FEATURE20 December 2022

‘No rocket science’: The bid for inclusive nat rep samples to become standard

Features Trends UK

A recent study from the MRS Representation in Research steering group found no cost penalty or negative impact on research results when inclusive parameters are included in nationally representative samples. Research Live spoke to Matt Reynolds of Vitreous World to discuss the findings and why fully representative sample is the future.

People in a train station

In recent years, some in the research industry have started to question the current industry standard for ‘nationally representative’ research sample, which is typically restricted to age, gender, social grade and region, and does not include minority quotas – for example, ethnicity or sexual orientation.

In early 2021, as part of its wider diversity and inclusion strategy, MRS set up a steering committee, chaired by Rebecca Cole of Cobalt Sky, to examine the issue of representation in research and how the industry engages with it. A big part of the Representation in Research (RinR) group’s work is identifying concerns – and ultimately dispelling some myths – about what exactly constitutes inclusive research and the perceived barriers involved.

Qualitative research with clients, agencies and panel companies undertaken by RinR in 2021 found an inherent frustration with the nat rep status quo, which involves data points being added as needed.

As part of the RinR work, Matt Reynolds and Bex Grove of Vitreous World conducted a research project to examine the difference in results and insights when undertaking research using only the ‘traditional’ nat rep sampling parameters of age, sex, geography and social grade compared with when additional inclusive parameters of ethnicity, gender identity, sexual orientation and physical disabilities and/or mental health conditions are applied. The study used census data from ONS to add quotas on ethnicity, sexual orientation and disability, with the three combined comprising 30% of the inclusive sample.

The study also sought to investigate whether there is a commercial benefit of more representative research. It focused on omnibus research, a popular means of reaching a sample of 2,000 UK respondents with results returned within 48 hours on average, and which traditionally uses standard demographic quotas, reflective of census data.

Discussing the context for the work, Reynolds, who is a member of the RinR steering group, says “inertia” in the industry had created a barrier to addressing the inclusive sample issue. “Suppliers were saying that clients didn’t really want these quotas and clients were saying that suppliers said they couldn’t reach them under the same budget. You had this kind of stand-off, with everyone saying ‘it can’t be done’. The vast majority of people in senior roles aren’t from minority groups, and probably don’t want to upset the apple cart. Everyone was a bit too comfortable, really.”

Vitreous World originally looked into the issue in 2020, finding that it could deliver research based on sample inclusive of minority quotas without incurring any additional costs. The company later co-launched the Voices4all initiative, penning an open letter to the industry, with the aim of creating a minimum standard for nat rep UK research samples that also proportionately reflect sexual orientation, ethnicity and disability.

“There was no rocket science; we just went to all our normal suppliers and asked: ‘can we test if we can do this?’ In less than 24 hours, we found that we could do it,” says Reynolds.

Work from RinR found that while industry stakeholders were fundamentally supportive of exploring the definition of nat rep, they wanted to know how to address the perceived barriers (one of which is increased cost) and know if there was a potential commercial benefit to more inclusive research – the impetus behind the latest study from Reynolds and Grove.

The study found no cost penalty for the inclusive sample compared with the mainstream nat rep sample (fieldwork and methodology was designed to match the industry’s current commercial pricing structures). Specifically, there was no difference in the price per complete for a minority group interview, compared with a mainstream group interview, and all data sets studied had the same cost regardless of the level of inclusion.

Furthermore, the research has identified a commercial opportunity – that of accessing and understanding the views of a group that totals almost a third of the UK population, which Reynolds describes as “untapped”. While the inclusion of minority quotas didn’t significantly change the total country view of what the UK thinks, ringfencing the quota minorities highlighted important differences in attitudes – differences that would be missed had these quotas not been included.

When you look at the cross breaks and ringfence the quota minorities, the difference in insight and feeling around a wide range of things was wildly different from the mainstream,” says Reynolds.

An example of this was a question asking participants to describe their dietary habits. Respondents from any minority group (across LGBTQ+, ethnic minorities and disabled people) were far more likely to report that they were flexitarian, compared with the mainstream sample.

Reynolds says: “Currently, even the biggest players in the market are not including these people in their methodology. If there’s a third of the market you’re not speaking to, in any kind of business, it’s untapped.”

GDPR question

One of the barriers for the industry, highlighted by feedback shared with the RinR group, is the belief that due to the General Data Protection Regulation (GDPR), panel companies cannot profile individuals based on information about their ethnicity, sexual orientation or disability, because it is too sensitive to collect. This perception means panel companies are unable to profile respondents and therefore cannot predict feasibility, meaning costs remain high.

However, there is no restriction on researchers processing special category data, provided they meet the requirements of Article 6 and Article 9 of GDPR and practitioners document what they do when collecting special category data and how they do it. (MRS has addressed this, and other questions, as part of a ‘frequently-asked questions’ document produced by the RinR group). 

MRS’s current best practice recommendation is that the characteristics required to be ‘genuinely representative of the current UK population, and therefore truly representative’, are: age, gender identity, region, social grade, ethnicity, sexual orientation, and physical disability and/or mental health conditions.

A new standard

Reynolds hopes that use of minority quotas will become standardised across the board and says adopting this approach would give tangible proof of companies’ inclusion efforts. “What you are saying to anyone you work with is ‘we are truly inclusive in everything we do. In the work we do, we are working every day to make sure we are as inclusive as possible.’

“Everyone uses nat rep all the time, not just the omnibus companies. The successful outcome would be the standard quotas on an RFP that you receive will have the minority quotas – those things showing up in our day to day work and becoming the standard.”

If each panel company were to profile on ethnicity, for example, Reynolds believes that they would soon find out that they have many panellists from minority groups. In theory, if these respondents were generally more available, it would mean the price involved in conducting research with them would go down, ultimately making a more diverse group of people more accessible to clients.

“It opens up access to huge swathes of respondents, which is just good for everybody because it means that products being refined are representing everyone – not just middle England,” says Reynolds.

In addition to the frequently-asked questions document, the RinR group and MRS have developed best practice guidelines for inclusive research, with specific recommendations on collecting data on ethnicitysexual orientationextended gender and physical disabilities and/or mental health conditions.

Going forward, RinR will continue to build more resources to support the sector. The committee plans to re-run its company partner survey on current nat rep practices to track any shifts, and will also run a second phase of the commercial benefits research in 2023.

It also intends to set up a sub-group to support the sector with the integration of the new census results – due to be published in spring 2023 – as well as a ‘panel taskforce’ to investigate different ways in which sample providers can increase their representation of and engagement with minority groups.

Additionally, work will begin on looking beyond the UK, investigating how making nat rep research more inclusive can be done on an international level.

Demographics, by their very nature, continue to shift, so this issue is only going to become more pressing. A recent report from WPP, ‘The Consumer Quality Equation’, estimated that the number of people from ethnic minority groups will double to almost a third of the UK population by 2061, with an estimated annual disposable income of £575bn.

Brands risk being left behind if their decision-making is based on research sample using only traditional demographics and doesn't factor more representative parameters into the mix.

Reynolds believes the research industry must admit its unconscious bias in order to move forward in making research sample truly representative of the UK population. He says: “Companies can worry about commercial viability and the risk involved but everyone is involved in this change. We’ve all had this unconscious bias, but once we isolate it and investigate it, once we all agree, we can change very quickly. There’s actually more opportunity on the other side of this wall, and it’s not a big wall. You can step over the fence – you don’t have to jump.

“By 2060, a third of the population will be from some kind of ethnic background so if you really want to your brand to do well going forward, you’ve got to stop pretending everybody’s white.”

0 Comments