Your article is good but flawed.
Survery results and "unbiased studies" are inherently flawed.
You can skew the results to get the numbers you want.
Here is an example.
Lets say your conducting a poll on Social Security and whether or not it should be privatized. Thus eliminating the government burden and potentially altering the amount of income retireee's would receive. Some would receive less some would receive more. (Please not this is just a exmaple I don't know what would happen if Social Security is privatized).
YOu call up people and ask them to participate in a survey. However you only gather stats from retireee's.
You start off simoply asking basic infromation this is to determine if the interveiwee is a suitable candidate for your study and will answer in the way you want when you ask the big question for your "unbiased study."
In the course of your basic questions you determine that the retiree only recevies Social Security as a means on income.
Natually if you ask the question.
"Are you in favor of privatization of Socail Security if it means you would receive less money?" They would say no.
Alternatively if you asked
"Are you in favor of privatization of Social Security if it means you would receive more more money?" They would answer yes.
Another problem is in all these surverys when relasing data they don't say the questions that were asked they just quote the statics. "85% of Americans are in favor of Social Security reform."
But they don't say they only interviewed retiree's.
Kinda silly how people take these unbiased studies as completely infallable.