Although most studies that seek to measure participants’ judgments and attitudes regarding humanoid robots’ possessing (or appearing to possess) a mind or mental capacities have been based on verbal measures, there is yet no standard psychometric instrument for this end. Using a COSMIN approach, this critical review seeks to summarize the most valid and reliable self-report instruments that aim to measure mental state attribution to humanoid robots. 501 papers were reviewed, but only 11 were included, finding that: (1) The instruments do not usually measure mental state attribution toward robots as an exclusive phenomenon but as a factor associated with the tendency to anthropomorphize non-human entities; (2) There is a lack of consensus regarding a definition of mental state attribution and the psychometric dimensions that underlie it; (3) The tendency to anthropomorphize does not by itself imply the attribution of mind to robots. In our discussion, we delve into the general problem of mind perception/attribution and speculate on the possible theoretical basis for a multifactorial model for measuring mind perception as part of a broader phenomenon we term “psycheidolia.”

尽管大多数旨在测量参与者对人形机器人拥有（或看似拥有）思想或心理能力的判断和态度的研究都是基于言语测量，但目前还没有用于此目的的标准心理测量工具。本次批判性审查采用 COSMIN 方法，旨在总结最有效、最可靠的自我报告工具，旨在测量人形机器人的心理状态归因。审查了 501 篇论文，但只纳入了 11 篇，结果发现：（1）这些仪器通常不会将机器人的心理状态归因作为一种独有现象来测量，而是将其作为与非人类实体拟人化倾向相关的一个因素；(2) 对于精神状态归因的定义及其背后的心理测量维度缺乏共识；(3)拟人化倾向本身并不意味着机器人具有心智。在我们的讨论中，我们深入研究了心理感知/归因的一般问题，并推测了用于测量心理感知的多因素模型的可能理论基础，作为我们称之为“迷幻症”的更广泛现象的一部分。