Past studies have shown Twitter users are pretty good at predicting the performance of movies, but now a Carnegie Mellon University study suggests the network could someday be used to glean information about public opinion on all sorts of subjects.
As we note in the lead story of today’s SmartBrief on Social Media, Twitter isn’t without limitations as a polling mechanism, yet data based on tweets seem to track well with traditional polling over the long term. It will be interesting to see how these data are used as the methodology for surveying Twitter becomes more refined.
Would you trust public-opinion figures that were based on tweets? Do you see Twitter becoming a viable alternative to traditional opinion polling?
Image credit, Daisy Daisy, via Shutterstock
Twitter
Facebook
Linkedin
RSS
[...] This post was mentioned on Twitter by Ana Hoffman, Santi Chacon. Santi Chacon said: Would you trust opinion data that came from Twitter?: Past studies have shown Twitter users are pretty good at pre… http://bit.ly/azXXwS [...]
Not in most cases – the population that uses Twitter is not representative. If the topic relates to something that is only relevant to the kind of people that use Twitter, that would be a different story and I'd be fine with it.
I think the types of people Twittering is pretty diverse, but one thing we have in common that we are fairly "tech savvy" and share the common mind set of connecting, networking and publicising. Are Twitters a particular type of person? Perhaps we are? I wonder is there is an age cut off, if so that can't be representative of the general population. ie: If 70+ year olds are not twittering, they are still consumers and active members in society.
Not sure if Twitter can ever replace traditional research tools, it may be a preliminary snap shot to see if further research into a topic is worth while, and possibly representative of certain topics that are relevant to that "type of person on Twitter". ie: If you were to do research on the pension, if the core relevant group isn't on Twitter then how can that be representative.
My recent post Paper Pop-Up BooksA confession: I’ve always had a thing…
Twitter polls may be good for a "straw poll" type of situation but can not take the place of a scientific random poll. Twitter users are not proportionally representative of the entire population and responses in Twitter are from people who proactively give their opinion, which is different from being asked for your opinion because you were randomly (or otherwise) selected. In addition to Twitter, there are all kinds of quick "votes" in other Internet environments, such as AOL and others. The same is true of these responses — interesting, but not scientifically valid.
It's the wrong questions, but as others have pointed out, not representative at all. The 80% of accounts that are inactive aren't involved in the numbers, and Twitter folks not representative.
Why do we have these silly questions looking for black and white answers?
ONLY CHAT
[...] Would you trust opinion data that came from Twitter? (smartblogs.com) [...]
[...] Would you trust opinion data that came from Twitter? (smartblogs.com) [...]
I WANT TO JOIN TO TRUST OPINION CLUB