Consider the Democratic Primaries. Do we see any predictive power in internet traffic?
Quantcast Demographic Info:
Hillary
Barack
Compete.com: Hillary vs. Barack
Alexa:
Quantcast:
Conclusion:
It’s tricky! however, I think we need to normalize the traffic by demographic as raw volume is not a good predictor at all (very low correlation between results+exit polls and internet traffic). See here for detailed information on results and polls.
No conclusion yet…
Next Steps:
I will be mashing all this data together to show trends overtime. AND, i will be overlaying it on tools like PolicyMap to show how Internet (general and social networks) + real world policies + polling locations + business all works together.
Amazing that we have all these tools and an incredibly small set of people uses them. Oh, that’s not amazing nor surprising – perhaps frustrating.
Stop looking for the symbolism and attend to the data – the behavior itself. you have more than a irrevalent burb in those two graphs. They represent something! Coorelations are not causal anyhow – even in the traditional sense. There is behavior going on and it means something. Rather than bounce it off the wall against traditional things like the exit polls come up with what Internet traffic does…, not what it doesn’t do. Isn’t that how the talking foxy dicks got it wrong in the last 3 major primaries… they used flat media’s interpretations in an Internet pool…
consider what all those lurkers are doing…
consider what they are doing after they leave the respective sites.
a student of behavior said to me “they have the antecedents, the factors that drove them to the sites, you damn betcha they are doing something after they leave.
just a thought…..
Any symbolism or irrelevant burbs were unintentional. My post was more a laying out of some questions to investigate and some approaches to that investigation.
Website traffic is a response to non website behaviors (consequences of other media activity and so forth) AND a representation of new behaviors (antecedants to what the users will do next).
I want to understand the complexity of internet traffic and these websites impact on voters as well as how accurate they are at measuring voting behavior and/or other relevant behaviors influencing presidential politics.
There is no pundit approach here. My secondary point was that we have access to a ton of behavioral data and we should use it if we want to gain deeper insight into what’s going on in the elections and with internet media as it relates to these interactions.
A blogs informality is a blessing and a curse. It’s great to be able to respond quickly and enter the banter of popular discussion. It’s terrible for robust and detailed analysis. A blog author can always take more time inbetween posts but the longer it is inbetween posts the less frequently people visit and the less frequently people visit the sooner the blog falls out of the fabric of the internet. It’s a very contingent circle.
A thought for you: In addition to reprimanding me for my foxy symbolism, add your own data trails and specific conclusions to further the conversations. Your comment includes a lot of … which reads as if your conclusions are left as trivial exercises for the reader. Let’s spell it out so we can argue specifics.