A good question which bring out the random sampling hypothesis testing
maniac in me.
They site statistics could be a bit biased if you get lots of hits from
jaws users or something in particular that has resources for one
operating system or browser like a Mac operating resources site stats.
Statistically though, to sample the whole world with some accuracy you
only need a moderate size sample if the sample represents the wider
population.
I would not trust the stats even of a single search engine to be
unbiased and without commercial interests in selling clicks. Could we
trust any one vendor or researcher to report unbiased results from one
or a few sites?
You need to sample generalist site, with wide appeal, indexed in all
search engines wide appeal Widgets like History Science Religion
Medicine Psychology, pictures, law and order etc etc. I only get a
couple of thousand hits a month (I have banned almost all bots who did
not follow robots.txt on my site, but most bad bots might be User Agent
X who I have excluded. I must check that.
My site seems biased towards Mac hits, a publishing orientation maybe?
I'm sure many others would show different stats. There must be a sample
possible of members stats, perhaps 40 (the more sites sampled the
better) over different months of the year, then we could be sure of
browser stats.
Which group might be best placed to conduct such research:-)
Don't look at me, but I could design the basics of such a study.
If you pick generalist sites with wide appeal, you have millions of
hits to base your conclusion on that your sample is a true reflection
of world-wide browser stats.
Tim
http://www.hereticpress.com
On 01/08/2006, at 10:00 PM, Paul Collins wrote:
Hi all,
Just wondered if anyone has a good resource for Browser stats.
Currently I've got a few but most get their stats from visitors to the
site which can be a bit biased.
Currently I've got
http://www.upsdell.com/BrowserNews/stat.htm
http://www.thecounter.com/stats/2006/July/browser.php
Anyone got better?!
Cheers
******************************************************
The discussion list for http://webstandardsgroup.org/
See http://webstandardsgroup.org/mail/guidelines.cfm
for some hints on posting to the list & getting help
******************************************************
The Editor
Heretic Press
http://www.hereticpress.com
Email [EMAIL PROTECTED]
******************************************************
The discussion list for http://webstandardsgroup.org/
See http://webstandardsgroup.org/mail/guidelines.cfm
for some hints on posting to the list & getting help
******************************************************
