Development and validation of RATje: A Remote Associates Test for Dutch children

Abstract

The Remote Associates Test (RAT) is increasingly being used as a measure of convergent thinking, one of the pillars of creative problem solving. As creativity is gaining momentum in Dutch elementary education, there is a growing need for a Dutch version of the RAT suitable for children. We therefore developed a child-appropriate 10-item RAT and examined its psychometric properties in a study with 531 Dutch fifth graders (Mage = 10.35, 52% boys). Results showed that the items were internally consistent (α = 0.73) and accorded with a 1PL-logistic IRT model. Item difficulty ranged from −0.82 (easy) to 1.49 (hard), item discrimination was 1.19 and there were very few signs of differential item functioning based on children's gender, socioeconomic status and language proficiency. However, as strong validity-supporting evidence was lacking, a second study was conducted with 202 fourth, fifth and sixth graders (Mage = 10.82, 46% boys). Internal consistency and item properties were essentially equivalent to Study 1. Validity of the RAT was supported by significant correlations with scores on other convergent thinking tests (0.39 ≤ r ≤ 0.42) and a lack of association with achievement on a divergent thinking test (r = 0.07).