techPresident - Does Facebook Support Generate Votes? [UPDATE] - Comments

You are correct about correlation vs. causation

jgulati — Fri, 31 Aug 2007 23:42:25 -0400

Patrick: You are correct to point out the difference between correlation and causation, and we make that point very clear in our paper. A candidate did not even have to access their profile to gain supporters. But even after controlling for other factors that we know influences vote shares (i.e., partisan advantage in the district, voting record, money raised, quality of challenger, and scandal), we still find that Facebook support has a strong postive effect on the final vote. Thus, we argue that something is going on here, and as you already have mentioned, we think that it is capturing something else that is going on in the campaign. Maybe that "something else" is underlying enthusiasm for the candidate, intensity of support, or organizational strenght of the campaign. We'd appreciate any additional thoughts you might have on this.

Jeff Gulati

Link to the excutive summary of the study

jgulati — Fri, 31 Aug 2007 23:31:36 -0400

Thanks Josh for posting the url for the complete paper. I have posted the text of the 3-page executive summary of the paper on a temporary blog site that I hastily created. It's a pretty sad effort, but at least it is readalbe. If you want a cleaner version of the summary, please email me, and I'll get one out to you as quickly as possible.

http://www.bentley.edu/academics_research/faculty_research/faculty_datab...

Jeff Gulati

Download the paper

Joshua Levy — Fri, 31 Aug 2007 15:20:27 -0400

You can download a PDF of the entire paper here.

The study isn't online

Joshua Levy — Thu, 30 Aug 2007 23:07:42 -0400

Unfortunately the study isn't online, though I recommended to Jeff that he upload it somewhere. I'm sure we can arrange for him to send it to interested people, and I'll make a note of it if he makes it public.

This issue of correlation vs. causation is indeed tricky, which is why Jeff made it clear that they don't believe causality is a part of it. Their numbers need to be dug into and analyzed more before we conclude anything one way or the other, but in any event I was surprised that the study raised even the possibility of a connection, which is more than we've seen thus far.

+1

Fred Stutzman — Thu, 30 Aug 2007 18:11:21 -0400

I'd have to agree here. It wouldn't surprise me that greater voter support correlates to greater online interest, but I doubt there's any explanatory power in the actual number of friends.

It might be interesting to build a model where new media metrics (FB friends, blog reach, Youtube views) are factored in with more traditional metrics (fundraising, polling, etc) to see if there's anything interesting there.

Link to study?

bradlevinson — Thu, 30 Aug 2007 17:04:40 -0400

Hey Josh,

Would you mind sharing the link to the study? Would love to look at it.

Thanks!

Correlation vs. causation

Patrick Ruffini — Thu, 30 Aug 2007 16:55:58 -0400

I think there may be some confusing of the two going on here.

I love Facebook, but I can't make the argument with a straight face that 500 more friends = 10,000 more votes or whatever the number is.

More than likely that Facebook support is a reflection of broader overall support.

Does Facebook Support Generate Votes? [UPDATE]

Joshua Levy — Thu, 30 Aug 2007 16:45:24 -0400

When we talk about social networking apps and elections, the question often arises of how online support translates into votes. Now a new study has, for the first time, linked online support on Facebook to actual votes.