Twitter as Pollster
Over at Pollster.com, I've got a piece that digs a little deeper into the recent study by Carnegie Mellon researchers that uses data from Twitter to approximate various measures of public opinion.
In discussing the implications, I cover a fair amount of ground, but one part in particular merits special attention here on our pages:
"Can the shortcomings of dirty data be overcome by a mix of sheer volume, sound data preparation/manipulation and savvy analysis?"
I believe that ultimately it can and it will. We aren't yet at a point where pollsters need to take down their shingles (nor do I think we will ever be), but we need to expand our purview and begin to incorporate alternative measures of sentiment, attitudes and opinions into our analysis so as to provide more accurate insights into the state of public opinion.
In some broad sense pollsters are entering into their own version of "the end of theory" in which "the data deluge makes the scientific method obsolete." It's time to position ourselves and our industry so that we are able to capture, collect, refine, measure and interpret these new data in a way that can create true knowledge for our clients.
- Alex Lundry
Comments on this entry