Contact Information

Call us today to see how
we can help your organization.

Election Insight, LLC.
18283 Bentwood Dr.
Doylestown, OH 44230

P: 330.715.2307

Clinton.Cooper@electioninsight.biz

Elections 2012 Update:

FINAL PREDICTION: Ohio's 16th District Congressional Race Prediction

I am predicting that Jim Renacci will hold the Ohio 16th District Congressional seat by defeating Betty Sutton by a 51% to 46% margin. The race has tightened significantly since my last prediction of two weeks ago due to a significant increase in Congressional approval by the Democrats and Independents, an influx of contributions by Democrats to liberal organizations, and a slight improvement in the unemployment rate which helps Democrat candidates.

I am also predicting a 64.45% turnout for individuals in the Congressional District. This is slightly below 2000 (64.85%) and 2008 (65.86%), and well below the high in 2004 (68.64%).

Ohio 16th Congressional District Map

 

 

Ohio's 16th District Congressional Race Prediction

If the election were held today, Jim Renacci would defeat Betty Sutton by a 57% to 43% margin. I am predicting a 64.5% turnout percentage for the 16th District, which is a bit above 2000 (64%) and well below the high of 68% in 2004. I will have one more prediction that I will publish on the eve of the election. Stay tuned and check back again on November 5th! For detailed results, see the map below.

Ohio 16th Congressional District Map

 

 

Current Ohio Early Voting Polls May be Flawed

There has been a lot of publicity over the last few days about a new poll coming out of Ohio done by SurveyUSA that shows that nearly 26% of Ohioans have already voted and that the vote is breaking 54% for Obama and 39% for Romney. A Suffolk Poll a few days also shows the same early voting pattern with 20% of Ohioans already voting with 54% going to Obama and 41% to Romney. Similar polling done in 2008 by Survey USA also shows the same early voting trends. Obama led McCain by 57% to 39% in Ohio, and an ABC/Washington Post poll reiterated that Obama led McCain in early voting by a 58% to 40% nationally.

At first look, these numbers are deflating to the folks supporting Romney. Contrary to what is being reported, closer examination of Survey USA’s current statistics may not be an accurate portrayal of what is really going on in Ohio.

First, Survey USA is showing that 26% of the Ohio Registered voters have already voted. In contrast, current early voting statistics for Ohio show that only 14% of the total Ohio Registered Voters have actually voted early. Joel Fagin, an Engineer from Kansas City Missouri, has developed a spreadsheet that shows Ohio’s county by county, up-to-date early voting statistics. It not only lists how many people have voted early, but also the Party Affiliation of those who have voted. SurveyUSA and other polling operations are vastly over estimating the number of early voters.

Second, the Survey USA poll says that Obama is winning the early vote 54% to 39%. This huge Obama advantage would seem contrary to current state and national polling that shows Romney is holding an 8% edge with Independent. Based on 2008 Party IDs, Obama won the early vote by nearly 16%. Today, the Obama 2008 lead has been cut down to 6%. While it is clear the number of Democrats and Republicans that have voted early, it is not certain how many of those with no Party IDs/Independents voted for one of the Presidential candidates. This can be deduced by making a few assumptions.

Before going into the analysis, some background information on how Ohio determines party affiliation is needed. While Ohio does not have a true party affiliation figure for each individual, Ohioans are assigned a party affiliation ID based on the last even year primary in which they voted. If someone voted a Democratic ballot in 2010, then they are assigned a Democrat party ID. If an individual did not vote in a long period of time, or choose to vote a ballot that is neither a Republican nor Democratic ballot, they are assigned a No Party/Independent ID. Because Ohio does not require a Party ID for all registered voters, early voting statistics show that nearly 52% of the early vote in 2008 and 51% of 2012 are by individuals with no Party ID.

Party Identification Assumptions for 2008: If every person who early voted with a Democrat Party ID voted for Obama, and every person who voted early with a Republican Party Id voted for McCain, it can be determine how many of the No Party Ids/Independents voted for each candidate. We can get these proportions from the polling data that SurveyUSA uncovered.

In 2008, according to the Ohio early voting spreadsheet, 1,386,683 Ohioans voted early. 324.43% were Democrats, 18.78% were Republicans, and 46.79% were votes with no Party ID/Independents.

Using the 2008 SurveyUSA early voting preference, we can extrapolate how many people voted for Obama and McCain.

If we assume that all the Democrats voted for Obama and all the Republicans voted for McCain, we determine the number of those with No Party ID/Independents that voted for Obama and for McCain.

In 2008, 52.75% of those individuals with No Party Id/Independents voted for Obama and 47.25% voted for McCain. It was estimated from exit polling that Obama won Independents by +8% in 2008, which is confirmed by this early voting analysis.

Running the 2012 numbers from the Ohio early voter spreadsheet and the SurveryUSA reported polling data it follows:

The SurveyUSA data suggests that 60.32% of the early vote from those with No Party ID/Independents have voted for Obama. As of today, most polls suggest that that Romney is leading Independents by +8%. Even in Survey USAs poll, they report Independents breaking for Romney by +8%. We can only conclude that early voting polling, at this point, cannot be trusted to provide accurate results. The flaw may lie in the oversampling of Democrats to Republicans and overestimating who has already voted or who will be voting on Election Day.

With a 6% lead in Registered Democrats over Republicans, there is no doubt Obama is probably winning the early voter turnout. However to get a +8% Independent support for Romney, Obama’s early lead would probably be around 50% to 46% or 47% for Romney. Perhaps a better indication of how the early voting is going, and for whom it is breaking , is to look at the total number of Democrat and Republican Party affiliated voters and infer the total early votes received. Given how early voting turned out in 2008, Romney is in good shape to carry Ohio in 2012.

Probably 2012 Early Voting Split that would also show a +8% Independent advantage:

Update 10/28/2012: Bias Confirmed. 

A new poll conducted by Time shows Obama leading Romney by 5% points in Ohio.  Not only is this poll ludicrous, there is no defense for the internals for the methodology of the poll.  Time reports that Obama will win Ohio by 5% points with a party id split of 37% Democrat, 28% Republican and 29% Independents.  This poll predicts the Democrats will outperform what occurred in 2008 by nearly 4% (Ohio 2008 Exit Polls, 37.5% Democrat, 32.5% Republican and 30% Independent). 

Second, the poll shows that 20% of the respondents have voted early.  This is closer to reality than the SurveyUSA poll which was discussed earlier.  But what is astonishing is who the early voters are supporting.  The poll reports that 60% of early voters voted for Obama and 30% voted for Romney.  This is a down right mistruth.  On page 11 of the report, the poll shows that the party id of those who have voted early is as follows, 46% Democrat, 22% Republican, and 27% Independent.  This is not even close to an accurate portrayal of what happened in 2008 or the current dynamics of the race in 2012.  Furthermore, there is a plus 45% (20 to 11 total votes) Independent voter advantage for Obama over Romney.  In reality, poll after poll shows that Romney holds a +8% lead of Independents over Obama.

We must take all the early voting polls with a grain of salt.  They are not sampling enough or providing an accurate picture of who early voters are voting for.  

Cuyahoga County Voter Registration and how it will impact Ohio's Electoral Votes

Cuyahoga County has historically been a bell weather county for determining the outcome of elections in Ohio. Not in direct terms of saying that if Candidate A wins Cuyahoga County by x percentage of points that he will win Ohio by x percentage of votes, but that Cuyahoga County has a large enough number of registered voters that you can statistically determine how many total votes each candidate will get in Ohio based on how many votes are cast for each party in Cuyahoga County.

Fox News recently reported that Ohio lost 490,000 registered voters since the 2008 election, with 44% lost from Cuyahoga County. According to the Cuyahoga County Board of Elections, that equals 208,207 fewer total registered voters in 2012 compared to 2008. The Cuyahoga County Board of elections also reports the breakdown of party affiliation for registered voters. Since 2008, Cuyahoga County has experienced a loss of 48,872 Democratic registered voters and 194,199 Independent Registered voters. That total does not add up to 208,207 voters. The Fox News article fails to mention that Cuyahoga County has gained 34,864 Republican voters since the 2008 election. Based on presidential elections from 1980 to 2008 an average 17% of the total Democrat votes cast in Ohio were cast in Cuyahoga County. In recent years, this has declined (to around 15.6% in 2008) as Cuyahoga County continues to lose population to surrounding counties. During the same time period, on average, 9.3% of the total Republican ballots cast in Ohio were cast in Cuyahoga County. However, due to population loss, the number has declined to 7.5% in 2008. Since 1980, the average voter turnout for Cuyahoga County is 64.57%. Past election results help predict future election results. Using this knowledge the number of votes that the Republican and Democratic candidates may receive on Election Day can be predicted. Because of Cuyahoga County’s aforementioned historical accuracy in predicting elections it is assumed that future election Ohio vote totals can be predicted based on Cuyahoga County.

The key to predicting each candidate’s vote share is determining the Independent vote split for each of the Republican and the Democrat candidates. Because of voter confidentiality it is not known precisely how many Republicans, Democrats or Independents voted for any particular candidate in the previous elections. However, statistical analysis can estimate voter percentages based on historical voter turnout. The first step in this analysis is to take the number of registered voters by party then proportion the total votes received for each candidate by the turnout percentage on Election Day. For example: in 2008, Cuyahoga County voter turnout was 60.34%. That 60.34% multiplied by the total number of registered Democrat voters (395,514) equals 238,656 registered Democrat voters in 2008. In 2008, Obama received 458,204 votes in Cuyahoga County. To estimate how many Independents voted for Obama in 2008, first take the total votes for Obama (458,204) less the proportion of registered Democrats (238,656) to get the approximate 215,548 Independent voters that voted for Obama. This is roughly 35.1% of total registered Independents in Cuyahoga County. Performing the same calculations for Republicans voters shows that 55,161 registered Republicans and 144,703 Independents (23.1%) voted for McCain in 2008. This same procedure can be performed for each election since 1980 showing that on average 28.2% of the registered Independents historically vote Democrat and 30.3% historically vote Republican. The 2012 vote totals for Romney and Obama can be predicted using this historical 64.57% average turnout for elections since 1980. The projection is that Obama will receive 345,693 votes to Romney’s 212,178 in Cuyahoga County.

Using the Cuyahoga County voter numbers discussed at the beginning of the article, Obama and Romney 2012 Ohio vote totals can be determined. In 2008 Cuyahoga County accounted for 15.6% of all the votes that Obama received (this number may in fact be closer to 15% as the population in Cuyahoga County continues to shrink). Based on this the prediction can be made that in 2012, Obama will receive 2,215,978 votes and Romney would win 2,829,037 votes if 7.5% of the total votes cast for Republicans were cast in Cuyahoga County. Romney would win 56.1% of the Ohio’s vote to Obama’s 43.9% with a turnout percentage of 64.67%. This analysis includes the assumption that total turnout will be much lower than past Ohio elections, confirming reports that overall enthusiasm is down for this presidential election compared to 2008.

Critics of this technique will say that too many assumptions are made in the analysis, including assuming the turnout of Cuyahoga County in 2012 as well as the possible independent vote split for the County. However, even if same turnout as in 2008 is used and the same vote split of Independent for Obama and Romney in 2012 as it was for Obama and McCain in 2008, then Romney would win 50.36% of the vote to Obama’s 49.64%. Simply stated, Cuyahoga County has lost enough Democrats and Independents since 2008, that when projected across the state of Ohio, Obama’s big 2008 gains are all gone.

The analysis does not guarantee that Romney will win Ohio by 12%, but with certainty the data does support that the changes in voter registration in Cuyahoga County have made it extremely difficult for Obama to win Ohio. This simple model estimation shows that with certainty current polling estimations in Ohio are based on outdated or ill-informed assumptions resulting in wrong conclusions. Given how drastically voter registration has changed in Cuyahoga County since 2008, it is likely that Romney can carry the State of Ohio, which is quite contrary to what popular media polls are saying.

Ohio's 2012 Possible Presidential Election Outcome based on Past Voter Registration History in Cuyahoga County Data

Ohio's 2012 Possible Presidential Election Outcome based on 2008's Voter Registration History in Cuyahoga County Data

Ohio's 16th District Congressional Race Prediction

If the Ohio Congressional 16th District Election were held today, September 28, 2012, my analysis shows that Jim Renacci would win his re-election by defeating Betty Sutton 52% to 42%. For detailed results, see the map below.

Ohio 16th Congressional District Map

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Elections 2010:
A Case of the Good the Bad and the Ugly!

The Good

Election Insight, LLC is very proud to have helped elect the Ohio State Senator of the 27th District on Tuesday night. He won with nearly 57% of the vote. Election Insight assisted his campaign with modeling expertise and GIS mapping. 

The Bad

Election Insight, LLC would like to congratulate Amy Schwan on running an outstanding campaign despite falling just short in her bid to win one of the Summit County Council At-Large seats. Picking up almost 16% of the vote, there was just a 4,000 vote difference between the second place and the last place candidates. We wish Amy the best of luck in her endeavors as she moves forward.

The Ugly; How did Election Insight do on Election Night?

In Election Insight’s first attempt at modeling an election, we had mixed results sprinkled in with some optimism for future modeling campaigns.

Turnout
Election Insight, LLC predicted that there would be between 54% to 59% turnout in Summit County for the 2010 Elections. The actual turnout was 52%. We were a bit generous with our prediction.  In 1994 Summit County had nearly a 60% turnout rate which led us to believe that we would see that same turnout rate in 2010. Since our model does not breakdown who (Republicans, Independents, or Democrats) voted, we cannot specify which groups did or did not come out to vote last Tuesday. In hindsight however, several other different variables were found that accurately predicted the 52% turnout. These variables will be used for future modeling and will give us precise predictions in the future. 
 
Ohio State Senate, 27th District
Election Insight, LLC predicted that The Ohio State Senator from the 27th District would win with a 63% vote share. While we were correct in our estimates that Frank would win, we overestimated that result by about 6%. 

Summit County Council At-Large
Election Insight, LLC predicted that all three Republicans would win their respective Council seats with 21%, 18% and 16% share. The highest percent share the any one of the Democrat candidates would get was 15.5%. In actuality however, Two Democrats won with 18% and 16.5% while Bill Roemer, the lone Republican to get elected, pulled in nearly 17% of the votes.  The total number of potential votes cast for this races was predicted to be around 641,000 and 504,000 being the actual total of votes cast for all six candidates. This equals a 78% voter fatigue ratio. In reality, the potential number of votes for this race came out to be 578,000 with only 407,041 actually being cast. This equals a 70% voter fatigue number. Therefore not only did we overestimate the turnout of voters for the election we also overestimated the number of votes that would be cast for this race. The 70% voter fatigue number represents the lowest voter fatigue that has been seen for all of the elections since 1994. In fact, the lowest voter fatigue number we had encountered for the span of elections was 76% in 2006. The highest voter fatigue was 82% in 1998. Therefore predicting a 78% voter fatigue was not out of the ordinary and it was a very safe assumption to make.

Reasons for Optimism

In Election Insight LLC’s first attempt at predicting elections, an earlier model used for both the Ohio State Senator for the 27th District and the Amy Schwan campaigns was almost exactly correct in predicting the actual vote share and turnout for both campaigns. This model was prematurely discontinued due to a misinterpretation of two variables. When the original model was revamped and the new variables that increased the accuracy and robustness of the model were included, the predicted results nearly mirrored the election results.

 

 

Future model predictions will be more accurate based on the learning from this election and new assumptions and we are optimistic about providing the most accurate results for our clients in the future.

© 2010 Election Insight, LLC All Rights Reserved.   |  Web Design and Development by WebScapes HD, LLC   |  Sitemap