OPINION5 August 2020

Challenging the status quo in methodology

Opinion UK

Inclusion is the hallmark of credible data, and market researchers must rise to the challenge by reviewing their approach to research design and methodology, writes Tara Watkins.


Minority ethnic groups account for 8.1 million people in the UK: that’s 13% of a population that’s becoming increasingly diverse. Yet no matter how sophisticated its techniques, consumer research is somehow failing to break through and extract insights from one of Britain’s fastest-growing and most influential demographics.

The same oversight applies to other minority groups, including individuals with mental health issues (comprising 25% of people in England each year), people with disabilities ( 19% of working age adults) and those who identify as LGB ( 2.2% of the UK population).

These are all significant consumer groups. So how can brands move from ‘we think’ to ‘we know’ – the golden mantle of all robust market research – if these voices aren’t included as an integral and assumed part of that knowledge?

Research cannot claim to be nationally representative if it’s not based on an effort to reach all sections of society. Worse still, by failing to do this it runs the risk of making clumsy and misleading generalisations.

The financial and legal barriers to achieving diversity in data are well documented. But perhaps we can achieve more tangible progress by dissecting the hurdles inherent within research design and methodology.

For example, English is the default language for most consumer surveys, meaning that at least some of the 4.2 million people who speak immigrant languages in the UK are being routinely overlooked by research efforts. 

If market researchers and advertisers are to make good on their desire to be more representative, they need to focus more energy on accessing people who don’t speak English as a first language. Careful consideration is required to recruit, communicate with and serve questionnaires to these panellists; it demands conscious effort and no small amount of resource and logistical planning.

Additionally, greater thought is needed in terms of who conducts research. For example, Islam is the second religion of the UK, equating to 4.8% of the population. It’s pertinent, therefore, to ask whether a greater number of Muslim moderators are needed in focus groups, in order to accrue deep-rooted understandings that extend into this important group.

We also need to reassess how we conduct research. For instance, methodologies should factor in how to reach the 22% of people who have a disability: in some cases this may involve prioritising online interviews over face-to-face canvassing (current Covid-19 restrictions aside) to allow stronger participation from this group. But equally, it’s a question that involves preliminary research into how people with various disabilities would best like to be approached.

Clearly, creating changes like these in market research is a large and multifaceted process: one that demands our collective focus. But researchers can pave the way to a pivotal shift in data acquisition and analysis by playing close attention to these four areas:

Who we speak to: making a conscious effort to ensure minority quotas are included in consumer research, with the outcome of greater nationally representative data or representation within specific regions of focus. This can be achieved by working closely across fieldwork, agency and brand right at the start of the research design process.

How we speak: building greater awareness of the research approach, including the language(s) used to conduct the research and who conducts the research. Researchers also need to consider the kind of methodology, for example online focus groups, that will allow them to connect with the minority groups who wouldn’t usually be able to participate in research.

Reporting on the data: when segmenting results, it’s good to be mindful of cutting the data by a multitude of audiences where appropriate, to capture a wider variety of perspectives. Brands should also interrogate the data that they receive to fully understand which groups were part of the research, and check whether they were fully represented.

Sharing experiences: change like this is not always easy to bring about, so it’s important that we keep the dialogue open. Researchers need to share our learnings across the industry, and seek out advice and learnings from one another.

These actions will carve out the ability to collect new perspectives and insights. They will also ensure that advertisers and market researchers bridge the representation gap so we truly hear all voices.

Tara Watkins is insight lead at the7stars


4 years ago

The reason we don’t recruit our samples for these groups or break down our analyses to reflect their input is that it is rarely relevant to the marketing issue at hand. When we draw a sample, we are supposed to draw it to represent the population we are interested in. We utilize quotas for demographic, psychographic, and/or lifestyle factors that differentially impact behavior, whether that be purchasing, reactions to advertising, or whatever is appropriate. If there’s no demographic or psychographic or lifestyle differences that affect behavior, it only makes sense to ignore these sub-groups as separately-identified entities in your sample and in your quotas. In other words, if sexual orientation has nothing to do with evaluating flavors of toothpaste, you don’t need have a readable subset of each orientation when testing new toothpaste flavors.

Like Report

4 years ago

In some instances we purposely bias the sample to recruit certain demographics, users and communities but even in those instances we should be mindful for both qualitative and quantitative methodologies to be as inclusive as possible; from the language the survey is served in to who runs the focus groups/in-depth interviews. However, when we are running nationally representative projects, in order for us to claim the study is representative at a national level, that requires more sensitivity beyond traditional and somewhat outdated definitions that no longer reflect modern society. Currently, here in the UK the definition for nationally representative research requires quotas on the following demographics: age, gender, social grade and region. It doesn’t include ethnicity and given that 13% of the UK population (increasing in cities) are from an ethnic minority it should be included within the definition of nationally representative research along with sexual orientation. This will ensure that all voices across society are represented within research.

Like Report