The FAIR-2 (‘Frankfurter Aufmerksamkeitsinventar’) is a pen-and-paper test of visual attention in which participants have to search for targets among distractors. For similar pen-and-paper tests of attention (e.g., d2), the repetition of the test causes large improvements in performance that threaten both its (retest) reliability and validity. We investigated the size and possible sources of practice effects in the FAIR-2 in three experiments. In Experiments 1 and 2, participants were tested twice using the original FAIR-2. We compared how performance changed after 2 weeks (Experiment 1) or 3 months (Experiment 2), when the test was repeated (complete repetition), or when targets and distractors changed their roles (test reversal). For Experiment 3, we used self-constructed versions of the FAIR that allowed for a third neutral condition (complete alternation) without any stimulus overlap between the two tests. The complete repetition condition produced strong performance gains (25–35%) that persisted for 3 months. For the complete-alternation condition, we observed small to moderate improvements, suggesting that stimulus-independent learning had occurred in session 1. Finally, performance did not differ between test reversal and complete alternation, therefore, suggesting that improvements in target processing had caused the large improvements in the complete-repetition condition.

FAIR-2（'Frankfurter Aufmerksamkeitsinventar'）是一种视觉注意力的纸笔测试，参与者必须在干扰物中寻找目标。对于类似的纸笔注意力测试（例如，d2），测试的重复会导致性能大幅提高，从而威胁到其（重新测试）的可靠性和有效性。我们在三个实验中调查了 FAIR-2 中练习效果的大小和可能的来源。在实验 1 和 2 中，参与者使用原始 FAIR-2 进行了两次测试。我们比较了 2 周（实验 1）或 3 个月（实验 2）、重复测试（完全重复）或目标和干扰物改变角色（测试逆转）后性能的变化情况。对于实验 3，我们使用了 FAIR 的自构建版本，允许在两个测试之间没有任何刺激重叠的第三种中性条件（完全交替）。完全重复条件产生了持续 3 个月的强劲性能提升 (25–35%)。对于完全交替条件，我们观察到小到中度的改进，这表明在第 1 节课中发生了与刺激无关的学习。最后，测试逆转和完全交替之间的表现没有差异，因此，表明目标处理的改进导致了完全重复条件的大幅改进。