From the comments on this page...
SEOmoz | How Reliable is Google Analytics?:

Any solution that doesn't use page tagging is going to overreport the number of visitor sessions and page views. This occurs because the log analyzer cannot distinguish between human and robot traffic. It also can get fooled by proxy servers that distribute load (something AOL has been known for).

When we installed Urchin Software for the first time with page tagging, we were shocked at the difference between the reported levels in Urchin and the previous (very old) version of Web Trends. The old software was not using any form of page tagging.

The great thing about our transition was that we were using Urchin Software instead of the ASP service (now called Google Analytics). Because Urchin Software referenced the log file, we were able to look through the log file on a very low traffic site and count by hand the cookies indicating individual visitor sessions. The numbers which we thought were wrong because they had dropped so much were actually accurate.

The bad thing about our transition was that we had to explain to our customers why their numbers dropped so much. It wasn't fun.

One thing you could do if you want to test this theory would be to modify your log files to write cookies to your logs. You should then be able to configure webalizer to ignore the cookies. Then you could grep the log files yourself to see if the numbers make sense.

Or you could seek an alternative hybrid solution. A hybrid solution is one like Urchin Software that combine page tagging with log analysis. Current versions of Web Trends, if set up correctly, offer this capability.

Regardless, the biggest difference you are going to see is between solutions that don't use page tagging and those that do. Those that don't use page tagging are going to be horribly inaccurate and overstate traffic. At this point, I don't trust any numbers reported by solutions that aren't doing page tagging."

and

"You'll get under-reported activity with page-tagging because approximately 10-12% of users have Javascript turned off (these numbers are provided with some cautionary statements). "

and

"Actually, you have somewhere around 11 to 15% under reporting with page-tagging. In addition to the javascript issues, you also have some users who block first-party cookies. If the page-tagging solution does not use first party cookies, the percentage of people blocking the cookies jumps substantially.

There is no accurate solution. Just ranges of accuracy. Page tagging is considerably more accurate. "

From...
The importance of the body of the web page.:

"The direct response business, you know, those companies that make infomercials and run classified ads in papers across the country, have studied and mastered the art of writing headlines. What they learned is that headlines are most effective when they accomplish 4 things:

1. Solve a problem
2. Solve that problem quickly
3. Solve that problem for what appears to be a small or reasonable amount of money
4. Make the reader curious to learn more..."

14% say http://www.w3schools.com/browsers/browsers_stats.asp
(based on a web techie user base)

17% says Mr Nielsen - http://www.useit.com/alertbox/screen_resolution.html

5% says Thomas Baekdal - http://www.baekdal.com/reports/actual-browser-sizes/abs-display/
(based on users of fashion sites mainly)