FEATURE12 July 2021

Does the qual and quant divide still matter in research?

AI Features Technology Trends UK

At last week’s inaugural Big Qual Summit, held virtually by MRS, speakers discussed the definition of ‘big qual’ and whether binary methodological divides between qualitative and quantitative research are still important.

Hand holds magnifying glass over a jigsaw puzzle with a piece missing

Researchers today have access to multiple sources of data and approaches to extract value from that data. So, what exactly is ‘big qual’ and what does it mean for traditional methodological distinctions and research skillsets for the future?

During a panel session on the notion of ‘big qual’ and how it can be used, Andy Dexter, managing director, Signoi, said: “‘Big qual’ applies to anything that involves large-scale unstructured data, for example, conversations that people are having on social media around a particular topic. There’s the opportunity to scale it up to millions of tweets – something that you would never deploy in a mainstream research project, but very quickly get a sense of the dominant themes and also surface very small granular themes. In a research sense, it always needs focus, so an example might be big-scale NPS surveys which come with comments.”

Clients increasingly don’t care about the distinction between qualitative and quantitative research, according to Louise Horner, research director, Acacia Avenue.

“We’re hearing clients increasingly saying ‘I don’t care, just give me great insight which has reassurance of some concrete numbers but also the more human perspective and that deeper understanding of motivations and a little bit of consultancy and judgement’,” said Horner. “Quant research done well can create qualitative data and give you structured but deeply human insight and I don’t think the delineation is there in a way in which it once was.”

From clients’ perspective, noted Horner, there is always a need for “very reassuring numbers”, but that qualitative insights can be equally powerful factors. “There’s some sort of alchemy happens when very senior people, often at a great distance from their own customers, are confronted with their own customers’ opinions. Those are the two elements that ultimately influence decision-making in the boardroom.

“Ultimately, it’s all in the service of the same result. It’s just another way of approaching the problem – it’s where you place the emphasis in terms of whether you need mainly numbers or a whether you need a  focus on qualitative insight, but it’s just the emphasis is really rather than it being a totally delineated piece.”

Dexter said: “All of our data is qualitative in research – we might be doing a survey but it’s not necessarily resulting in objectively factual conclusions. It’s based on interpretation of the brief, question, how it’s framed, what the sample is, who’s taking part in the research.”

Gill Ereaut, consultant at Linguistic Landscapes, argued that quantitative research is also interpretative and that researchers should acknowledge the interpretative nature of quantitative research. She said: “Quant can be hideously misleading – you can count the wrong thing and clients get a sense of security based on a flimsy base. So having the overtness and honesty about the interpretative quality of quant research is quite a helpful thing.”

Big qual extends the reach of qual, said Ereaut, who added: “When we use enormous corpus linguistics data sets, sometimes it’s like the NPS set, we can interrogate billions of words. When you’re doing that at scale you can flip it back around and end up with 99% significance rates. This pattern is there in the data to that degree of significance. The client can feel pretty certain that that pattern was there, and then, like all qual, you construct the narrative and meaning in a way that has to pass other forms of validity testing.”

Ereaut added: “Qual is a conscious, systematic approach – it’s not like we’re taking something fluffy and making it rigorous. For me, it’s about looking further and finding more patterns.”

So, what does this mean for the research skills that will be needed in the future? 

Horner discussed the value of integrating approaches such as discourse analysis within quant projects. “When you’re dealing with words on a page in a survey environment, it’s difficult to understand why people are responding in the way they are when you’re looking at scaled data. But if you use discourse analysis properly within a survey you can start to see language patterns, whether that’s done by an algorithm by an individual.

“People might give it the same relevance score but there’s other data underneath that that tells you how they feel through the language they’re using. I don’t know what that means in the longer term for skillsets within an organisation, but I think we all have to be open-minded because it helps us move at pace in a more cost-effective way to give a single project a very big qualitative lens on a quant study.”

While he acknowledged there are clear specialisms and not everyone can be an expert, Dexter said: “I think there’s a false delineation between qual and quant skillsets. Surely as researchers we have to be in the business of understanding information of any sort? That seems to be absolutely core to me.”

Ereaut said: “You have to be nosy, curious and a bit geeky to be any type of researcher. Researchers are inherently interesting because they are inherently interested.”

1 Comment

5 months ago

Great discussion - qual at scale is a big topic, but I wonder if it's more about looking at patterns, accessing outliers rather than really getting to the why, exploring behaviours and motivations in depth? I agree it would be great if more researchers could wear both anlaytics and in-depth hats, but @andydexter - how many qual researchers would be equipped to get a job in an anlaytics environment? The skill sets are different, of course definitely learnable - but i see more specialisation emerging in MR rather than the arrival of the polyglot insight person.

Like Report