CAN we quantify the impact of playing from memory on listeners' evaluations of a performance? Kopiez et al. (this volume) attempted to do so by using clever audiovisual manipulation techniques to compare the ratings across pairs of performance videos with and without a conspicuously present music stand. The audio tracks were identical in all other ways and were synchronized with the performer's movements. In contrast to an earlier study by Williamon (1999), which the authors sought to replicate, this experimental design ensures that the sound, as well as the temporal association between sound and movements, were exactly the same in both conditions.

THE EFFECT OF PLAYING FROM MEMORY: CAUSES AND CONSEQUENCES

As noted by the authors, playing from memory is nowadays seen as a sign of virtuosity (Chaffin, Demos, & Logan, 2016) and maybe even creativity (Czerny, 1991, p. 70), although this practice was not universally endorsed in the 19^th century (Bitzan, 2010; Ginsborg, in press) and tends to be associated with specific musical genres (Ginsborg, 2004). Whereas an abundant musicological literature discusses the perceived advantages and disadvantages of playing from memory, there is very little empirical research on the topic, with the exception of Williamon (1999). However, the authors rightly point out that, considering that a number of studies have shown that visual information contributes significantly to the evaluation of a live music performance, the presence of a score or music stand may be expected to affect listeners' impressions.

In my view, we should nevertheless make a distinction between the visual impact of performers' expressive movements and that of a music stand (or score). In the former instance, expressive movements or gestures (Vines, Krumhansl, Wanderley, & Levitin, 2006; Wöllner, 2008), as well as facial expressions (Thompson, Graham, & Russo, 2005), have been shown to contribute to the emotional impact of the performance. However, in the case of playing from memory (corresponding visually to the absence of a music stand), we are likely dealing with a second-order, cognitive effect on the evaluation of a performer's virtuosity or (apparent) creativity. In that view, the visible score or music stand may perceptually act as a marker of insecurity, lack of confidence, or lack of virtuosity on the part of the performer, thus negatively affecting the performance's evaluation. In other words, unlike expressive movements, the presence or absence of a stand is not likely to have a direct emotional impact but rather an indirect, cognitive effect. To be sure, playing from memory may allow the performer more freedom with respect to expressive movements, given that he/she is not constrained by the music stand or by looking at the score, thereby possibly indirectly enhancing the expressivity of the performance. However, the design used by Kopiez et al. attempted to minimize this potential confound by asking the performer to synchronize his movements to the same playback track for both conditions (playing from memory and playing with a music stand), by making sure that the music stand did not obstruct the listeners' view of the performer (unlike Williamon's study), and by selecting the most persuasive versions in a pre-test.

As pointed out by the authors (Platz & Kopiez, 2013), the ability to play from memory would thus seem to fall into the category of "impression management", meaning that the presence of a music stand affects the evaluation of a performance through its influence on the performer's perceived competence or "appropriateness" with respect to the audience's expectations, rather than by modulating the perceived expressive qualities of the performance itself. Even though the effect size associated with the presence of a music stand remains relatively modest, this finding, along with earlier results, indicates that future research on music performance evaluation should take into account a broader range of modalities and psychosocial constructs instead of confining itself to a structure-oriented model of musical communication (e.g., Kendall & Carterette, 1990). For performers, these results also imply that the additional work required to learn to play a piece competently from memory is not done in vain, as it may ultimately lead to more positive evaluations or critical appraisals, especially with respect to performers' perceived understanding and spontaneity (Alessandri, Williamson, Eiholzer, & Williamon, 2015).

METHODOLOGICAL CONSIDERATIONS

Turning our attention to the study design, the authors should be praised for including detailed calculations of effect size and a priori statistical power. Furthermore, although replication studies such as this one may lack the glamour associated with more "cutting-edge" research, they are absolutely necessary in a field which has its fair share of issues with reproducibility (Frieler, Müllensiefen, Fischinger, Schlemmer, Jakubowski, & Lothwesen, 2013; Open Science Collaboration, 2015). In these respects, Kopiez et al. establish a model to follow for all researchers in the field. Nevertheless, the experimental design was perhaps suboptimal in some aspects, and I outline a few suggestions for improvement below.