This post was co-authored by Natasha Stevens and Michelle Morgan
At a time when surveys seem to be under a kind of siege – viewed by some as backward and outdated – let me go out on a limb: Surveys have never been more important or relevant.
Yes, behavioral data can tell us exactly what people do – no guessing or memory jogging required. But there are also restrictions; we may only know, for example, what people are doing within a single environment – online or in store. And while behavioral information can provide extremely rich consumer insight, it often cannot tell us why people do things: what they were hoping for, whether they were disappointed, and their feelings about the brands in their lives.
The challenge for researchers
Surveys can help us fill in all of these gaps; and yet we also know that consumers’ patience with long questionnaires – especially on smartphones – is shrinking. The challenge for smart researchers, then, is: How can we use surveys only when they will provide unique and indispensable information, but quit before our returns start to diminish?
The answer is doubling down on a skill unfamiliar (and perhaps unsettling) to many researchers — data curation. Here we use different data sets, often from very diverse sources, to create the complete picture we need of consumers’ preferences and behavior. By linking two or more data sets through carefully developed criteria, we can focus our survey takers on giving us only the information we can obtain nowhere else.
Data curation in action
One recent example of data curation in action was inspired by the looming changes in US healthcare and insurance. Survey data capturing public opinion on US healthcare reform is abundant – much of it focused on the specifics of the policy itself, with respondents generally profiled according to their political affiliation. We wanted to develop a profile of survey respondents that went beyond party politics and looked more deeply at motivations and personal characteristics around health.
Using KnowledgePanel®, the largest probability-based online panel in the US, supplemented with key health psychographic variables from MRI’s gold-standard Survey of the American Consumer®, we were able to develop a more nuanced picture of our survey takers. The MRI-KnowledgePanel® fusion allowed us to integrate health-related profile data for KnowledgePanel® members – such as body mass index (BMI), information about chronic physical and mental health conditions, and health insurance status – with 25 health psychographic variables from MRI.
We found that those who disapprove of the Affordable Care Act are less likely to believe that generic drugs are as good as brand-name drugs.
In addition, they are more likely to be the first to try advanced medication and more likely to agree that medication has improved their quality of life.
Using the integrated databases, we were able to add a number of high-value characteristics to the mix without additional questions or fees; these included presence of chronic health conditions, medication compliance, body-mass index, and body image.
Deeper insights from fused data
By mastering data integration and curation, we can deepen our insights from any one source. In our healthcare example, the fused data allowed us to develop a richer and more robust profile of survey respondents than we could achieve with KnowledgePanel® data alone. With the right data resources and expertise, this new approach creates almost infinite possibilities for expanded insights.
Natasha Stevens is Executive Vice President of Digital Experiences at GfK. Michelle Morgan is the Research Director of Data and Insights Integration. To share your thoughts please email firstname.lastname@example.org or email@example.com.