Abstract
The Berlin Numeracy Test is a psychometrically sound instrument designed to quickly assess statistical numeracy and risk comprehension in educated samples (e.g., college students or medical and business professionals). The test is available in multiple languages and formats including an online adaptive test that automatically scores data ( http://www.riskliteracy.org ). In this chapter, we review results of a validation study (n = 300) documenting convergent (e.g., cognitive ability, numeracy), discriminant (e.g., personality, life satisfaction), and predictive validity (e.g., numerical and non-numerical risky choices). The Berlin Numeracy Test was found to be the strongest predictor of a battery of everyday risky decisions (e.g., evaluating claims about medical treatments, consumer goods, and interpreting forecasts), providing more than twice the predictive power of other numeracy instruments. The Berlin Numeracy Test also accounted for unique variance beyond other related cognitive tests (e.g., cognitive reflection, working memory, and intelligence). Twenty additional validation studies (n = 5,036) indicated that the Berlin Numeracy Test maintained psychometric discriminability across 15 countries (e.g., China, England, Germany, Japan, India, Pakistan, Spain, Sweden, and the USA) and various samples (i.e., community samples, Mechanical Turk web panels, medical professionals). Discussion centers on construct validity and the benefits and limits of adaptive testing.
