I've been talking to Anne Crassweller of NADbank about research and the challenges faced by companies like hers, so I know this is going to be an interesting session.
NADbank collects the data about readership for the daily newspaper industry, how many people read a newspaper every day and throughout the week, and all the related data.
Frank Graves is a major figure in political polling, and has been a big advocate of IVR polling. This is where you pick up the phone, and get to respond to a very small number of polling questions through an interactive voice response. I kind of enjoy these, they are fast, and usually interesting. And faster than a real person, frankly.
Anne: It's all about the sample. But that doesn't mean we can do, or want to do things, the way we did them in the past. Consumers do not all use land lines, so land line methods are inherently biased now.
She says her stakeholders want the data really fast, overnight if possible. So they have been working with Frank Graves at Ekos to develop alternative approaches.
Frank Graves is now up.
[BTW - first presenters to use Prezi]
Online panels are not reflective of households, and have a gazillion other problems for sampling. Challenges for declining response rates, sampling, commercial pressures for faster and cheaper ... So, they have been playing around with IVR, "robo-calls."
They have been refining the methodology to get better and better samples.
Best lines of the day: "Now that's an exciting table!" "Prediction is really hard, particularly about the future."
Anne is back up, talking about how the HD-IVR delivers results for them.
- Gender results are a closer match to census than any of the other methods compared
- Age results are much closer to census than any of the other comparison methods. This is the only method to really get 18-24 people effectively
- Education is very close to the census, again much better than the other methods compared to.
Part of the challenge is that the CATI surveys are much longer, so they have been trying to figure out what questions to shift to IVR.
Cell-phone only households are the challenge, because so many people are getting rid of their landlines. The debate is, does cell-only define your media behaviour? Their tests have showed that, NO, your choice of being a cell-only household is not a major determiner of your media habits.
My quant buddies are probably following the stats quicker than me, but I think this is quite interesting from the standpoint of just getting good, real, data. There really is no substitute for actual facts, is there?