?

Log in

No account? Create an account

Previous Entry | Next Entry

Account Stats Pt. 2

I wanted to know the various statistics regarding account types on LJ, but unfortunately, they don't make the numbers public anymore. So, I wrote a perl script to crawl userinfo pages and grab the account types of a random selection of them. Here are the results.

Population size: 7483488
Sample size: 4164

Confidence level: 99%
Confidence interval: ±2%

Account TypeNumber in SamplePercent of SampleApproximate Number in Population
Free Account369488.71%6638602
Paid Account4159.97%746103
Permanent Account310.74%55377
Early Adopter150.36%26940
Paid Account
previously an Early Adopter
70.17%12722
Permanent Account
previously an Early Adopter
20.05%3742

Comments

( 5 comments — Leave a comment )
ydna
Jun. 21st, 2005 03:30 am (UTC)
Outstanding!

But what was your randomization method? (i.e., what are the biases in your sample)
rfreebern
Jun. 21st, 2005 11:17 am (UTC)
I seeded with shadesong, jwz, and vaginapagina, scraped every username from every page I saw, and randomly chose approximately 1/5th of all the distinct usernames seen.
two_star
Jun. 21st, 2005 04:08 am (UTC)
Hmm. I'd expect crawling userinfo would bias you toward people with larger numbers of friends-of, who are probably more likely than the general population to have paid accounts. Did you have some way of correcting for this?
ydna
Jun. 21st, 2005 05:56 am (UTC)
Unless he used the Latest Posts or Random Journal links as sources or occasional seeds. Of course, he's being very hush-hush about his methodology so far.

I smell a conspiracy!

(just kidding)
rfreebern
Jun. 21st, 2005 11:22 am (UTC)
Not really, no.
( 5 comments — Leave a comment )

Latest Month

February 2011
S M T W T F S
  12345
6789101112
13141516171819
20212223242526
2728     
Powered by LiveJournal.com