NH08DemPrimary,Media Missed that the Polls got it 80% Right

After the New Hampshire 2008 Democratic Primary in January 2008 in which Hillary Clinton pulled a surprise upset victory over Barack Obama, I heard a lot on TV and radio about Obama losing votes in the last couple days before the primary or the pollsters being off about his estimates. I looked at the polls and the actual voting results and Obama did as expected. 12 polls cited below predicted what Obama was expected to get. The lowest prediction was 32% and the highest at 39%. Obama received 37%. Since polls usually concede a + or- 4% margin of error, having only one of the 12 polls outside that margin of error (and that poll only off by 5%) is pretty good prediction work for Obama’s vote outcome.
Hillary Clinton did much better than expected,but I would contend it is likely she did not get her votes from potential Obama voters since his final number was close to the poll prediction. 11 of 12 polls had the combined total votes among Edwards, Richardson and Kucinich as more than the 23 percentage points these three candidates actually received and I would contend she received some of these votes. If she received 2-3 percentages points from each of these candidates Clinton’s percentage would increase 4-6 % (before we even factor in the likelihood she won a majority of the undecided vote), but each of these candidates’ results would remain within the 4% margin of error since their result would only be 2-3% different than the poll numbers.


Clinton’s lowest poll number among these polls was a 28%. If she got 6-9% from Edwards, Richardson and Kucinich, her number rises to 34-37%. She then needs to pick up only 2-5% of the overall vote from the undecideds (who made up 3-14% of those polled) to get the 39% she won the primary with and the pollsters were still correct for four of five candidates on the ballot. Unfortunately, the one candidate they were outside the margin of error for happened to win. If Kucinich had somehow gotten 8% it would have barely registered in the media,but it also would have been outside the margin of error. The story was purportedly about the pollsters being wrong, but the focus was only “who won”. The pollsters actually (within the generally accepted 4% margin of error) got it 80% right.


http://www.presidentpolls2008.com/primary-election-poll-results/new-hampshire-democratic-republican-polls.html
Poll Obama Clinton Edwards Richardson Kucinich
American
ResearchGroup 39% 28% 22% 4% 2%
CBS 35% 28% 19% 5% 4%
CNN/WMUR 39% 30% 16% 7% 1%
Marist 36% 28% 22% 7% 3%
Suffolk Univ 35% 34% 15% 3% 0%
Fox 32% 28% 18% 6% 2%
Franklin PierceU 34% 31% 20% 6% 1%
Zogby 39% 29% 19% 6% 2%
Strategic Vision 38% 29% 19% 6% 3%
USA Today 41% 28% 19% 6% 3%
Research2000 34% 33% 23% 4% 3%
MSNBC 33% 31% 17% 7% 1%
http://www.cnn.com/ELECTION/2008/primaries/results/dates/index.html#20080108
Actual Vote
Obama Clinton Edwards Richardson Kucinich
37% 39% 17% 5% 1%

Here are links to some media stories about the New Hampshire primary polls:

http://www.cbsnews.com/stories/2008/01/14/opinion/pollpositions/main3709095.shtml

http://www.aapor.org/newhampshirepollswhatwentwrongupdated

http://www.pbs.org/newshour/bb/politics/jan-june08/gang_01-09.html

1st State Senate 2006 Primary Results

All numbers cited are available at the State of Delaware Election Department website:

http://elections.delaware.gov/information/electionresults/election_archive.shtml


Carroll/1st Rep(201),3rdRep(6),4thRep(71),6th Rep(39)11thRep(13),TOTAL VOTES 330

Hendrix/1stRep(76),3rdRep(2),4thRep(95),6th Rep(50),11thRep(8),TOTAL VOTES 233

McDowell/1st Rep (450),3rd Rep(15)4th Rep(620),6th Rep(351),11thRep(41)TOTAL 1477

Potter/1st Rep (566),3rd Rep(13),4th Rep(243),6th Rep(116),11thRep(6),TOTAL 944

Carroll (11.1%), Hendrix (7.8%), McDowell (49.5%), Potter (31.6%) of 2984 votes cast

Voter Turnout 2002&20061stSenatePrimary(Dems)

All numbers cited are available at the State of Delaware Election Department website:

http://elections.delaware.gov/information/electionresults/election_archive.shtml



Turnout by Representative District in 2002 & 2006 -1st State Senate Democratic Primary
2002
1st Representative District 1214 votes 55.6% of total turnout

3rd Representative District 34 votes 1.6% of total turnout

4th Representative District 675 votes 30.9% of total turnout

6th Representative District 231 votes 10.6% of total turnout

11th Representative District 29 votes 1.3% of total turnout

Total Turnout 2181 votes 100% of turnout
_________________________________________________________________________________

2006

1st Representative District 1293 votes 43.3% of total turnout

3rd Representative District 38 votes 1.3% of total turnout

4th Representative District 1029 votes 34.5% of total turnout

6th Representative District 556 votes 18.6% of total turnout

11th Representative District 29 votes 2.3% of total turnout

Total Turnout 2984 votes 100%

What If 2006 1st State Senate Race Was 2 Way?

After hearing talk about whether Harris McDowell would have beaten Charles Potter if it had been a two-way race, I looked at the voter turnout in the2002 and 2006 Democratic primaries for the 1st state senate district.
The first state senate district consists of parts of five state representative districts. The 1st District includes the area of Ninth Ward area of Wilmington where State Senator McDowell lives and Brandywine Hills where Wilmington City Council member Potter lives, as well as adjacent suburban areas north of the city. The 4th representative district includes the area around Trolley Square & Rockford Park in Wilmington and adjacent suburban areas to the west of the city. The section of the 3rd District is in 2 nonadjacent areas in the city of Wilmington west of Adams St. The 6th and 11th Representative districts are suburban districts near Philadelphia Pike north of Wilmington.

McDowell won the 4th, 6th, and 11th districts handily. He beat the combined totals of his three opponents and beat Potter better than 2-1 in each of these districts. He edged out Potter by 2 votes in the 3rd district which only represented 1.3 % of the total district –wide turnout. Potter won the 1st representative district with 566votes to McDowell’s 450,Thornton Carroll’s 201 and Les Hendrix’s 76.
The theory that Potter would have won in a two-way race seems to be based on an assumption that Potter would have gotten the votes that went for Carroll and Hendrix. It also assumes that turnout would remain constant despite 2 of the 4 candidates not being on the ballot.
The results indicate that while the vote count increased by 79 voters in the 1st rep dist, Potter’s stronghold which includes part of his city council district, the 1st district’s overall influence decreased because of larger relative increases in voter turnout in areas that supported McDowell heavily, especially the 4th & 6th representative districts.
The 1st representative district has been represented in Dover by a Democrat since at least the late 1960s and the district is very used to having Democratic primaries that make a difference in who is sent to Dover. This may mean that the maximum voter turnout in an off-year election may have been reached by repeated primaries-the people that would vote, do vote.
The 4th & 6th districts have Democratic representatives in 2008, but had been represented by Republicans for over 20 years. 2006 was the first off-year ( therefore no Governor,County Executive or Wilmington Mayor race on the ballot)primary where the winner of the Democratic primary had a realistic shot of winning the general election. In the fourth district Gerald Brady and Loretta Walsh, both Wilmington City Council members for several years, opposed each other for the right to replace Republican Joseph DiPinto ,who retired. In the 6th district Diana McWilliams ran for re-election against Michael Dore for the seat McWilliams had won in 2004 which became open when Dave Ennis became the Republican nominee for Insurance Commissioner in 2004. These two races contributed to drive up the combined turnout in the 4th & 6th by 679 more voters in 2006 (1585) than had voted in 2002 (906). People who suddenly had more reason to vote , acted on that reason.
This resulted in the combined turnout of the 4th & 6th going from representing 41.5% of the district-wide total turnout in 2002 to 53.1% in 2006 and the 1st representative district’s relative turnout decreasing from 55.6% to 43.3 % of the total district turnout.
The numbers imply that for Potter to win he would have had to have win over all the Carroll-Hendrix voters among voters in the 4th & 6th and dramatically increased his vote count in the 1st,since the most he could pick up in the 3rd is 8 votes and even if he picked up all the nonMcDowell votes in the 11th he would lose the 11th 41-27. Even if Potter had gotten every vote cast for Carroll & Hendrix in the 4th & 6th he still would have lost those 2 districts to McDowell (971-614). If Potter got all the Carroll-Hendrix votes in the 1st he would have beaten McDowell (843-450).

If the same 2984 people voted and all the people who voted for someone other than McDowell voted for Potter, Potter would have won 1507-1477. To do this Potter would need to win all 255 nonMcDowell voters in the 4th & 6th Rep districts, voters in an area where McDowell beat Potter 971-359. If McDowell wins only 31 of those 255 voters (12% or 1 out of 8 voters) and turnout stays the same, McDowell wins.

If voter turnout goes down due to Carroll & Hendrix not being on the ballot, the off setting vote split has to be even more lopsided for Potter to win. Since McDowell had 49.5 % with the 2984 turnout, he only needs to keep those 1477 voters and have overall turnout decrease to 2954 ( a 1% drop in turnout) and McDowell has 50% +1. Carroll & Hendrix drew a combined 563 voters. 30 voters is only 5% of that 563 voters and it is within the realm of possibilities that the absence of either of these candidates may have convinced 5% of their combined voter base that they did not have enough reason to vote. It is possible that without Carroll and Hendrix the 18% gap between McDowell and Potter may have been larger.
It is possible that in a two way race Potter might have won,but he would need to have either massively energized turnout in the 1st or made massive inroads in the 4th & 6th Districts.

Disclaimer-First Post

The purpose of the first post is not to endorse any political candidate or political party, so no endorsement is stated or implied. I am not saying who would serve the consituents better,only reviewing the election results in an attempt to make an educated guess of whether the outcome might have been different if circumstances had been altered.

Purpose

The purpose of this blog is took look at numbers that have an impact on politics. Some of these items are election results, campaign finance information, polling data and whatever numbers seem relevant.
The focus will mostly be Delaware, but numbers from other areas and national numbers may also be the focus.
Numbers in politics represent behavior, so the numbers may tells us something about how some people may be behaving. Of course the numbers will always be open to interpretation.