[starbase]

Not true, it depends on what information you are looking at. Let me use the thermometer example. You got a good one that reads 70 degrees duringthe day than drops to 65 degrees at night. Then you got bad one that is off by 10 degrees. It may read 60 degrees and drop to 55 on the same night as the good one, but they both accurately recorded the 5 degrees drop.

But that's a loaded example. A thermometer, unless it is truly wacky, is still made of mercury which must go up or down when heated or cooled. It may not be very accurate, but it will follow that trend every time.

But traffic-measuring tools like Alexa, that depend solely on the behavior of those who have it installed on their browser, could very well move in the opposite direction of a given trend. If the small subset of people with Alexa log off when a big rush of conventional users logs on, Alexa will say traffic is down when, in fact, it's up. Alexa could also say traffic is up when it's down, or that it's up when it's up, and also that it's down when it's down.

In short, I just don't think these measurements are dependable enough to even look at seriously. Any time we start to rely on all our data being 'equally flawed', then it's time to chuck the data out and find another way to measure our hypothesis, IMHO.

[McGavin999]

In short, I just don't think these measurements are dependable enough to even look at seriously. Any time we start to rely on all our data being 'equally flawed', then it's time to chuck the data out and find another way to measure our hypothesis, IMHO.

Those of us who have been here for a long time can tell just by the activity on the latest post page and by the number of posts on various threads. When you hit an "active" thread and find most of the posts saying things like "Did too!" and "Nuh uh, did not" then you recognize the problem.

[Bob J]

John R. used to have an FR traffic website. Don't know what happened to it.