Clarity Strategic^TM

There's a "study" out there that is sweeping the media. I first saw it on Mashable, and in the last several days several people I consider to be thought-leaders have been quoting it and passing it on. It was conducted by one Soraya Mehdizadeh, a Psychology grad at York University in Toronto. The major result was a strong correlation between low self-esteem and high Facebook activity.

I give Mehdizadeh high marks for not overestimating the value of this study, but the rest of the social media world gets a failing grade for jumping on the result at though it means something.

Because it doesn't. The study is so flawed - with respect to what we think it is telling us - as to amount to little better than a book report. It reminds me of all those years teaching rhetoric, when students would try to back up a claim with what they called a "random survey," when all they did was run up and down the hall of their dorm floor and poll the residents.

And that's what was done here. 175 million subscribers to Facebook, which is several years (a decade?) removed from being primarily a social platform for college students, and the "research" was conducted on 100 (York) college students. Which would be the population of Mehdizadeh's dorm if she were living in one. Which probably IS the approximate number of students in York's psychology program. And Mehdizdadeh actually calls the sample "random," which is true only in a fever dream.

The students used were between the ages of 18 and 25. That age range is responsible for maybe 10 to 15 of my 650 Facebook friends - including my two kids. I'm over 50 - most of my friends are over 40. Hey, I admit there's a bit of difference in attitudes between someone 18 and someone 25. Are they REALLY the same as the differences between someone 18 and someone 45? Does Facebook FORCE us into the same behaviors as young people? I say no. Do I really have to convince readers that age (which the "study" essentially ignores) is at LEAST as strong a factor in determining social media behaviors as is gender (which is considered)? Really?

There are other schools in the Toronto area besides York, so it's possible Mehdizdadeh gathered data from other campuses. But I'd bet a Barberton chicken dinner that she didn't consider one single data point from outside of her region. So now we're supposed to suppose that social media behavior is the same in Toronto as it is (say) in Melville, Saskatchewan? Never mind in locations where the language is different, the climate is different, the Internet connections are different (for crying out loud!). Really?

And I'm not even addressing such issues as race, economic background, actual computer literacy, grade point average, campus activity, blah blah blah.

Here are the facts:

• The sample in the study isn't "random." A random sample reflects the structure of the actual audience. The sample in the study is one of convenience to the researcher only. 
• There's no "blocking" - some effort to eliminate sources of variation that don't matter. So we're not even sure what they are. Our sources of variation are the students' gender, and that's it.

Hey, Mehdizdadeh did what she was supposed to: developed and took the survey, processed the statistics (she's probably good as SAS then), came to a cautious conclusion and suggested more work. That last is something research scientists have to do, you know, to justify continued funding. She did her job. It's not her fault that the result has little more real meaning than any undergrad's research paper.

No, THAT is OUR fault. WE are the ones who made a big deal out of something that's so totally not. Maybe we - all of us active in the social media - all know in our hearts already that we're egomaniacs. Maybe this "study" justifies to us what we know about ourselves. But I for one think it doesn't. It's a bit of fluff, and we're throwing a lot of attention at it.

Can I go back to "Can I Has Cheezburger" now?