The idea that a lead this great is a “statistical dead heat” is as false as it is true. That is, the margin of error that is “plus or minus four points” refers to a an interval of two “standard deviations.” People talk of two standard deviations as a statistically significant margin because there is a 95 cases out of a hundred — 95% of the time — the results will fall within the predicted range. However, a single standard deviation — two points in this poll — is correct 50% of the time. So there is a 50% chance that Bush is at least two points ahead of Kerry, and essentially no chance that Kerry is ahead of Bush.
Of course, I’m just a lowly corporate tool, so if any of you statisticians out there think I’m wrong and want to bitch slap me, whack away.
I think the DNC has planted a story on these polls and Drudge is running with it here.
PaulOctober 30, 2004
Dan I commented on your blog but I’ll paste it here as well.
———————
Dude, you completely missed what was going on. Drudge indicated it was a Dem polling agency. If you look at the bottom of the press release it says “Paid for by Kerry/Edwards.”
It is a political hack poll that is indeed intended to “Keep hope alive” but it did not require your BS detector, it is blatant.
Hey gang, I was at the Republican headquarters in Hackensack, NJ working the telephones today for the Bush campaign(forget NYC, I working in Jersey because Bush has a shot in that state). Guess who paid a surprise visit? The Bush daughters. They are so cool.
thatcoloredfellaOctober 31, 2004
I love the selective use of polling utilized! It’s like Fox on the Internet!
Where are Bush personal numbers? Right/Wrong Track? Independents? Battleground states? You know the significance of the incumbent’s 50% percent bar?
This election is a referendum on the GWOT. We know where voters stand on that.
There’s a good chance Bush will take IA, WI, and MN redering Ohio unimportant. If you ignore the Trib poll and take Zogby with a grain of salt, Bush looks very good in those states.
– Erm…Paul….nothing to click?
oh geez! same mistake 2 times in 2 days. fixed.
must… get… sleep…
The idea that a lead this great is a “statistical dead heat” is as false as it is true. That is, the margin of error that is “plus or minus four points” refers to a an interval of two “standard deviations.” People talk of two standard deviations as a statistically significant margin because there is a 95 cases out of a hundred — 95% of the time — the results will fall within the predicted range. However, a single standard deviation — two points in this poll — is correct 50% of the time. So there is a 50% chance that Bush is at least two points ahead of Kerry, and essentially no chance that Kerry is ahead of Bush.
Of course, I’m just a lowly corporate tool, so if any of you statisticians out there think I’m wrong and want to bitch slap me, whack away.
I think the DNC has planted a story on these polls and Drudge is running with it here.
Dan I commented on your blog but I’ll paste it here as well.
———————
Dude, you completely missed what was going on. Drudge indicated it was a Dem polling agency. If you look at the bottom of the press release it says “Paid for by Kerry/Edwards.”
It is a political hack poll that is indeed intended to “Keep hope alive” but it did not require your BS detector, it is blatant.
Hey gang, I was at the Republican headquarters in Hackensack, NJ working the telephones today for the Bush campaign(forget NYC, I working in Jersey because Bush has a shot in that state). Guess who paid a surprise visit? The Bush daughters. They are so cool.
I love the selective use of polling utilized! It’s like Fox on the Internet!
Where are Bush personal numbers? Right/Wrong Track? Independents? Battleground states? You know the significance of the incumbent’s 50% percent bar?
You know the significance of the incumbent’s 50% percent bar?
We know what the pundits tell us the significance is — but them folks don’t have much credibility anymore.
I still say Bush 52-47.
This election is a referendum on the GWOT. We know where voters stand on that.
There’s a good chance Bush will take IA, WI, and MN redering Ohio unimportant. If you ignore the Trib poll and take Zogby with a grain of salt, Bush looks very good in those states.