Participants rated Intuition, Suddenness, Pleasure, and Certainty accompanying their solutions to items of a popular fluid intelligence test – Raven’s Advanced Progressive Matrices (RAPM) – that varied from easy (around 80% correct) to difficult (around 20% correct). The same ratings were collected from four insight problems interleaved with RAPM. Suddenness and Certainty substantially decreased from easy to difficult matrices (Pleasure strongly overlapped with Certainty). In easy matrices, subjective experience matched that observed during insight problems, suggesting the highly fluent processing resulting in vivid and univocal solutions. By contrast, processing difficult matrices seemed to involve effortful incremental combination of complex information that yielded uncertain outcomes, resembling full-blown analytic problems. Only Intuition, generally rated low, was unaffected by RAPM difficulty. These results suggest that RAPM constitutes a heterogeneous test, with easy vs. difficult items involving relatively distinct types of processing. This novel knowledge can help in understanding the processes underlying solving Raven’s matrices. The study also contributes to the understanding of the validity of subjective ratings as measures of metacognition.

参加者对他们的解决方案与一种流行的流体智力测试（Raven的高级渐进矩阵（RAPM））的解决方案一起对直觉，突然性，愉悦性和确定性进行了评分，从容易（正确率约80％）到困难（正确率约20％）不等。从与RAPM交错的四个洞察力问题中收集了相同的评分。突然性和确定性从容易矩阵到困难矩阵都大大降低了（乐趣与确定性强烈重叠）。在简单的矩阵中，主观经验与洞察力问题期间观察到的主观经验相匹配，这表明高度流畅的处理过程产生了生动而明确的解决方案。相比之下，处理困难的矩阵似乎需要复杂的信息复杂组合，从而产生不确定的结果，类似于成熟的分析问题。只有直觉 通常评分较低，不受RAPM难度的影响。这些结果表明，RAPM构成了一个异类测试，涉及相对不同类型的处理的简单项目与困难项目。这些新颖的知识可以帮助您理解解决Raven矩阵的基础过程。该研究还有助于理解主观评定作为元认知量度的有效性。