitive assessments are expected for the conjunctive conditionals. The main experiment included 30 participants. The design of the experiment was 2x2x2, for which the factors of the type of conclusion (MP-MT), the validity of the conclusion (valid-invalid) and the type of conditional (a conditional which includes either a conjunction or a disjunction) were manipulated. Each participant completed a total of 64 tasks. Two premises were presented on the screen and then the conclusion followed. Participants’ task was to verify the conclusion as quickly and correctly as possible, and after this to evaluate the confidence of their answers (from random guessing to complete confidence). The participants reacted faster to MPs compared to MTs (1629ms and 2518ms), to valid compared to non-valid conclusions (2043ms and 2499ms), and for tasks that included conjunctions compared to those that included disjunctions (1982ms and 2589ms).The results show the significant effect of conclusion type and the conditional type on the response time (F(1, 29) = 34.99 ; p < 0.01) ; (F(1, 29) = 26.27 ; p < 0.01) and accuracy (F(1, 29) = 12.61 ; p < 0.01) ; (F(1, 29) = 127.37 ; p < 0.01). The judgment confidences, in accordance with the starting hypothesis, were significantly greater for conditional tasks which include a conjunction than for disjunctive conditionals (F(1, 29) = 17.37 ; p < 0.01). This result is in line with Koriat’s theoretical framework because the conjunctive conditionals have greater consensuality than disjunctive conditionals and therefore higher confidence judgements.

deductive reasoning ; conditionals ; metacognitive assessments ; judgment of confidence

nije evidentirano