For years, clients have been asking me for the ability to predict outcomes based on exsiting data. When I read this abstract, I thought that maybe all my dreams were going to come true. But, since I don't have a PhD in research, I checked it out with some of my friends who do. And sadly, we're along way from any magic bullets here.
Tina McCorkindale who has done some wonderful research into the Fortune 500's use of Twitter and Facebook (stay tuned for a synopsis of her latest) gave me this perspective:
I appreciate their efforts trying to predict sales based on sentiment, but I found so many flaws in the analysis. First, they generated 2.8 million tweets to analyze using a computer model . There’s all sorts of flaws in using computer analysis especially analyzing sentiment (not to mention the unreliability of even accurately searching for the movie titles). Next, their hypothesis that most of the tweets prior to release of the movie should be promotional and stop, followed by positive and negative tweets I disagree with as well because they didn’t even look at the content of the tweets, but a computer did. Plus, they didn’t have any research to back this up. Some people may say they are going to the movie or they heard it wasn’t very good (there’s all sorts of error entered). Also, the sample size is entirely too small (as far as the number of movies) and there could be other factors as well such as a high holiday time. I can’t imagine there being an enormous amount of tweets about Twilight (or even draw an analysis) because the movie is geared toward tweens and teens, who aren’t really even on Twitter. So I think for some of the movies, Twitter probably isn’t the best source because it’s not the movie’s target audience. I can also see drawing a parallel with the news media analysis, but it’s an entirely different ball of wax. It’s just flawed in the definitions, methods, analysis, and conclusion. I really liked the attempt and I think it’s a great way to measure, but I think they went about it all wrong and drew conclusions when they shouldn’t have.
Thanks Tina, I couldn't have said it better myself.