2008 US Presidential Election. Comments http://2008uselection.blogsome.com The United States Presidential election of 2008 will be held on November 4, 2008 Sun, 22 Nov 2009 13:00:22 +0000 http://wordpress.org/?v=1.5.1-alpha by: Nathan W. http://2008uselection.blogsome.com/2007/02/09/gore-clinton-lead-romney-by-double-digits/#comment-5 Fri, 09 Feb 2007 16:09:27 +0000 http://2008uselection.blogsome.com/2007/02/09/gore-clinton-lead-romney-by-double-digits/#comment-5 For further understanding of polls see: <a href="http://www.sfgate.com/cgi-bin/article.cgi?file=/chronicle/archive/2004/10/10/EDGMV952LR1.DTL" rel="nofollow">http://www.sfgate.com/cgi-bin/article.cgi?file=/chronicle/archive/2004/10/10/EDGMV952LR1.DTL</a> <blockquote>Based on interviews, outside reading and my two attempts to pass a Statistics 101 class, I suggest that you look for at least the following: Margin of error: When a story tells you that Thomas Dewey is leading Harry Truman by a 52-48 margin, ask yourself: What's the margin of error ("confidence interval" in stats-speak)? If the paper doesn't answer the question, the numbers are useless. A 4 percentage-point margin of error, for example, means that Dewey could be as much as 8 points ahead. But he also might be in a dead heat. Sample size: The story should tell you the size of the sample, which can relate to the accuracy of the poll. As the experts say, you don't have to eat a whole pot of soup to know what it tastes like. A spoonful will do, so long as you stir the pot. So it is with polls. A truly random sample, sufficiently big, will give statistically reliable results. Pollsters can scientifically sample fewer than 1,000 people nationwide and be 95 percent sure their results are accurate to plus or minus 3 or 4 percentage points. </blockquote> For further understanding of polls see:

http://www.sfgate.com/cgi-bin/article.cgi?file=/chronicle/archive/2004/10/10/EDGMV952LR1.DTL

Based on interviews, outside reading and my two attempts to pass a Statistics 101 class, I suggest that you look for at least the following:

Margin of error: When a story tells you that Thomas Dewey is leading Harry Truman by a 52-48 margin, ask yourself: What’s the margin of error (”confidence interval” in stats-speak)? If the paper doesn’t answer the question, the numbers are useless. A 4 percentage-point margin of error, for example, means that Dewey could be as much as 8 points ahead. But he also might be in a dead heat.

Sample size: The story should tell you the size of the sample, which can relate to the accuracy of the poll. As the experts say, you don’t have to eat a whole pot of soup to know what it tastes like. A spoonful will do, so long as you stir the pot. So it is with polls. A truly random sample, sufficiently big, will give statistically reliable results. Pollsters can scientifically sample fewer than 1,000 people nationwide and be 95 percent sure their results are accurate to plus or minus 3 or 4 percentage points.

]]> by: Nathan W. http://2008uselection.blogsome.com/2007/02/09/gore-clinton-lead-romney-by-double-digits/#comment-4 Fri, 09 Feb 2007 16:01:29 +0000 http://2008uselection.blogsome.com/2007/02/09/gore-clinton-lead-romney-by-double-digits/#comment-4 Wow, I was wrong! Which is even better news for Romney than I thought. The Rasmussen poll has an accuracy of + or - 4% points, by contacting just 800 random people in the US. Simple Statistics 101 says that due to the accuracy of the poll, +4% to Romney and -4% to Hillary that they could be in a dead heat as of last month. It's always fun to realize how inaccurate polls are :P Wow, I was wrong! Which is even better news for Romney than I thought. The Rasmussen poll has an accuracy of + or - 4% points, by contacting just 800 random people in the US. Simple Statistics 101 says that due to the accuracy of the poll, +4% to Romney and -4% to Hillary that they could be in a dead heat as of last month. It’s always fun to realize how inaccurate polls are :P

]]> by: Nathan W. http://2008uselection.blogsome.com/2007/02/09/gore-clinton-lead-romney-by-double-digits/#comment-3 Fri, 09 Feb 2007 15:52:25 +0000 http://2008uselection.blogsome.com/2007/02/09/gore-clinton-lead-romney-by-double-digits/#comment-3 Wow, this is amazing! This is better news for Romney than I thought. Ironically, you do not state the accuracy of the Rasmussen poll. Most Rasmussen polls spend the required amount to get to the + or - 3% points, by contacting over 1000 random people in the US. The moves in the poll you site: Gore (who is not running yet) beats Romney by 11 points last month and 14 points this month. That slide is within the accuracy of 3% points for the poll, thus it's a wash. Hillary's numbers are less impressive, 8 points a month ago and 11 points this month, also a wash due to poll accuracy. The most telling aspect of this poll is that a well known Vice President and a well known First Lady are within striking distance to an obscure Governor from Massachusetts. Now that's amazing. Wow, this is amazing! This is better news for Romney than I thought. Ironically, you do not state the accuracy of the Rasmussen poll. Most Rasmussen polls spend the required amount to get to the + or - 3% points, by contacting over 1000 random people in the US. The moves in the poll you site: Gore (who is not running yet) beats Romney by 11 points last month and 14 points this month. That slide is within the accuracy of 3% points for the poll, thus it’s a wash. Hillary’s numbers are less impressive, 8 points a month ago and 11 points this month, also a wash due to poll accuracy.

The most telling aspect of this poll is that a well known Vice President and a well known First Lady are within striking distance to an obscure Governor from Massachusetts. Now that’s amazing.

]]> by: Brandon http://2008uselection.blogsome.com/2007/02/08/us-polls-boost-clinton-obama/#comment-2 Fri, 09 Feb 2007 00:58:46 +0000 http://2008uselection.blogsome.com/2007/02/08/us-polls-boost-clinton-obama/#comment-2 The 92% is encouraging, should be 100. I wonder what it would be for a minority candidate? The 92% is encouraging, should be 100. I wonder what it would be for a minority candidate?

]]>