Challenges and problems for the development of an instrument for the self-evaluation of musical abilities in elementary school age

Abstract

On the basis of previous theoretical and empirical works by i. a. Marsh (1992), Vispoel (1993) and Kröner, Schwanzer and Dickhäuser (2009) an instrument for the assessment of the musical self-concept with the two subscales ‘singing’ and ‘making music’ of primary school pupils was developed. The instrument was tested for validity during a pilot study (n = 88) and a longitudinal study, which is part of the JeKi-program ("For each child an instrument" [Jedem Kind ein Instrument]), whose three measurements have been evaluated (n = 826). Factor analysis provided support for the theoretical assumptions of the structure of the scales. Retest-reliabilities increased during a period of three years. For the subscale singing a retest-reliability of r = .61 for the self-concept in the second and third class could be reported. Correlations between the scales and the external criteria support the validity of the instrument. Furthermore measurement invariance was checked for the three underlying measurement points as well as for gender differences and the grouping factor with children who are experienced in instrumental lessons and children who are not. There is a minimum of metric statistical invariance for the different time points and the grouping factor instrumental lessons and a minimum of strong statistical invariance for the grouping factor gender. Mean-differences between these groups indicate reciprocal effects of social and dimensional comparisons, but couldn’t be explored on the basis of this study.

Keywords: gender, instrumental instruction, longitudinal study, self-concept