Abstract The present study aimed to investigate practice effects associated with the abstract rules involved in the Raven's Advanced Progressive Matrices (RAPM) under standard administration conditions. To that end, a linear logistic test modeling approach was used in combination with Carpenter, Just, and Shell's (1990) taxonomy of rules. Several operation-specific learning models were used in order to test different contingent and non-contingent learning hypotheses. The models were fitted to a sample of responses from 293 participants to Sets I and II of the RAPM. A Bayesian framework was adopted for model estimation and evaluation. The perceptual variables involved in the items were included in the analyses in order to control their influence on performance on the RAPM. The results did not provide evidence of rule learning during the RAPM. Instead, they suggested the existence of fatigue effects associated with each of the rules. Interestingly, the results revealed the existence of learning effects associated with the items' perceptual properties.

中文翻译：

---

在 Raven 的高级渐进矩阵中研究特定于操作的学习效果：一种线性逻辑测试建模方法

摘要 本研究旨在研究在标准管理条件下与 Raven 高级渐进矩阵 (RAPM) 中涉及的抽象规则相关的实践效果。为此，将线性逻辑检验建模方法与 Carpenter、Just 和 Shell (1990) 的规则分类法结合使用。使用了几种特定于操作的学习模型来测试不同的偶然和非偶然学习假设。这些模型适用于 293 名参与者对 RAPM 的第一组和第二组的回答样本。采用贝叶斯框架进行模型估计和评估。项目中涉及的感知变量包括在分析中，以控制它们对 RAPM 性能的影响。结果没有提供 RAPM 期间规则学习的证据。相反，他们建议存在与每条规则相关的疲劳效应。有趣的是，结果揭示了与项目感知属性相关的学习效果的存在。