Maximum representation: A better way to represent all voices?

Representation in research has increasingly been at the forefront of conversation, with an industry wide recognition that while research aims to understand all people, there have been flaws in both methodologies and mindsets. This has been challenged further by end users seeking to be more inclusive and unbiased in their thinking and outputs. Therefore, the need for market research samples to be unbiased and balanced to deliver truly representative data is paramount.

Steps have been made to have stronger representation within research, with Voices4All launching in the summer of 2020, and the MRS subsequently adjusting their guidelines for nationally representative research to include sexual orientation, disability and ethnicity in quotas, alongside age, gender, social grade and region. This shift to ensure stronger diversity on sampling quotas means that a more accurate and realistic understanding of consumer preferences are collected, that in turn are representative of the actual population.

But we have to stop and pose the question about whether nationally representative research, even when we have quotas in place, is enough to hear all voices? For example, using the guidelines outlined by Voices4All, fielding a 1,000 nationally representative survey, would result in 20 respondents who fall into the quota of LGBT (the quota definition used based on census data), and 30 respondents who fall within the ‘black’ ethnicity group.

While the quotas ensure representation at a total level, because the base sizes for certain audiences are so small, you lose any detail and variance on them as responses are flattened out in the average. Additionally, these base sizes are not statistically robust enough to allow the opportunity to cut the data by these groups to understand how they compare or contrast to the national average.

While it has been fundamental to update the nationally representative quotas, other measures need to be considered when looking to drive inclusive research. Two methods, which can be used within quantitative research include maximum representation (a less common practice within market research) and oversampling.

Maximum representation is the concept and practice whereby we count a person’s multiple characteristics rather than counting them as a single unique individual, essentially double or triple counting an individual depending on the demographic profiles they fulfil. This approach allows the opportunity to increase the volume of underrepresented audiences’ opinions, behaviours and views within a piece of research. This approach works well on standalone pieces that require a broad sample that needs to be cut by sub-audiences and when budgets are constrained.

Similarly, oversampling on particular audiences provides the opportunity to expand audiences beyond bucketed quotas so that we are able to cut the data beyond a total sample to understand how particular audiences over index or under index against key areas of interest.  

How is the right route determined?

• A clear understanding of the research hypothesis is key, as this helps determine what is most important to understand from the insight and, therefore, whether nationally representative research or boosting audiences is more critical
• Knowing where the insight is going to have impact will provide understanding into how the insight is going to be used, i.e. is it for product development, marketing or sales. Understanding this again helps prioritise the focus for the research
• Identifying the consumer opportunity, and therefore who needs to be understood from the research, will inform whether the research will provide sufficient insight into that audience

Alongside quantitative methods, qualitative methodologies can play a pivotal role in helping to unpick the nuances between audiences and critically underrepresented audiences. It’s only by understanding the ‘why’ that we gain deeper understanding of both similarities and differences. Various methods, such as online communities, ethnography, focus groups, video diaries or in-depth interviews, are employed to achieve this understanding.

Only by continuing to put in place systems that improve the inclusivity of market research can we begin to glean the full range of rich insight offered by all the full range of voices.

Tara Watkins is head of insight at The7stars