French validation of the Barcelona Music Reward Questionnaire

Introduction

The rewarding effects of music are highly dependent on cultural and personal preferences. As a result, large differences in the way individuals experience musical pleasure are observed (Blood & Zatorre, 2001; Chanda & Levitin, 2013). While music can induce positive effects on mood and affect in some individuals (Juslin & Västfjäll, 2008), others seek the social bonding capacity of music (Cross, 2001). Conversely, certain individuals cannot experience any pleasure from widely different stimuli, including music—a disorder termed anhedonia (Assogna et al., 2011). Traditionally, it has been hard to assess the sources of this inter-individual variability in music-induced reward. Previous groups have developed questionnaires—such as the BIS/BAS scales (Carver & White, 1994) or the Sensitivity to Reward/Sensitivity to Punishment Questionnaire (Torrubia et al., 2001)—that assess individual differences to overall sensitivity to reward experiences (Carver & White, 1994; Torrubia et al., 2001). However, music is considered as a higher-order pleasure and, as such, might involve different processing mechanisms than basic rewards (Menon & Levitin, 2005). In addition, previous studies have supported a dissociation of music rewarding experience from other rewarding experiences related to other types of primary and secondary reinforcements, such as food, sex and money among others (Mas-Herrero et al., 2014).

In light of these findings, the Barcelona Music Reward Questionnaire (BMRQ) was developed by Mas-Herrero et al. (2013). This questionnaire is specifically geared towards assessing sensitivity to music reward and was a welcome addition to a limited choice of behavioral tools suitable for music reward studies. The BMRQ can serve as a valuable research tool in psychophysical studies addressing music reward in healthy individuals, hearing-impaired individuals or individuals affected with other conditions or pathologies. To date, only Spanish and English versions of this questionnaire are available in the literature, limiting its application. With over 200 million speakers worldwide, French is one of the most common languages in the world. In countries with several official languages including French such as Canada, Morocco and Senegal, it is all the more important for research tools to be available in all official languages to adequately test the population. In that context, a French version of the BMRQ was required to meet the needs of the numerous researchers in the French-speaking areas of the world. In this paper, we sought to translate the Barcelona Music Reward Questionnaire into an international French and to assess its construct validity and reliability.

Materials and Methods

Results

A total of 1,027 participants voluntarily completed the entire translated questionnaire (Mean age: 22.3 (SD 7.8) years, females: 64.7%). While participants were mostly from France (89.4%) and Canada (5.1%), 4% of our sample was obtained from 25 other countries such as Cameroun, Senegal and Egypt. Table 1 resumes the demographic statistics of the sample. The majority of our respondents were non-musicians (77.5%). While the questionnaire was primarily advertised in academic institutions, approximately a quarter of our participants did not complete a university degree. Overall, our French sample was similar to the Spanish and English samples in terms of age, gender and music training.

Table 2 shows the means and standard deviations of items of the French and the Spanish version of the test. As can be observed, the differences observed between the mean items in both cultures were not significant.

Once the polychoric correlation matrix was available, the observed Kaiser–Meyer–Olkin (KMO, Kaiser, 1970) index was computed: the 0.855 value obtained suggested that the correlation matrix was well suited for factor analysis (see Kaiser & Rice, 1974). The congruence values (Tucker, 1951) between the rotated loading matrix and the ideal loading matrix ranged from 0.88 to 0.96. As the coefficients were all above the threshold of 0.85, the factor similarity between the rotated loading matrix and the ideal loading matrix was fair (Lorenzo-Seva & Ten Berge, 2006). Table 3 shows not only the loading values after rotation, but also the loadings of items on the control scale (i.e., the acquiescence, AC). The procedure used to obtain a control scale was previously proposed by Lorenzo-Seva & Ferrando (2009). The method is based on the idea that in a scale where some items are worded in the opposite direction to the other items, it is possible to identify acquiescent response style. In such a balanced scale, the centroid helps to estimate the overall tendency of individuals to use systematically a particular value of the response scales independently of the worded direction of the items (i.e., to show an acquiescent response style). In an initial step, the first centroid is computed, and it is taken as an estimate of the loading values of items on an underlying acquiescent factor. If the scale is partially balanced, a subset of balanced items is used to compute the first centroid, and then the unbalanced set of items is projected on the first centroid. The variance explained by the first centroid is then removed from the correlation matrix, and the residual correlation matrix is factor analyzed in order to estimate the loading on the content factors. In addition, this first centroid can be understood as a control scale: a scale that accounts for the variance due to the acquiescent response. Our results show that some of the items properly loaded on the AC scale. This finding confirms the appropriate use of a model in which AC response bias style was controlled. We were also able to validate that the loadings of items on the content factors were free of AC using this model. Lastly, using the loading values on the content factor, we demonstrated that the items in our adapted instrument were well related with the corresponding expected scale.

In addition, the inter-factor correlation values between content factors ranged between 0.22 and 0.32. While these inter-factor correlations are in general slightly lower than the original version of the test by Mas-Herrero (0.22–0.46), our results demonstrated that the scales were also correlated in the French adaptation.

Finally, the reliability estimates computed on the basis of the factor scores of the scales were 0.93, 0.96, 0.88, 0.91, and 0.93 for Musical Seeking, Emotion Evocation, Mood Regulation, Social Reward and Sensory-Motor, respectively. None of the reliability estimates obtained in our analyses were below the threshold of 0.80. In comparison, the corresponding reliability estimates in the original pooled English and Spanish samples were 0.89, 0.88, 0.87, 0.78, and 0.93, respectively (Mas-Herrero et al., 2013). Furthermore, the overall test (Music reward) in the French translation showed an acceptable reliability (0.91), concordant with the reported value by Mas-Herrero et al. (0.92). The distribution of the overall test scores (global sensitivity to music reward) using the French translation was centered on a mean of 50, similar to that of the original instrument (Fig. 1). Likewise, the internal consistency for the overall French scale was 0.852, with a 95% confidence interval [0.839–0.865]. Globally, all our analyses demonstrated that the fit obtained in the French translation was similar to that of the original English and Spanish samples, indicating that the factorial structures are equivalent.

Discussion

Our study described the translation and adaptation of the BMRQ into French and provided analyses of the psychometric properties of the translated scale. Our results demonstrated that the translated BMRQ has acceptable construct validity while keeping the factorial structure of the original English and Spanish questionnaires. In general, the results that we obtained were similar to those reported by the developers of the original instrument (Mas-Herrero et al., 2013). This suggests that our translation procedure was successful.

The geographic distribution of French speakers encompasses over 30 countries throughout all five continents (L’observatoire de la langue française, 2014). With such a diverse speaker population, significant regional differences in the spoken language currently exist. In that context, an internationally acceptable French adaptation of the BMRQ was required to accommodate researchers and clinicians across the French-speaking regions. This translated BMRQ is born from a collaborative work between two bilingual groups in North America (Montreal, Quebec) and Europe (Lyon, France). Efforts were made during the translation process to remove all regional French influences. Each group first independently translated the original English BMRQ into a locally acceptable French. Then, a consensus between the two translators was obtained to produce the final international French version of the questionnaire. We believe this collaboration was necessary to adapt the original BMRQ into a French that would be easily understood by speakers around the French-speaking world. This belief is echoed in our results: over 30 French-speaking countries are represented, and 5.5% of participants learned French in countries other than Canada (Quebec) or France. Finally, our association with the developers of the original Spanish instrument (UL) ensured the French adaptation remained faithful to the initial questionnaire.

While we collected responses from over 1,000 participants, the majority were obtained from France (89.4%). This is partly a reflection of the differences in the number of French speakers between the regions sampled: 6 million in Quebec compared to more than 77 million in the European Union (L’observatoire de la langue française, 2014). In an effort to reduce sampling bias effect, the music focus of the study was not explicitly stated in the test instructions when administered to participants. This can be seen in the number of non-musicians among our participants (77.5%), a proportion that is similar to what has been reported in the original version of the BMRQ. We therefore believe our sample is representative of the general French speaking population and that sampling bias was not significant.

Previous work by Ayotte, Peretz & Hyde (2002) and Peretz, Champod & Hyde (2003) have established that approximately 4% of the population suffers from congenital amusia, a disorder of music processing that hinders their ability to perceive, produce and enjoy music. In contrast, some individuals suffer from general anhedonia, a deficit in experiencing pleasure from widely different stimuli, usually in the context of depressive disorders or neurodegenerative diseases such as Parkinson’s (Loas et al., 1994; Assogna et al., 2011). Three case studies have also reported a form of acquired anhedonia specific to music that resulted from strokes in limbic structures such as the amygdala, as well as areas of the temporo-parietal cortex, inferior parietal cortex and insula (Mazzoni et al., 1993; Satoh et al., 2011; Griffiths et al., 2004). In those neurologic and psychiatric patients, the use of a standardized tool such as the BMRQ will help determine a loss in the capacity of feeling emotions through music. However, the BMRQ can also be employed to explore music reward in healthy individuals. In fact, a recent report by Mas-Herrero et al. (2014) was the first to identify a group of healthy people for whom music is not rewarding. The term coined—specific musical anhedonia—refers to a unique subset of the population that draws no pleasure at all from music despite being perfectly able to experience pleasure in other ways. Using a stepwise regression analysis, Mas-Herrero et al. found the BMRQ score to be the only predictor of high-pleasure or chill responses in all their participants (compared to other reward scales such as the BIS/BAS). Their work has shown that the ability of music to induce pleasure may not be universal, and that there may be individual differences in access to the reward system (Mas-Herrero et al., 2014). To further understand the neural correlates behind musical pleasure and reward processing, further studies in that population are required and the BMRQ could prove to be a very valuable tool.

Conclusion

The French version of the BMRQ appears valid and reliable. The addition of the French adaptation to the previously available English and Spanish versions significantly increases the reach of this scale. We believe it can not only serve as a valuable psychophysical tool in music reward and emotion research, but its use could also be extended to emotion and reward research in other domains and modalities in which music can be used to test the specificity of a given deficit. Clinical applications of the BMRQ include the examination of musical pleasure experience in healthy individuals and in patients suffering from a wide variety of cognitive, neurologic and auditory disorders.

The French BMRQ test is available online at the following URL: www.brainvitge.org/bmrq_french.php.

Supplemental Information