Producing polls, plotting personalities and predicting preferences

Some people only take part in research for financial benefit. There’s no point pretending this motivation doesn’t exist, because it does. Instead, it should be recognised and marginalised. This is done through a combination of the metaphorical carrot and stick.

I won’t dwell on the stick, since these mechanisms should be hygiene factors for any research project – triple-checking the legitimacy, consistency and consideration given to responses.

Beyond the obvious requirements of running an interesting research project, it is with the carrot that researchers can get creative. I personally disagree with the ban on client products as incentives, but there are many non-monetary forms of gratitude that help facilitate a relationship based around mutual respect and reward. The opportunities to plant some carrots will vary depending on the nature of the research and the methodology utilised but a couple of (quantitative) examples include:

• Share basic data: Despite Marina Hyde’s protestations, lists are popular. So popular, that they are starting to eat themselves (which is why I like that a blog well-known for hosting a list of lists is called Fimoculous). Where opinions or intentions are aggregated, why not share the results? I might think Newcastle United are going to win the Championship this year, but is it a consensus opinion?

• Provide a basic segmentation: We may not necessarily appreciate them, but labels and stereotypes (or “pen portraits”, as we like to call them) are a useful way to communicate information at a broad level. We may even be curious to see which characteristics best fit our personality, and the promise of answers can be a strong motivation to participate with careful consideration. Quizilla has formed a business out of “What twilight character are you?” type surveys and Channel 4 ran a great example with their UK tribes project (Disclosure: We work with Channel 4, though we didn’t contribute to this project. And I’m an Indie Kid).

• Offer recommendations: This could be either editorial or response-generated. The latter would be more likely to produce “better” results, but the size of the data set needed limits usefulness to  only the biggest projects. In the same way that Last.fm can tell me that as I am listening to Lift to Experience I may also enjoy Six By Seven (I do), we can correlate consumption information to display the “nearest” options to those initially answered

There are multiple ways that we seek to maintain interest and involvement among our respondents, and we are regularly trying new methods to share and exchange . Doing so limits professional respondents, and encourages considerate participants.

sk

Image credit: Ryancr