Evaluation of pitch accuracy in solfeggio examinations: What about non-musical variables?

Background and aims
In experimental settings, 81% of the variance of judges’ rating of pitch accuracy is explained by musical variables (precision of pitch intervals and respect of tonality) (Larrouy-Maestri, Lévêque, Schön, Giovanni, & Morsonne, 2013). In ecological settings, non-musical variables influence judgments of music performances (i.e., McPherson & Thompson, 1998; Platz & Kopiez, 2012). This study aims to better understand the evaluation of pitch accuracy in the context of formative and summative solfeggio examinations at tertiary music level, for which live performances are evaluated.

Method
Twenty-one participants of conservatory were asked to learn simple melodies during solfeggio classes. They were evaluated two times (formative and summative examinations) by 3 judges. Each performance was also objectively analyzed regarding pitch accuracy (number of contour errors, precision of pitch intervals and respect of the tonal centre) with a computer-assisted method (Larrouy-Maestri & Morsomme, 2014).

Results
The 3 judges provided strongly and significantly correlated ratings. The musical criteria objectively analyzed explained 56% of the variance of the jury’s rating when the examination purpose was formative (p < .001) and 31% of the variance when the purpose was summative (p = .009). Interestingly, the predictive musical criteria differed depending on the examination’s purpose. In addition, the variance explained by our statistical model increased (from 56% to 67% and from 31% to 46%) when taking into account non-musical variables such as the gender of the music students.

Conclusions
Besides the educational perspectives, the proposed method appears interesting for examining the influence of non-musical variables on the pitch accuracy assessment in ecological contexts.