The numeric understanding measures: Developing and validating adaptive and nonadaptive numeracy scales

Abstract Numeracy—the ability to understand and use numeric information—is linked good decision-making. Several problems exist with current numeracy measures, however. Depending on the participant sample, some existing measures are too easy or hard; also, established often contain items well-known participants. The article aimed develop new understanding (NUMs) including a 1-item (1-NUM), 4-item (4-NUM), adaptive measure (A-NUM). In calibration study, 2 samples ( n = 226 264 from Amazon’s Mechanical Turk [MTurk]) each responded half of 84 novel items. We calibrated using 2-parameter logistic item response theory (IRT) models. Based parameters, we developed 3 measures. subsequent validation 600 MTurk participants completed Berlin Numeracy Test, Weller Rasch-Based in randomized order. To establish predictive convergent validities, also judgment decision tasks, Raven’s progressive matrices, vocabulary test, demographics. Confirmatory factor analyses suggested that 1-NUM, 4-NUM, A-NUM load onto same as NUM scales showed similar association patterns subjective cognitive Finally, they effectively predicted classic effects. fact, based power analyses, 4-NUM appeared confer more detect effects than Thus, IRT, brief without sacrificing construct scope. can be downloaded Qualtrics files https://osf.io/pcegz/ ).