QUOTE (Balta1701 @ Aug 23, 2011 -> 05:10 PM)
The standard deviations presented in Figure 1 are on the order of ~1.0 points in IQ or less, while showing 10 points of spread, and the very liberal and very conservative groups are at least 5 error bars apart. Are you bringing outside info here or is there something in the article I'm missing suggesting that the actual error bars in that figure should be 5x larger than presented?
The IQ test is relatively unique because it has been used so extensively that we know the range of scores. Basically, we (psychologists that are not me) know that the test itself has a 15 point standard deviation. So, if we gave the test to everyone in the country 68% would score between 85-115 and 96% between 70-130. (This is actually how the range for mental retardation and "gifted" were determined.) Of course this is only true for individual scores, group means have a much smaller likely range. (I used to know this, but it's crazy rare to have a mean be more than 5 or so points away from 100.) Basically, I find it unlikely that this was due to anything other than chance. Maybe if I read the original peer reviewed article, I would be less skeptical. But this seems to prey on people that don't know a ton about the IQ distribution, statistics, critical thinking, or how not to piss me off.
My question was more if those are standard deviation, standard error, or some other type of confidence range.
Plus, it's psychology today. Weekly World News : Reality :: Psychology Today : Psychology.