Errors in subtractions: judged and observed frequencies.

Area of Science:

• Cognitive Psychology
• Mathematical Cognition

Background:

• Understanding cognitive processes in solving basic arithmetic problems.
• Investigating the distinction between reconstructive and retrieval-based problem-solving strategies.

Purpose of the Study:

• To analyze error patterns and self-correction in subtraction tasks.
• To compare participants' estimations of error rates with actual performance.
• To explore the relationship between metacognitive judgments and the underlying solution processes.

Main Methods:

• Adult participants solved subtraction problems (M-N, M,N <= 13) multiple times, with response times and errors recorded.
• Retrospective verbal reports identified solution strategies: reconstructive (conscious) vs. reproductive (unconscious retrieval).
• A second group estimated error rates for these problems, which were then compared to actual error data.

Main Results:

• A low overall error rate (2.8%) was observed, with a high self-correction rate (92%).
• Participants systematically underestimated errors on difficult problems and overestimated errors on easy problems.
• Accuracy in predicting error-prone problems was moderate (0.67 hit rate), but poor for predicting error-free problems (0.20 hit rate).
• A significant correlation existed between judged and actual error rates for reconstructive solutions (r=0.69), but not for retrieved solutions.

Conclusions:

• Basic subtraction involves both conscious and unconscious processes, with high accuracy and self-correction.
• Metacognitive judgments about performance are biased, particularly underestimating difficulty.
• The accessibility of the solution process to consciousness influences the accuracy of self-assessment.