Comment Re:Statistical Noise and Statistical Bias (Score 1) 557
The report could be dismissed as statistical noise if the slight decline weren't also reflected at sites like w3schools.
As we all know, surveys are no more reliable than their sampling procedures and reliable surveys are not necessarily valid.
For the time being, I'm inclined to accept the argument for noise, but there are also many anti-firefox biases in standard browser counting methods.
Most are relatively consistent over time, but there is one obvious one which increases with firefox use.
With experience, Firefox users increasingly find and deploy extensions which block the images and scripts that research companies use to count user behaviour.
I try to avoid blocking the more obvious ones, but it's far from easy to distinguish them from spam-minded market research.
As we all know, surveys are no more reliable than their sampling procedures and reliable surveys are not necessarily valid.
For the time being, I'm inclined to accept the argument for noise, but there are also many anti-firefox biases in standard browser counting methods.
Most are relatively consistent over time, but there is one obvious one which increases with firefox use.
With experience, Firefox users increasingly find and deploy extensions which block the images and scripts that research companies use to count user behaviour.
I try to avoid blocking the more obvious ones, but it's far from easy to distinguish them from spam-minded market research.
