CRAT-RUS: A Set of Remote Associates Test Items for Insight Research

The project contains data and supplementary materials related to the research study: Moroshkina N.V., Oshkanova A.P., Kosyakova A.V., Knyazeva I.S. CRAT-RUS: A set of Remote Associates Test items for insight research – theoretical foundations and validation results / Russian Psychological Journal. 2026. Vol. 23. No. 1. Introduction. Remote association tasks are a convenient and widely used tool for studying insight and creative thinking. However, numerous versions of these tasks exhibit significant linguistic heterogeneity, both within Russian-language sets and between versions in different languages, which hinders the comparison of research results and further progress in understanding the mechanisms of insight problem-solving. Methods. In this study, we developed and validated a new bank of 60 tasks in Russian, comprising two types (semantically convergent and divergent triads). We detailed the criteria for their creation, based on the contemporary understanding of the cognitive and metacognitive processes involved in insight. Task selection was guided by metrics derived from the Russian National Corpus (a representative collection of texts in Russian). The validation study involved 71 native Russian speakers (mean age: 23 years). Results. We found that convergent and divergent triads were balanced in terms of objective difficulty. However, in divergent triads, correct solutions were significantly more likely to elicit an aha! experience (an affective marker of insight) compared to convergent triads. Additionally, the likelihood of successful problem-solving increased with the higher word frequency of the answer and stronger collocational stability between the triad words and the answer. Conversely, the probability of a correct response decreased with the higher frequency of the first word in the triad. Solution time was shorter for triads with higher-frequency answers. Greater semantic distance between triad words predicted a higher incidence of the aha! experience upon correct solution. Low-frequency answers also increased the probability of the aha! experience. Discussion. Thus, we created the bank comprising two types of problems that are equivalent in difficulty but differ in their likelihood of inducing the aha! experience. We also identified linguistic predictors of objective problems difficulty and the occurrence of aha! experiences.