engleski

Mathematical reasoning in the Even-Odd task: Bolstering mindware through strategy training

Recently, we introduced the Even-Odd task within the dual-process approach to reasoning. In this task participants are simply asked to judge, as fast as possible, whether the sum of two presented numbers is even or odd. The task has three main conditions ; the even-even, odd-odd and even-odd combinations. It was proposed that the task cues different rapid processes in parallel with congruent or conflicting outputs. In the even-even (e.g. 2+6) condition both mathematical processing and simple matching indicate that the sum is even, however, in the odd-odd combination (e.g. 5+3) a matching bias would lead to the conclusion that the sum is odd, while it is in fact even. Our previous experiment confirmed that the influence of this conflict is clearly measurable both through accuracy and response time, and that the effect is stronger in more complex versions of the task (double-digit when compared to single digit numbers). In this experiment we introduce training with an additional strategy in order to strengthen what Stanovich (2018) would refer to as mindware and bolster mathematical processes. Participants (N=59) were given explicit instructions that the even-even, as well as the odd-odd combinations always result in an even sum, while the even-odd combination is the only one resulting in an odd sum. They then completed trials with textual combinations (e.g. “even + even”) for further training. After the training they completed 128 trials in a 2 (single-vs double-digit numbers) × 3 (even-even vs odd-odd vs even-odd) experiment. We analyzed the results by comparing them to our previous experiment. The results showed that introducing the additional strategy training eliminated the decrease in accuracy due to digit number which was reflected in an experiment by digit number interaction (F(1, 113) = 5.58, p<.05). The same finding was present for response times as well (F(1, 113) = 10.44, p<.01), the difference in response times between single-and double-digit versions of the task was significantly smaller in this when compared to our previous experiment. This effect was particularly stronger in the odd-odd condition reflected in a three-way interaction (F(2, 226) = 7.26, p<.01). However, introducing the strategy did not reduce the effect of heuristic reasoning through the matching bias. Participants were still less accurate and slower in the odd-odd and even-odd when compared to the even-even condition. The result is interesting because we expected the additional strategy to improve performance and reduce the impact of conflicting processes. This may be due to the new strategy not being fully incorporated and practiced. It also may be the case that the strategy simply produced responses in parallel rather than in conjunction with other mathematical processing. These and other considerations will be discussed within the dual-process approach to reasoning while demonstrating the robustness of the new task in differentiating between competing processes.

mathematical reasoning ; dual-process theory ; mindware ; matching bias ; Even-Odd task

nije evidentirano