With so much research taking place in the lead up to the royal wedding, we had a great opportunity to see how the public's claimed intentions compared with actual behaviour.

 In a YouGov poll in the days leading up to the big event, 42% of adults said they planned to watch the wedding; 31% of men and 61% of women. This was up significantly on the same study two weeks earlier, when just a third of Brits reported an intention to tune in.

 In the event, BARB data tells us that 53% of adults watched some coverage between the hours of 10am (when guests started to arrive) and 1pm (after the parade concluded). Interestingly, the main increase compared to the YouGov survey was amongst men - 44% of men tuned in, along with 63% of women. It should also be noted that BARB only measures in-home viewing, so does not include people watching out of home.

 So, in this case, it seems women were fairly consistent in predicting their viewing behaviour, while men tuned in in bigger numbers than their claimed intention (perhaps an indicator of who was in charge of the remote on the day!). As far as the finding for men is concerned, this is consistent with the commonly found perception amongst the public that they watch less TV than the actual viewing figures suggest - overall TV viewing has remained steady and even increased a little over the last five years, despite claimed behaviour studies often finding that people report watching less TV than they used to.

 The other big event of the last few weeks - the UK referendum on AV - allowed us to see how online behaviour compared to real-world voting patterns. In his book Click, Bill Tancer (General Manager, Global Research at Hitwise) reveals how he and his team are able to predict the winners of reality TV series with some accuracy by studying search volumes for each contestant. This methodology fell short for the final of X Factor (where a win for One Direction was predicted), but search volumes are in general a good indicator of the public's interest in a particular topic, individual or news story.

 As the votes on AV were being cast on the 5th May, Hitwise reported that: "On May 3rd, visits to the YES campaign website were 8.4 times higher than visits to the NO campaign website." In fact, the 'Yes' site received more traffic than the 'No' site for much of the lead up to the vote.

 It's easy to post-rationalise this and there are perhaps good reason for the imbalance between online interest and votes - AV is less understood than the current FPTP system, so perhaps more research was required. The No campaign was also more prominent in paid-for media, so perhaps there was less need for undecided voters to seek out information online. Whatever the motivations of 'Yes' site visitors, this imbalance between interest and ultimate voting behaviour shows the pitfalls on relying on indicators such as web traffic alone to gauge the public's intentions and opinions.

MEC Opinion:
While claimed and online behaviour can be useful and often accurate in predicting real-world behaviour, it's important to be mindful of the gap between what we say we'll do, what we do online and what we ultimately do off-line.