Spatial audio technologies become very important in audio broadcast services. But, there is a lack of methods for evaluating spatial audio quality. Standards do not take into account spatial dimension of sound and assessments are limited to the overall quality particularly in the context of audio coding. Through different elicitation methods, a long list of attributes has been established to characterize sound but it is difficult to include them in a listening test. A previous study aimed at clustering attributes in families. Thus 3 families of attributes were highlighted, ''timbre", ''space" and ''defaults". The overall quality and these three families were evaluated in the listening test presented in this article. The test protocol was based on the Mushra recommendation. However it included three anchors specific to each attribute and no reference in order to evaluate quality instead of fidelity. The aim of the experiment described in this paper was to verify the influence of those 3 attributes on the overall quality in a 5.1 reproduction system. It results that the defaults attribute has more influence on the overall quality than the timbre and the timbre. Moreover the presentation of the three attributes on the same interface adds no bias.