I'm having an interesting (to me) email exchange with a reader about the limitations of what NextStage's Evolution Technology can do regarding recognizing an individual on a website is a republican or a democrat. I started documenting the exchange in A Reader Asks "Do Republicans and Democrats Think Differently?" The exchange continues below. My responses are in italics.
Is there anything in that article below that would provide that marker? Something along these lines: The match-up was unmistakable: respondents who had described themselves as liberals showed "significantly greater conflict-related neural activity" when the hypothetical situation called for an unscheduled break in routine.
Conservatives, however, were less flexible, refusing to deviate from old habits "despite signals that this ... should be changed."
Howdy,
Oh, there's plenty in that article and we can detect what they're describing. My concern is that what they're describing is far too general. The argument would be basically the same for Neural Age and Neural Gender; sometimes an individual has male qualities and sometimes female, so saying with 100% accuracy that someone is male of female is a no-win situation.
We can state something like "right now, while that person was navigating your website, they were displaying more male than female behaviors", and similarly for age. Could we determine a marker and detect for it? Yes, definitely. Is there enough evidence that these markers are panthetic across the entire voting population? That I'm not sure about. Can we state that someone will definitely vote democratic or republican based on their navigation habits? No, only that at the time they're navigating your website they are thinking like a republican or democrat.
Now, the conjecture can be made that across all our user populations we could pool the data and come up with an indication of how people are likely to vote, but I would want a much broader set of sites in our inventory before doing that as anything other than an academic exercise.
So again, is it possible, yes. Are we capable of doing it? Yes. Right now? Yes. Would it be accurate? I don't think so, although you've given me something to consider.
Joseph
Please contact NextStage for information regarding presentations and trainings on this and other topics.
Links for this post:
- IMedia Brand Summit on 9-12 Sept 07
- XChange on 20-21 Sept 07
- DC Emetrics Summit on 14-17 Oct '07
- Society for New Communications Research Annual Research Symposium & Awards Gala on 5-6 Dec 07 in Boston.
Comment Preview