The Musicality of Non-Musicians:
Measuring Musical Expertise in
Britain
Daniel Müllensiefen
Goldsmiths, University of London
Why do we need to assess
musical sophistication?
Need for a reliable tool to quickly assess musical
expertise in music cognition research
We cannot rely solely on musical training as a proxy for
musical abilities
•
“Expert listeners” with limited formal training
No standardised questionnaire instrument to assess
skilled musical behaviours
Need to recognize multiple facets of musical expertise:
•
•
•
•
Music writers, commentators, critics
Music educators
Music producers, recording engineers
DJs
Earlier tests of musical
expertise
Self-report questionnaires:
Cuddy, Balkwill, Peretz, & Holden (2005), Ollen (2006), Werner, Swope, &
Heide (2006), Chamorro-Premuzic & Furnham (2007), MacDonald &
Stewart (2008), Chin & Rickard (2012)
Musical ability tests:
Seashore, Lewis, & Saetveit (1960), Wing (1962), Bentley (1966), Gordon
(1989), Wallentin et al. (2010), Law & Zentner (2012)
Conceptual suggestions:
Hallam & Prince (2003), Bigand (2006), Honing (2011), Levitin (2012)
Missing:
(Focus on musical expertise) x (Covering wide range of musical
skills) x (Combining self-report and objective testing)
What is the Goldsmiths Musical
Sophistication Index (Gold-MSI)?
A new definition of musical sophistication
A new self-report inventory
A new battery of listening tests
A lot of data
Joining self-report questionnaire and ability tests into
one research tool and make it freely available
A lot of data...
Pilot study self-report inventory with BBC LabUK (n = 488)
BBC LabUK online implementation How Musical Are You?
(n ~ 148,000)
Lab study testing reliability and correlation with other
tests (n = 53)
2 studies investigating the validity of self-report inventory
and the correlation with personality traits (n = 224, n = 44)
5 extended lab studies for optimisation of listening tests
(together: n ~ 600)
Online implementation for Channel 4’s Hidden Talent Show
(n = 3,793)
Testing tapping abilities of visitors at London Science Museum
(n ~ 300)
Analogues to musical expertise
Review of expertise research literature in other
domains
•
•
•
Art expertise (Augustin & Leder, 2006; Leder & al., 2004)
Wine (Hughson & Boakes, 2002, Hughson & Boakes 2009)
Physics (Chi, Feltovich, & Glaser, 1981)
Greater expertise in one domain is associated with
more structured cognitive representations that
exhibit a clearer hierarchical organization (Ericsson &
Smith, 1991; Glaser, 1994; Honeck et al., 1987)
A new definition of musical
sophistication
Musical Sophistication:
•
Psychometric construct comprising musical skills, expertise,
achievements and related behaviours across a range of
facets measured on different subscales.
Assumptions:
•
•
•
Facets of musical sophistication can develop through active
engagement with music in its many different forms.
Individuals vary in their level of sophistication on the different
facets.
High levels of musical sophistication are generally
characterised by
•
•
•
higher frequencies for exerting the musical skills or behaviours
greater ease, accuracy or effect of the musical behaviour when
executed
a greater and more varied repertoire of behaviour patterns
associated with it
Comparing with earlier models
Definition of musical sophistication builds on similar
concepts introduced earlier:
•
Hallam & Prince (2003) and Ollen (2006) also stressed the multidimensional nature of musical sophistication including aural skills,
receptive responses, and the different abilities to make music
What is new in our definition:
•
•
•
•
•
•
Emphasizes other skilled musical behaviours besides instrumental
practice
Is not biased towards art music
Assess real-world and high-level musical skills
Includes a self-assessment of musical skills
Models musical sophistication as a continuous parameter
Is explicitly linked to cognitive theories of expertise in other
domains
The BBC’s How Musical Are You?
test
The BBC’s How Musical Are You?
test
How Musical Are You? Test implemented by BBC Lab
UK and promoted on the BBC network in 2011
148,037 participants
70 items self-report inventory on musical background,
behaviour, self-assessed skills
Four ‘objective’ perceptual and production tests
Minimal socio-demographic data
Duration: ~25 minutes
Final feedback on “relationship with music” as well as
objective scores
The dimensions of musical
sophistication
Data: 147,633 participants responding to 70
question items
Analysis goals:
1. Identify latent factor structure and ‘cluster’ items
into subscales
2. Refine and shorten subscales
3. Assess internal reliability on fresh dataset
Techniques: Factor analysis, item response
models, structural equation modelling
Factor analysis on the How
Musical Are You? data
Best model: one general factor and 5 dimensions
•
•
•
•
•
Active engagement
Perceptual abilities
Musical training
Singing abilities
Emotions
Item response theory used to shorten the questionnaire
•
•
5 scales comprising 6 to 9 items; 38 items in total
Cronbach’s alpha between .789 and .900
Strong general factor of
musical sophistication
•
Evidence: High eigenvalue
of 1st factor, high inter-factor
correlations, high ωhierarchical
5 distinct dimensions of
musical sophistication
They can be measured by
6-9 items each
•
Evidence: Good overall
model fit (RMSEA=.06,
SPMR=.06, CFI=.88,
TLI=.87)
•
Evidence: High internal
reliabilities of subscales
(Cronbach’s α > .79)
Test-retest reliability and
comparison with other scales
Convergent
Validity:
Correlations with
AMMA (n=44)
Test-Retest
Reliability (n=53,
mean time lag=
64 days)
Convergent
Validity:
Correlations with
MEQ (n=141)
Active Engagement
.90**
.20*
.41**
Perceptual Abilities
.89**
.32**
.51**
Musical Training
.97**
.40**
.43**
Singing Abilities
.94**
.42**
.43**
Emotions
.86**
.19*
.32*
General Musical
Sophistication
.97**
.45**
.50**
Musical sophistication and
personality traits (TIPI)
n=141
Active
Engagement
Perceptual
Abilities
Musical
Training
Singing
Abilities
Emotions
General
Sophistication
Extraversion
.20**
.29**
.27**
.34**
.18*
.33**
Agreeableness
.10
.19**
.10
.19*
.14*
.18*
Conscientiousness
-.13
-.08
-.12
-.12
-.16*
-.16*
Emotional
Stability
.08
.18*
.13
.13
.04
.16*
Openness
.39**
.36**
.30**
.33**
.41**
.43**
Extraversion
(Eysenck)
.33**
.31**
.19*
.44**
.28**
.35**
Musical sophistication and
personality traits
High scores on Openness to Experience associated with:
•
•
•
Cognitive ability (Chamorro-Premuzic et al., 2004)
Aesthetic interest (McManus & Furnham, 2006; Furnham &
Chamorro-Premuzic, 2004)
Powerful emotional reactions when listening to music (Vuoskoski
& Eerola, 2011)
Interim summary
The Gold-MSI self-report inventory is a valid and
reliable measure of different facets of musical
sophistication
It comprises 5 factors and 1 general factor
It is based on self-assessed skills and self-reported
behaviours
⇒
How does self-reported sophistication compare to
performance in listening tests?
Listening tests
13-item Melodic Memory test:
•
•
•
AB comparison
novel folk tunes
akin to Dowling & Bartlett (1982) and Cuddy & Lyons (1981)
Listening tests
17-item
•
•
•
Beat Perception test:
correct/incorrect judgement
unknown instrumental tunes from rock, jazz, popular classical
variant of Iversen & Patel’s (2008) Beat Alignment Test
Listening tests
16-item Sound Similarity
test:
•
•
•
800ms audio excerpts from
typical rock, pop, hiphop,
jazz songs
Sorting paradigm similar to
Gingras et al. (2011)
Inspired by Gjerdingen &
Perrott (2008) and
Krumhansl (2010)
Listening tests
129,560 participants provided complete data for all three
tests plus the self-report inventory
Musical Training benefits
melodic memory (β=.22 ) and
beat perception performance
(β=.17)
Modelling self-reported and test
performance
Perceptual Abilities benefit
melodic memory (β=.15) and
beat perception (β=.17)
Accurate sound similarity
judgements are related to
Active Engagement (β=.11) but
not to Musical Training (β=.03)
Modelling self-report and test
performance
The general musical
sophistication factor
indexes all three test
scores
The three listening
tests measure
different abilities
=> low inter-test
correlations (r < .15)
Musical sophistication and
socio-economic variables
How does self-reported musical sophistication relate to
socio-economic variables?
Data: 90,474 Brits from How Musical Are You? sample
Analyses:
•
•
•
Random Forest regression
Post-hoc analysis based on permutation tests
Regression tree models (Strobl et al., 2009)
SES affecting self-reported
General Musical Sophistication
Higher musical sophistication: Younger people, ‘creative’ professions (media,
music) and flexible occupational status (at school, uni, self-employed)
But: R2 = .045
The role of wealth
How does wealth of local area affect test scores and selfreported Musical Sophistication?
Data:
•
Averages (self-report, test scores) by local authority of ~90,000
Brits from How Musical Are You? Sample
•
Median income by local authority from 2011 Annual Survey of
Earnings (National Office of Statistics and Ordnance Survey)
Correlations with income
Musical Training ~ Annual Income (r = .31)
Active Engagement ~ Annual Income (r = .01)
Correlations with income
General Sophistication ~ Annual Income (r = .14)
Combined test score ~ Annual Income (r = .40)
‘Musical landscapes’
Summary
Gold-MSI is a valid and reliable self-report inventory and test
battery for musical skills and expertise.
Identification of psychological and social conditions related to
musical sophistication:
•
•
•
•
Intensive engagement with music is highest in early, flexible periods
of life
Musical behaviour and skills seem to change with life
circumstances (occupation, age)
Substantial correlation of regional income with musical training and
objectively assessed listening skills
But: Main proportion of individual differences not explained by SES
factors
All components of the Gold-MSI:
•
•
•
Are freely available for research purposes
Are fully documented
Have data norms derived from an adult population
Next Steps
Causes, Consequences and Correlations:
Genetics:
Heritability of hearing ability and musical listening skills: Twin
study with DTR, St Thomas‘ Hospital
Neuroscience:
Neuroanatomical correlates of individual differences in basic
auditory and high-level musical skills: MRI study with ICN
Psychological and social factors:
Secondary data analysis of joined datasets from Personality
Test (Cambridge) and Great British Class Survey (LSE/Uni
Manchester)
How to get the Gold-MSI
Materials for download (documentation and audio
stimuli):
http://www.gold.ac.uk/music-mind-brain/gold-msi/
Online version of the tests:
http://www.gold.ac.uk/music-mind-brain/goldmsi/online-tests/
BBC version: “How musical are you”
https://www.bbc.co.uk/labuk/experiments/musicality/
Thank you!
The Gold-MSI Core Team: Very helpful people:
Amit Avron
Daniel Müllensiefen
Thenille Braun
Monika Ruscynski
Jason Musil
Naoko Skiada
Lauren Stewart
Katharina Bauer