Tweeters Are Different
It's axiomatic in consumer product circles -- movies, toothpaste, that sort of thing -- that people who Tweet are representative of people in general. Know what they're saying about your product in Twitter World and you've got a direct link to your customer.
Not so fast, apparently. From Cornell University comes this study:
Data from Online Social Networks (OSNs) are providing analysts with an unprecedented access to public opinion on elections, news, movies etc. However, caution must be taken to determine whether and how much of the opinion extracted from OSN user data is indeed reflective of the opinion of the larger online population. In this work we study this issue in the context of movie reviews on Twitter and compare the opinion of Twitter users with that of the online population of IMDb and Rotten Tomatoes. We introduce new metrics to show that the Twitter users can be characteristically different from general users, both in their rating and their relative preference for Oscar-nominated and non-nominated movies. Additionally, we investigate whether such data can truly predict a movie's box-office success.
Which means, if true, something we've all suspected. The world online and the world offline are two very different places.
- Comment (12)
- · Quote
- · UnfollowFollow (1)
Comments:
May '11
Re: Tweeters Are Different
The authors analyzed 10,000 tweets which I assume means they read 10,000 tweets. They have my sympathy if not respect.
Re: Tweeters Are Different
Reminds me of this about the shocking downfall of the show Community:
Community Engagement: How the Internet Ruined My Perception of What's Popular
Mar '11
Re: Tweeters Are Different
Actually, the study appears to show that one part of the on-line world is very different to another part, and probably intersecting to some extent, of the on-line world.
If that difference is measurably large, then how large must be the difference between both and offline?
Sep '10
Re: Tweeters Are Different
Rob, a small correction: the paper's authors are at Princeton. The arxiv.org site is managed by Cornell but anyone can upload a paper to it.
Mar '11
Re: Tweeters Are Different
I think this study says more about Twitter users versus other users of online media than it says about online users versus the offline population.
Edit: Thanks, Stephen. You made the point first.
Edited on April 3, 2012 at 2:16amMay '10
Re: Tweeters Are Different
I hope it wasn't a government-funded study.
Feb '12
Re: Tweeters Are Different
The only tweet I know. And he tweeted off too
He rocks in the tree-top all a day long
Hoppin' and a-boppin' and a-singin' the song
All the little birds on J-Bird St.
Love to hear the robin goin' tweet tweet tweet
[Chorus]
Rockin' robin (tweet tweet tweet)
Rockin' robin (tweet tweet tweet)
Oh rockin' robin well you really gonna rock tonight
Every little swallow, every chickadee
Every little bird in the tall oak tree
The wise old owl, the big black crow
Flapping them wings sayin' go bird go
May '10
Re: Tweeters Are Different
What about Ain't She Tweet (A'coming down the street)?
Feb '12
Re: Tweeters Are Different
EJ- I raise ya..Elvis with Send me the pillow that you tweet on.. Jo stafford doing The Teena-tweet waltz.. and then there's always Julie London with The more I tweet you..
May '10
Re: Tweeters Are Different
I've always been skeptical of the people who actually take time to write reviews on imdb and Rotten Tomatoes. I give little weight to their reviews, just like I don't pay much attention to Yelp reviews or debates on the "talk" pages of Wikipedia. All of those media are opportunities for self-absorbed people to establish a fiefdom where they have the authority they always wish they had.
I'm not saying all users of those sites are like this, but for my time it's just not worth the effort to try to distinguish them. You can never know which reviewers share you tastes. I trust the opinions of people I know in person.
Of course, I could have forgone writing all these words and just linked to this xkcd comic.
May '11
Re: Tweeters Are Different
Don't forget Rocky Top...Home Tweet Home to me.
Mar '11
Re: Tweeters Are Different
Self-selected samples are never representative, regardless of size. Even when researchers make good-faith efforts to fashion samples that may serve as accurate proxies for whatever population is being examined, there is almost always a significant margin of error.
If the purpose of the Online Social Networks study is to deepen our understanding of social networks, its methodology dooms it to failure. I’d be surprised if the percentage of IMDb and Rotten Tomatoes users who’ve ever written reviews exceeds single digits. What can be learned from comparing the responses of one self-selected sample to two other self-selected samples? Certainly nothing that can be said to be representative of the users of those sites, much less of movie goers in general.
However, if the purpose of the study is to generate some attention without too much effort, it may be a modest success.