Psychometric properties and heritability of a new online test for musicality, the Swedish Musical Discrimination Test

Psychometric properties and heritability of a new online testfor musicality, the Swedish Musical Discrimination Test. Access to the published version may require subscription. a b s t r a c t We examine, in 6881 twin individuals, the psychometric properties of a new test (the Swedish Musical Discrimination Test, SMDT) that was developed to tap auditory discrimination of musical stimuli. The SMDT consists of three subtests measuring discrimination of melodies, rhythms, and single pitches, respectively. Mean test taking times for the subtests were 3.0–4.6 min. Reliability and internal consistency were good with Cronbach's alpha values and Spearman–Brown split-half reliabilities between .79 and .89. Subtests correlated positively (r values .27–.41). Criterion validity was demonstrated in three ways: individuals that had played a musical instrument scored higher than individuals that had not (Cohen's d .38–.63); individuals that had taken music lessons scored higher than individuals that had not (Cohen's d .35–.60); finally, total hours of musical training and SMDT scores correlated (r values .14–.28) among those participants that had played an instrument. Lastly, twin modelling revealed moderate heritability estimates for the three sub-scales. We conclude that the SMDT has good psychometric characteristics, short test taking time, and may serve as a useful complement to existing tests of musical ability. Music is a human universal of profound significance for most people. At the same time, musicality, broadly defined as the capacity to learn and perform music-related tasks, seems to vary substantially between individuals. There have been many endeavours to objectively measure musicality since the early 20th century. Several standardised, explicit forms of musicality tests have been constructed, both with practical aims, such as selection of students for musical training, and for research pur-Exactly how musicality is measured can make substantive differences to the information one obtains and hence to the aspects of the phenomenon one can study. Correlations between different tests and between tests and criteria such as teacher's ratings and music school grades tend to be in the range of .4–.6 (Shuter-Dyson & Gabriel, 1981). One important reason for these relatively moderate correlations is that different tests use different operationalisa-tions of musicality. Indeed, within musicality testing there are two strong traditions, which differ in various characteristics. These are the 'atomistic' tradition of Seashore and the 'omnibus' approach of The atomistic approach is based on the assumption that musi-cality is made up of several relatively narrow and distinct musical abilities. This leads to …