|
|
05 Jan 2015
Market research projects rarely have the luxury of undertaking a census in the truest form of the word, i.e. interviewing every single person that makes up a given group or sample.
Consequently, the concept of a sample being ‘representative’ appears frequently as researchers strive to deliver answers that speak for the population more broadly.
However, it is valuable to discuss the concept of a ‘representative sample’ in qualitative research. Unlike quantitative research that creates a representative sample through weight of numbers, including statistically measurable levels of confidence based on the size of that sample, qualitative research does not afford researchers the same surety, operating as it usually does in numbers <100.
Over the years, Icaro have seen a number of researchers who believe that ‘representation’ = ‘representative’ where qualitative research is concerned. For example “BME people make up about 10-12% of the population so I need one in my focus group”. This tokenistic approach applies equally to, for example, disabled people, single mothers or those on extremely low income and ends up with the researcher in question claiming their sample is “representative”.
However, it is at best misguided to believe that one disabled person will speak for the entire disabled community. Indeed, by trying to create a qualitative sample in the same way as you would a quantitative one, the risk of falling into this trap is extremely high and the researcher ends up trying to count up quotes or keywords or answers from specific groups to present in some pseudo-quantitative manner.
At Icaro, our approach centres on an experienced moderation and analysis team to ensure feedback is representative, when the actual make-up of the groups may only be broadly technically representative at best. What this means in practice is that if a specific sub-group of people are important enough to the outcome of the research to be included at all, then they are important enough to be included in viable numbers where differences in opinion may start to be observed. Compared to official ONS population data this may mean that the group is over-represented in that sample by a considerable margin – however, by doing this we can see the range of views (if there is one) across these participants and compare and contrast them with the views of technically larger groups in society.
So, the next time you plan a qualitative sample, ask yourself if you’ve fallen into the trap of thinking that representation equals representative…