October 06, 2005
Shirking the Edelman and Technorati blogger survey analysis
UPDATE (October 8, 2005): For those of you still following this one, Phil Gomes has released some more analysis, based on the raw data.UPDATE (October 7, 2005): Wow. For a much more salient, intelligent, and engaging treatment of the questionable nature of the Edelman and Technorati survey and any analysis derived from it, I highly suggest you read Constantin Basturea’s posting on the topic. Stop reading this. Go there. Now.
ORIGINAL POST
So there’s a great deal of analysis swarming around the Edelman and Technorati blogger survey I posted earlier this week.
Now, I’m the first to admit: I’m not a statistical wiz. But here’s the one thing that sticks in my craw about this survey and any derivative analysis thereof: 821. That’s eight hundred twenty one survey responses. 821. And you can round it to 820 for all I care, because my responses are no better than any of the things I write here.
820 bloggers.
In fact, there are, no doubt, a small army of folks like me, who when approached with a survey begin bouncing and chattering excitedly like a cage full of howler monkeys, thinking that someone, anyone, is actually interested in their opinion. Is our opinion valid? No probably not. I may have just answered the survey to get the results. Who knows?
So let’s cut all those folks. Let’s say that there are about 20 people in addition to me who were completely useless respondents.
800 bloggers.
Now, let’s account for the bold faced liars, the wannabe influencers, and those people who take the Cosmopolitan personality quizzes with the intention of getting the answer they want. That’s about 50 people.
750 bloggers.
Okay, that’s enough cutting of the survey respondents. Why? Because I said so. I could whittle it down to about 25 without even trying, so let’s move on.
How does that number stack up against the number of active bloggers out there in the world of blogging? Technorati, one of the survey sponsors, claims to be tracking upwards of 18.9 million blogs. 18 million. I heard something the other day along the order of 80,000 blogs created per day. That’s eighty thousand, kids.
For a point of reference, I’m currently tracking about 350 very active bloggers in my Bloglines account. So, roughly half of our survey responses could be people I read. Hmmm. Alas, I digress.
Now, granted, a huge chunk of those exponential blog numbers are the spammers gumming up services like Blogger and Technorati. I’ll give you that. So let’s cut the Technorati real number to, oh, say 9 million. Cut it in half, right off the top.
9 million. Nine with six zeros. Got it?
Now, let’s all look deep inside ourselves and ask, “Am I really a serious blogger?”, “Do I blog as often as I should?”, “Do people remember that I wore this shirt on Monday?”, “Am I an active blogger?”
You, you, and you. Liars. Get out. You and you. You like to think you’re active, but once a month does not mean active. The rest of you, I’m willing to take at your word. Especially those of you who said, “No.”
So let’s cut it to a third.
3 million bloggers.
Hmm. 750 respondents divided by 3 million potential respondents. That’s about 0.025% of active bloggers who actually felt like wasting the time to respond to the call to provide feedback. That is, they didn’t have anything better to do when the request arrived. Or they had bloggers’ block. Or something. Whatever the case, I don’t think that cuts it. In fact, I think it misses by a factor of 4.
So I had planned to do some analysis. But I’ve got nothing. Nothing with which to work, I mean. It’s obvious that I can talk for hours about nothing. It’s not the nothing that’s stopping me.
The best take I’ve seen on trying to extract something relevant out of this data is Phil Gomes swag on so-called “authority bloggers.” And Steve Rubel has a good take on what would be more helpful, using the results as a launching point.
Long story short: It’s not statistically relevant. And there’s only one thing worse than poorly analyzed statistics. No, it’s not my writing style, gentle reader. It’s statistically irrelevant data being proffered as evidence.
What we got, was a hook for a press release and a way for generating buzz. But we didn’t really get any good data. So toss it.
Think I’m wrong? 24.7% of the readers I surveyed generally do.
Shirking the Edelman and Technorati blogger survey analysis
| | Subscribe:
RSS, Bloglines
