Petition to keep non-instanced dungeons.

[LemmingKingXXX]

And Nielsen ratings only account for 0.02183% of American viewers. But it works because they use sample sizes.

Carefully, meticulously selected sample sizes that strive to capture as much of a complete picture as possible of the population as a whole without actually polling 330 million people. I don't recall ever hearing Rokkien talk about his statistics training or how he took multiple census, analyzed them for demographic data and chose a select %1 of the population, asking each of them to read the thread and "like" if they agreed with his point of view.
Sample sizes are only worthwhile when they are accurately created, and even when that is the case they can contain at least some error, that's why all the formulas governing samples have mathematical "corrections" in place to help minimize and account for that ever present error.

[Sephrick]

Carefully, meticulously selected sample sizes that strive to capture as much of a complete picture as possible of the population as a whole without actually polling 330 million people. I don't recall ever hearing Rokkien talk about his statistics training or how he took multiple census, analyzed them for demographic data and chose a select %1 of the population, asking each of them to read the thread and "like" if they agreed with his point of view.
Sample sizes are only worthwhile when they are accurately created, and even when that is the case they can contain at least some error, that's why all the formulas governing samples have mathematical "corrections" in place to help minimize and account for that ever present error.

Not to mention he didn't even read the wiki article from which he pulled that percentage, because it was from the section criticizing the flaws of on of Nielsens many varied methods of collecting data.

And yet in 2009 of the 114,500,000 U.S. television households[14] only 25,000 total American households (0.02183% of the total) participated in the Nielsen daily metered system

0.02183% isn't the total they poll. Only the amount that bothered to participate in an experimental new system.