« General Election Cattle Call, September 15 | Main | Final State Polls from the 2000 Election »
Friday, October 15, 2004
Who Is Insider Advantage, and What Do They Know About Florida?
Posted by DavidNYCInsider Advantage is a polling firm I've heard very little about, yet they've popped up with not one but two Florida polls in just three days' time. Their website doesn't offer very much - I'm getting everything from Race2004.net, which is in turn getting its info from the National Journal.
Anyhow, in a poll that concluded on October 12, Bush was up 47-44. Just two days later, they had Kerry up 48-44. A seven-point swing in just two days seems quite improbable to me, though one major event (the final debate) did take place during that time. Both of these polls apparently have 5% margins of error, which is bigger than any other pollster's margin in any Florida poll, ever.
What do you think?
P.S. A totally unrelated topic: If you have subscribed to the Bloglet e-mail updates for this site, please let me know if you've been receiving them lately. Thanks.
Posted at 02:50 PM in Florida | Technorati
Comments
Here's a poll that might indicate a trend:
http://money.cnn.com/2004/10/15/news/economy/confidence/index.htm
Posted by: Marsden at October 15, 2004 03:35 PM | Permalink | Edit Comment | Delete Comment
Here is the bio of Insider Advantage's co-founder:
Matthew Towery, co-founder of InsiderAdvantage.com, now serves as Chairman. Matt served in the Georgia Legislature for six years and was his party's nominee for Lieutenant Governor of Georgia in 1990. He served as campaign chairman for U.S. Speaker of the House Newt Gingrich from 1992 to 1998, and has worked as an attorney with the Atlanta/Washington-based law firm Long, Aldridge, & Norman. Matt owned, operated, and sold Color Graphics, one of the Southeast's largest sheet-fed printing companies. He has written several books, currently serves as the chief political analyst for a major NBC news affiliate in Atlanta, and is a brilliant governmental affairs and campaign strategist.
And a quick glance at its most prominent board member:
The InsiderAdvantage.com Board Of Advisors
InsiderAdvantage has established a Board of Advisors, made up of established government and business leaders, to provide guidance and counsel to the management team. The Board of Advisors is committed not only to advising the management team on critical business and political issues, but also to increasing awareness of InsiderAdvantage.com through established networks in their communities. The InsiderAdvantage Board of Advisors includes the following leaders:
* Newt Gingrich, Senior Advisor, CEO of the Gingrich Group and former U.S. Speaker of the House of Representatives.
Posted by: rsmart at October 15, 2004 03:43 PM | Permalink | Edit Comment | Delete Comment
Here's their web site.
http://www.insideradvantage.com/
I remember reading abotut them before. They're a firm founded by a Dem and a Republican for polling and research. No idea how reliable they are.
Posted by: erg at October 15, 2004 03:44 PM | Permalink | Edit Comment | Delete Comment
To know what to make of the poll, I think we need to compare it to the other poll out today, from Strategic Vision. Since they are a partisan pollster, and a very consistent one, I've figured out that their bias averages a consistent net +4 for Bush. So if you adjust them accordingly (subtract 2 from Bush, add 2 to Kerry), then Strategic Vision's numbers for Florida are a tie at 47%.
So my opinion is that they are very close to being right about Kerry's number, but are understating Bush's support.
Plus, with the 5 point moe, you can show Bush -3 and Kerry +4 without there being any change in the "true" support for the candidate.
Perhaps the best thing to do is to take the two numbers and average them.
Posted by: KariQ at October 15, 2004 03:57 PM | Permalink | Edit Comment | Delete Comment
I dont know Im standing up for IA, but could this be a lingering effect of the hurricanes. Polling might be weird in the Sunshine state.
Posted by: J at October 15, 2004 04:05 PM | Permalink | Edit Comment | Delete Comment
This could just be sampling error. Comparing the two Insider Advantage polls, the change for Kerry's percentage is 4 points. When comparing two polls like this, the cumulated margin of error in measuring a bounce is 1.41*MoE. In this case that's 7 points.
Crunching some numbers, The probability that Kerry gained (anything ranging from no change to a big bounce) is about 80%. To my mind a 4-point bounce is not out of the question. Unfortunately, this is the only firm that has data both before and after. Give it a few days.
Posted by: Sam Wang at October 15, 2004 04:06 PM | Permalink | Edit Comment | Delete Comment
My guess is that maybe Kerry picked up 1 or 2 points on Bush -- due to some combination of the debates and possibly some post-hurricane frustration about dealing with the govt -- and the rest of the 7-point swing due to sampling error. [Remember that a 5-point margin of error means a swing in the difference of as much as TEN points is only borderline significant.]
Posted by: Jason at October 15, 2004 04:27 PM | Permalink | Edit Comment | Delete Comment
Jason is almost correct. If the margin of error is 5 points, then a difference of 5 points is 84% likely to report the direction correctly. A difference of 10 points is 95% likely to report the direction correctly.
When comparing two polls taken on different dates, the interpretation is more complicated - see my post above.
Posted by: Sam Wang at October 15, 2004 05:26 PM | Permalink | Edit Comment | Delete Comment
I have been recieving the updates.
Posted by: David Trinh at October 15, 2004 06:24 PM | Permalink | Edit Comment | Delete Comment
ANY poll with a 5% margin of error is extremely suspect--regardless of whose poll it is.
Posted by: Suzanne DeBolt at October 15, 2004 11:12 PM | Permalink | Edit Comment | Delete Comment
Don't know who these guys are but I wouldn't
worry too much about the 5% margin of error
as it could be attributable to their being
more honest about it than others.
Looking at the big swings in the Gallup
polls, for example, makes it clear the
of error for them is much higher than
what they claim.
Posted by: The Other Rob at October 16, 2004 10:26 AM | Permalink | Edit Comment | Delete Comment
I have three thoughts about Fla, two hurricane related.
1) Maybe a lot of Floridians, due to one hurricane or another, lacked an opportunity to really pay attention to the previous debates. The third one, however, was well-viewed.
2) Maybe now that Kerry can actually campaign in Fla., and Bush doesn't get points for doling out federal aid left and right, and holding hands with his brother and disaster scenes, the recent hurricane skew will reverse.
3) Fla is full of old people. The last debate sharply pointed out differences in domestic policy regarding health care, medicare, SS, etc. Maybe the content of the last debate was more influential on the generally older crowd in Fla.
Look for the shifts post-debate in Fla. I think the momentum could change suddenly.
Posted by: PAVoter at October 16, 2004 12:42 PM | Permalink | Edit Comment | Delete Comment
I've been getting the bloglet updates.
Posted by: Joel Caris at October 16, 2004 04:04 PM | Permalink | Edit Comment | Delete Comment
I would ignore anyone or any poll claiming to have the state of FL's pulse. With all they have been through, FL is more of a wild card than ever. FL is totally up in the air, and will be until Nov. 2nd--but hopefully not up in the air the day after and beyond!!
Posted by: pepe at October 16, 2004 04:46 PM | Permalink | Edit Comment | Delete Comment
Being born, raised and living in FL, and having family in different parts of the state, I don't think the FEMA stuff has any impact.
Pretty much everyone who got impacted by the storms was insured. Pretty much everyone who got impacted is hurting due to the deductibles. FEMA is not offering to pay those; honestly I don't know what FEMA does, and neither do my folks in WPB, who got hit by two storms and went 12 days without power in the first one, and three weeks without cable in the second.
People think Jeb did a pretty good job, think the power companies suck, and hope it doesn't happen again.
But nobody is expecting FEMA to pay their bills at the end of the month.
Posted by: The Other Rob at October 16, 2004 06:26 PM | Permalink | Edit Comment | Delete Comment
To clarify MoE a bit:
MoE ONLY refers to statistical margin of error. It is the unavaoidable error created when you pick a random sample out of a larger group. It is related ONLY to sample size, nothing else.
There are several other sources of error -- for example, its pretty hard to get a random sample. The polling companies have been working hard at it, with different methods, but the fact that many of them weight samples to match population indicates that this is still an art, not a science.
When it comes to estimating turnout (LVs vs RVs), there is still more modelling of turnout rates etc. so theres even more possiblity of error beyond MoE.
Re: this poll. Insider Advantage did a 3 day poll. The full sample showed a 47-44 Bush lead, but the last day's sample (after the 3rd debate) showed Kerry up 48-44. The margin of error is higher on this because its a smaller sample. It may be that the debate caused this swing. I sure hope so, but one day polls are not always reliable.
Posted by: erg at October 16, 2004 07:11 PM | Permalink | Edit Comment | Delete Comment
Thanks for the explanation. I think publishing the MOE is somewhat misleading in that it makes non-math oriented people think that it means more than it does.
Posted by: The Other Rob at October 17, 2004 09:48 AM | Permalink | Edit Comment | Delete Comment