Personality and music preference

This week I had a very nice email from a gentleman who is kind enough (much like you, dear reader) to glance over my blog now and then. This student was embarking on a music psychology project at his college and he wanted to ask me a simple question to help him develop his project ideas: Why do different people like different types of music?

Well, after reading the question I sat at my computer for a minute or two spouting off ideas like a slot machine fires out coins when you hit the jackpot. Where to begin?! So may things influence how people come to form musical preferences; passive exposure during childhood, the influence of culture, the development of self concept in the formative years, the influence of memories, emotional style, listening habits….and so on!

I tried to be helpful and pointed my young enquirer in the direction of some literature on musical preferences. My advice then was for him to pick one or maybe two of the influences, whatever he was intrigued by the most, and to investigate hypotheses directly from within his chosen dimensions. He thanked me for my reply and that was that.

Then, low and behold, the very next day I can across and article on one of the dimensions which I have not yet mentioned; personality. So in honour of that young man, here is a summary of a recent article entitled ‘Toward a better understanding of the relation between music preference, listening behaviour, and personality’

The article was written by a group from the Philips Research Lab in the Netherlands and they had a rather intriguing method for their study. They wanted to better understand how reported musical preferences related to actual listening behaviour, as attitudes are not always consistent with behaviour in this world. They tracked what people listened to and how long they listened while they were at work (at Royal Philips Electronics) from a database of 70,000 audio recordings over three months. Remember my blog last week about music listening habits at work? This is a simpler form of recording listening habits, but much more extensive –  a rare source of rich data!

The recordings from their database were grouped using the ‘All Music Guide’ (2008) into 16 music genres. They measured the music preferences of 395 people (using the STOMP – Short Test of Musical Preference) of which 267 people provided enough listening behaviour data for further analysis.

The authors originally intended to try to support the findings of Rentflow and Gosling (2003) who had set out four types of musical preference;

• Reflective and complex
• Intense and rebellious
• Upbeat and conventional
• Energetic and rhythmic

However, try as they might (and they tried a LOT of factor analyses!) the present authors could not replicate their framework. Their data pointed to a 6 factor preference model instead:

• Rhythm and Blues
• Hard Rock
• Bass Heavy
• Country
• Soft Rock
• Classical

It is interesting that they chose to stick more closely to genre type labels instead of the more esoteric concepts employed by Rentflow and Gosling (2003). And I wonder how well these models hold up against different cultures, as well as across different periods of time. The authors acknowledge in their paper that a lot of the music listening reflected the current state of industry music sales in the US and UK.

When they looked at musical preference according to the STOMP and listening behaviour they found strong correlations, meaning that the individuals were reliably listening to the types of genre that they reported liking on their questionnaire (measured using listening duration).

This seems a straight forward finding, but I wonder how it would change if you asked people of a younger age group, perhaps one that was less gender biased (the majority of participants in the present study were male), and how it would vary if people had to fill out the STOMP in group situations. That would be a good way to get at social influences on reporting of musical preference.

And what about personality? They found a number of correlations between personality traits (as measured by the Big 5) and musical preference which replicated previous work. These included positive correlations between extraversion and pop, dance, rock and religious music, and a positive correlation with agreeableness and soundtracks. But the only consistent correlations that were found after they considered both reported preferences and listening behaviours were between neuroticism and classical music, and openness to experience and jazz.

The authors make an additional point throughout the paper which I think is very worthy of future consideration.  They argue that studies in this field are hampered by the use of genre labels. Genre’s, they argue, are evolving, overlapping, ambiguous, and vary across individuals as to how they are defined; these are all potential reasons why their data did not replicate those of Rentflow and Gosling (2003).

After less than 10 years the groupings of musical genre seemed to have changed quite a lot between the two studies. In the present study rap and dance genres grouped together in the analysis, whereas in 2003 they were easily separable. Also, blues and jazz were distinct from classical music in the present study whereas in 2003, apparently, they grouped together!  So the use of genre labels may be too unstable a notion upon which to measure musical preference. What might we use instead? The authors suggest trying to relate personality to specific, objective audio characteristics rather than genre. An intriguing idea…

Conclusions

This study has shown that musical preferences can match closely with listening behaviours in an occupational environment (although, as I said, I would love to see this tried out in different populations and situations) and that there are relatively few reliable correlations between personality and musical preference/listening. However, the problem of using genre labels must be considered before drawing conclusions on the latter point. Overall, this is a very interesting study and certainly an eye-opener when it comes to ways in which you can potentially measure listening behaviour when you have access to music industry level technology.