Optimization strategies for the German version of the soundscape affective quality instrument

Abstract

Semantic differences in the soundscape circumplex instrument may have an impact on affective quality assessments of acoustic environments, hampering comparison in cross-cultural settings. Although an adopted version of the ISO/TS 12913-2 exists in Germany, it has not yet been sufficiently validated. In a listening experiment conducted as part of an international round-robin test, a total of 27 soundscapes were evaluated using alternate German translations of the English instrument. Results from two different optimiziation strategies were compared: (1) maximized circumplexity of the German scale version and (2) language invariance between German and English scale versions. Confirmatory factor analyses and invariance testing were employed to determine optimal translations and calculate resulting goodness of fit. Results demonstrate that strategy 1, with two new item translations, yields better results in terms of circumplexity and language invariance than the current DIN ISO/TS 12913-2 version. However, to allow fully unbiased cross-cultural comparisons of soundscape impressions, drawing on the instrument developed with strategy 2 is more suitable, since it produces scores strictly corresponding to those of the English scale. As the analysis demonstrates, this seemingly paradox finding results from a lack of empirical circumplexity of the English original.