Not Liking something is not the same as not “liking” something

Posted on | June 2, 2010 | No Comments

I’ve noticed a worrying trend among news agencies. It seems that it has become de rigeur to use the number of members of a facebook group as an indication of the popularity of a trend or movement.

For example, here is a quote from a recent CNN.com article discussing the “quit facebook” movement:

“More than 2,700 people had pledged to quit Facebook on the group’s website Monday morning…By contrast, more than seven times that many people are fans of former “Top Chef” contestant Kevin Gillespie’s beard. A fairly random user-created page called “I Love Facebook” had roughly the same number of members as the “Quit Facebook” page.”

Other major magazines are skipping the more accurate random poll and simply cite the number of facebook members a group has to indicate support. Business Week cherry picked a single “Boycott BP” group (out of many) to illustrate a ground swell of support.  Even ReadWriteWeb gets in on the fallacy of using Facebook group data to critique another inaccurate survey.

So where is this going?
Clearly the fallacy that CNN and other news agencies using the number of people in one  Facebook group to compare against another is that the nature of the “like” button is not taken into account. Facebook group counts suffer from “selection bias.”

Firstly, users are more likely to be exposed to causes for which they are already sympathetic too from friends with similar view points. The same person who signs up for Pro-Obama groups may not be exposed to Pro-Republican groups. Secondly, there is no “dislike” button; therefore many news agencies commit the fallacy of using this data while ignoring that a lack of members indicates a lack of support or that a fast growing  group indicates a groundswell of support. This is simply not true. Many users do not like anything; therefore their views will not be counted. If one out of every 3 people disagree with a cause, but it still gets one Million members, those 3 million go completely unmeasured. If only 1.4 million people Support our vets on Facebook, does that mean that the other 298.6 million automatically do not?

I think we have to be careful of using Facebook group information to tell us anything at all, and I think our news agencies have to be a little less hasty to use Facebook group data when a real survey of people’s opinion is what the story really needs.

Do more people want to quit Facebook than support Kevin Gillespie’s beard? Unfortunately, I guess we’ll never know.

Comments