What work is the project doing?
Accent Bias in Britain brings together three separate academic disciplines – sociolinguistics, social psychology, and labour market economics – along with a carefully designed experimental approach – to examine the presence and impact of accent bias in the real world.
Our primary activities
The project is composed of three pieces of work.
Examining attitudes to accents in the UK
We carried out two large-scale national surveys to explore prevailing attitudes to British accents across a representative cross-section of the UK.
One looked at how people react to accent categories or labels: what does the term “Cockney” or “Queen’s English” make you think of? The other examined how people react to actual recordings of people speaking in accents, and how these attitudes affect perceived job suitability, according to the age, region, social class, and personal beliefs of the listener.
The first – looking at accent labels — gives us a picture of the stereotypes that people have about different accent groups at a very general level. The second – using actual speech – tells us how people react to the sound of a particular accent whether or not they know what accent they are listening to. This gives us an idea of how people react to each other in real social interactions, rather than more pre-packaged attitudes that might circulate via the media.
Examining the effects of accent bias in professional contexts
This study builds on the tools and techniques used in the national survey, but focuses on accent bias in a specific employment sector – law – and looks at the extent to which accent interferes with lawyers’ objective judgments of entry-level candidates. The study goes beyond simply reactions to accents and incorporates implicit differences in the quality of answers, to test lawyers’ ability to identify competence independent of accent.
The legal profession has been identified as particularly lacking in diversity and prone to subjective measures of aptitude (Ashley et al. 2015).
Developing tools and training resources
This work will test different forms of anti-bias intervention to see which, if any, is most effective. We will produce and implement training materials that convert key findings into action, and directly tackle obstacles to social mobility. Training materials will be targeted at three of the project’s core beneficiaries: the legal sector (lawyers and law firms); HR professionals; and university groups (students, instructors, careers advisors).
Working with project partners and participants, we will design a set of bespoke training resources for each group. We will maintain contact with users of these materials and track progress in order to assess their efficacy, and make further recommendations.
Key research questions
Our research has been designed around five main questions:
To what extent does a candidate’s accent interfere with objective assessments of their knowledge and professional competence?
Is the degree, type, and direction of bias different in Southern and Northern England?
Are patterns of bias similar across age, class, gender, and ethnicity in the UK?
The basis of bias
Do patterns of bias correspond to an individual’s own background (prior exposure to accents, personality) or to prevailing social norms?
Can interventions mitigate the effects of bias?
Our interdisciplinary design
To address our research questions and the gaps in the existing literature, we have brought together theories and methods from sociolinguistics, social psychology, and labour market economics.
Sociolinguistics provides us with a detailed understanding of existing patterns of accent variation across regions and social groups in the UK.
Social psychology provides us with an experimental method for testing listener attitudes to such variation, and a model of how attitudes are formed in a person’s mind.
Finally, theories of discrimination in labour market economics provide us with a framework for conceptualising and pinpointing bias, and allow us to go beyond simply identifying preferences for one accent over another, and investigate how such preferences may impede fair access to employment.
Drawing together these complementary insights, we developed an innovative experimental design that represents a novel approach to the study of accent bias in professional contexts.
Using an experimental approach
We opted for an experimental approach, derived from social psychology, rather than field observation, as the latter is problematic for research on attitudes to accents. This is because in an observational study it is impossible to control for fluctuations in, for example, speech rate, intonation, hesitation, volume, and voice quality – all of which are known to influence listeners.
In an experiment, we can ensure that the audio signal is identical in all respects except for accent.
The experiments were carefully designed to target natural responses, using identical audio stimuli that vary only in accent and use native speaker voices. Experiments were conducted in the workplace, mimicking hiring situations, and involved recruitment staff from the participating firms.
Examining five British accents
We chose to study five easily recognised British accents:
These accents were chosen because together they allowed us to examine listener evaluations across a number of fundamental social contrasts, including region (North vs. South), prestige (standard vs. non-standard), localness (local, supralocal, national), age (established vs. newly emergent varieties), and both ethnic- and class-based associations.