It’s the afternoon of the 19th of January as I write this, not long after Lord Ashcroft released his latest voting intention poll for the general election. According to the New Statesman’s database, this is the 22nd opinion poll released so far this year, and there’s at least one more to be released today. We’re still almost four months from the election and we’re currently averaging over one opinion poll being released a day, not forgetting the many different polling aggregators and polls of polls that update regularly based on the new figures that come in.
Given that I’ve blogged, tweeted and discussed polls many times before, and I’m a student in a department that has consistently used various types of polling data, this may be a somewhat heretical question, but are there just too many polls out there now?
It seems to me that we’re in a situation where we’re getting much more polling information than we’ve ever had before but there’s so much that it’s almost impossible to distinguish signal from noise. Minor up and down movements in daily trackers are dissected and analysed by a Twitter crowd of thousands, and that discussion is then forgotten the next day when the reverse movements are treated with the same reverence, despite it being just a return to the norm. Most of the polls are all within the margin of error with each other, giving us the oddly paradoxical result that only those that are significant outliers from that norm get properly remembered.
The other problem is with that so much polling out there it comes to dominate the conversation because everyone can find a little snippet of polling data that supports the argument they want to make. Want to argue that everyone should note the rise of the Greens? Ashcroft’s latest has them at 11%. Want to argue that UKIP are the main threat to the establishment and the Greens are nowhere? ComRes have UKIP at 18% while the Greens are just on 3. Looking to argue that as the election comes closer, people will stop flirting with minor parties and go back to voting Tory or Labour? Populus have them both over 35%. Want to argue about different polling methodologies and the effect they have on the results? Take your pick, and go for it.
This is in January for an election in May. As everyone steps up their operations as the election comes closer, and more and more attention is paid to the polls that do come out, they’re going to dominate the election discussion even more. We could reach a point in April where there’s a different poll showing opinions UK-wide, in Scotland, for a region, a constituency or the important left-handers under 35 demographic every hour of the day.
Yes, data is a good thing, but too much data thrown at people without the time to process it all is going to lead to everyone cherry-picking the results they like best and ignoring the rest. No one needs to change their mind because everyone can find support for what they want to believe.
We’re about to be buried under an avalanche of data, from which I’m not sure anyone will be extracted without injury. At least the election actually happening might give us a break for a few days, right until someone decides they need to know what the result would be if the election had been held a week later.