|
InvisibleBlack -> RE: Who Is Smarter Than a Fifth Grader (6/10/2010 6:33:24 PM)
|
quote:
ORIGINAL: FirmhandKY
So ... Zogby isn't a real, scientifically based polling organization?
Are is it that just all polls are faked up?
You'd have a better chance of convincing me if you show me how the poll violates the science of polling and statistics.
I'm not going to look otherwise, because my biases says it's accurate. [8D]
Firm
I read this article earlier this afternoon and then tracked down the blog of the scientist who performed the experiment and downloaded his original report. I have a couple of comments:
1) This test suffers from serious confirmation bias in two areas. The first is that all of the questions are ones that a "right-wing" person would agree with and that a "left-wing" person would disagree with. This obviously skews the results. A fair test would have been something along the lines of 4 questions that a "left-wing" person would disagree with, 4 that a "right-wing" person would disagree with and then 4 that the "average" person would disagree with. This would have eliminated someone's innate political bias from skewing the results. Secondly - he included "Don't know/not sure" as a "correct" answer. This tilts the results to being ignorant of the answer and knowing that you're ignorant as being "not incorrect". A better option would have been to discard the "don't know" results from either being viewed as "correct" or "incorrect".
2) The designer of the test is, himself, obviously biased. A "pure" test would have had him come up with the concept, someone else design the test and then a third person, who either did not know what the test was about or had noleanings either way conduct the test and correlate the answers.
3) He goes out of his way to adjust for educational bias (in fact, there's an entire section on how the level of education does not correlate with how one answers the questions but political leanings hugely correlate) but he also gathers an enormous amount of personal information (gender, marital status, fear of death (I'm not kidding) and dozens of other things) and I would be curious to see if there's any correlation of any other area.
4) All of that being said, I truly appreciate his work. He included all of his raw data. You can download the spreadsheet with all his results, including the questions he threw out. You can run your own analysis, include or exclude any questions and develop your own correlation matrix. That's a scientist.That's someone who's looking for valid criticism and feedback. That's so damn rare these days that it's wonderful to see.
5) From a purely anecdotal standpoint - he may be right. Based solely on the people I know with degrees in economics or working in the field of economics, a greater number of them are right-leaning / conservative / libertarian than I tend to see in the general population. Whether this is because more right-leaning people seek careers in economics, or whether being trained in economics tends to skew one more towards the right-wing would be the question that would need to be investigated as correlation does not equal causality.
|
|
|
|
