When Statistics Tell You Nothing

According to eMarketer, as of 2008, 54% of Internet users read at least one blog one time per month, and they project that number to reach 67% by 2012. In addition, the number of bloggers will grow steadily between now and then (although the number is still very small). At first glance, this development seems to be very promising to the blogging community, but in reality, it doesn't tell us much at all.

The numbers are likely inflated, because the definition is too broad.

You have to read only one blog post once per month to qualify as a blog reader?

You have to update our blog once in the past three months to qualify as a blogger?

When analyzing and comparing statistics, one of the first things you should do is find out how the statisticians defined their terms. Otherwise, you leave yourself open to manipulation by those who, blinded by their own passion, see only that which advances their own beliefs. - Cam Beck

Comments

[Alan Wolk (The Toad Stool)]

Exactly Cam.

I'm always leery of survey type stats because they are so open to nuance and inflection.

To wit: "Is Cam Beck one of your favorite bloggers?" will get a very different response than "Do you enjoy reading Cam Beck's blog?" - even though they are essentially asking the same question. And if I wanted stats to support a proposition, I can just phrase the question that way.

Particularly apt to note during election season, when that sort of thing is rife.