(FMW)B Polling Accuracy Analysis Report

Foster McCollum White & Associates
______________________________________________________________________________________
1
Foster McCollum White Baydoun (FMW)B
Polling Accuracy Report
Polling Studies 2010 to 2012
And Methodology Overview
Conducted by
Report Completed
September 10, 2012

______________________________________________________________________________________
2
Foster McCollum White & Associates and Baydoun Consulting 2010 to
2012 Public Opinion Polling Accuracy Analysis
On behalf of Foster McCollum White & Associates and Baydoun Consulting, we are
proud to produce the following report on our joint public opinion polling projects and the
reliability of our statistical analysis and polling construct model. Our industry is an
evolving industry with the constant pace and need for information about voters,
consumers and businesses is growing at a rapid rate. Firms that are in a continuous state
of business process improvement are the firms that will continue to identify the models
for studying the various audiences and interpreting data in coherent, logical and ethical
manners. Our analytics service model is one that we strive to continuously improve our
processes and stay ahead of the trends within the analytics industry.
The following report highlights the polling studies that we have completed from the
Michigan Primary election of 2010 through the Michigan Primary election of 2012 and
the comparable election data from those completed election cycles. We are assessing our
polling methodology, statistical sampling models and analytic constructs to report
accuracy and effective in our model across the topline aggregate data and geographical
cross tabular data. We have complied direct election statistics for each election cycle to
measure the direct correlation between our poll respondents’ desired selections and the
actual election selections. Our findings within the report reflect a very high level of
statistical accuracy and predictive outcome modeling for determining the course of each
election that we conducted polling on. We believe that our previous predictive analysis
history and our continuing business process improvement model will allow us to continue
a high level of accuracy and data reliability for the 2012 General election cycle and future
election cycles.
We are in the midst of conducting multiple state polling studies for the 2012 General
election cycle and will update this report upon the completion of the November election.

______________________________________________________________________________________
3
Michigan Democratic Primary Governor’s Contest
July 29, 2010
FMW & Baydoun Consulting conducted a state-wide poll on the Michigan Democratic
Governor’s primary. A 1,648 sample survey of registered and likely Democratic voters
and hard partisan Democratic voters who have already voted or are certain or likely to
vote in the August 3rd
Democratic primary for Governor of Michigan was conducted on
July 29th
. This sample is reflective of voters who have a consistent history of participating
in Democratic Primaries over a minimum of the past four primary election cycles. The
margin of error for this polling sample is 4%. We have made no weighting adjustments to
the aggregate baseline responses. Our respondent demographics matched our PBVA
demographic models with a variance ratio of 5%.
Baseline question
If the election was held today in the Democratic primary for governor, who would you
vote for? Andy Dillon, Virg Bernero, or are you undecided??
(Andy Dillon): 21.8%
(Virg Bernero): 49.9%
(Undecided): 28.3%
Bernero margin in the FMW & Baydoun Consulting poll was 28.80 points.
The primary was held on August 3, 2010. The final results are listed below:
(Undecided): 0%
Bernero margin of victory in the Democratic Primary was 17.36 points. Bernero gained
8.93 points from the undecided voting universe (31.55% of the undecided respondents).
Dillon gained 19.67 points from the undecided voting universe (69.51% of the undecided
respondents).
Geographical Voting Regions
Bernero’s lead among hard partisan Democrats is virtually insurmountable in the other
three regions. They are as follows:
Central Michigan – July 29, 2010 Poll
Bernero 62.3%
Dillon 16.3%
Undecided 21.3%
Bernero margin in the FMW & Baydoun Consulting poll was 46 points.

______________________________________________________________________________________
4
Central Michigan – August 3, 2010 results
Bernero 72.15%
Dillon 27.85%
Undecided 0%
respondents).
Thumb region – July 29, 2010 Poll
Bernero 57.3%
Dillon 13.8%
Undecided 28.6%
Thumb region – August 3, 2010 results
Bernero 66.38%
Dillon 33.62%
Undecided 0%
respondents).
Southwestern Michigan Region – July 29, 2010 Poll
Bernero 50.8%
Dillon 16.2%
Undecided 33.0%
Southwestern Michigan Region – August 3, 2010 results
Bernero 58.61%
Dillon 41.39%
Undecided 0%
respondents).

______________________________________________________________________________________
5
Southeastern Michigan – July 29, 2010 Poll
Bernero 46.1%
Dillon 27.0%
Undecided 26.9%
Bernero margin in the FMW & Baydoun Consulting poll was 19.1 points
Southeastern Michigan – August 3, 2010 results
Bernero 54.44%
Dillon 45.56%
Undecided 0%
respondents).
Northern Lower Peninsula region – July 29, 2010 Poll
Bernero 42.7%
Dillon 17.7%
Undecided 39.6%
Bernero margin in the FMW & Baydoun Consulting poll was 25 points
Northern Lower Peninsula region – August 3, 2010 results
Bernero 56.51%
Dillon 43.49%
Undecided 0%
respondents).
Upper Peninsula Region – July 29, 2010 Poll
Bernero 46.8%
Dillon 14.9%
Undecided 38.3%
Bernero margin in the FMW & Baydoun Consulting poll was 31.90 points
Upper Peninsula Region – August 3, 2010 results
Bernero 54.88%
Dillon 45.12%
Undecided 0%

______________________________________________________________________________________
6
respondents).
FMW & Baydoun Consulting Polling Accuracy
The FMW & Baydoun Consulting polling model was extremely accurate when
comparing the aggregate and geographical cross tabular groups.
Our election poll results were:
(Undecided): 28.3%
(Undecided): 0%
Bernero margin of victory in the Democratic Primary was 17.06 points.
Our data model suggested that Speaker Dillon would win an average of 66% of the
undecided voters and Speaker Dillon won 69.51% of the undecided voters. Our polling
findings for Mayor Bernero’s lead among the aggregate voting universe and geographical
cross tabs were statistically accurate. We correctly projected that Virg Bernero would win
all six geographical regions of Michigan.

______________________________________________________________________________________
7
Michigan November General Election Polling Study for
Governor
October 11, 2010
FMW & Baydoun Consulting conducted a state-wide poll on the Michigan Governor
General election. A survey of registered and likely voters, equally pre-weighted towards
registered Democratic and Republican voter percentages and Independent voters who
have already voted or are certain or likely to vote in the November 3rd
general election
was conducted on October 11th
, 2010. 2331 Respondents participated in the survey; the
margin of error for this polling sample is 2.03%. We have made no weighting
adjustments to the aggregate baseline responses. Our respondent demographics matched
our PBVA demographic models with a variance ratio of 5%.
Baseline question
The general election for Governor is 23 days away. If the election for Governor was held
today, who would you vote for, certainly for Virg Bernero, leaning toward Virg
Bernero, Certainly for Rick Snyder, leaning towards Rick Snyder, another candidate or
undecided?
(Certainly for Virg Bernero): 26.51%
(Leaning toward Virg Bernero): 6.01%
Total Virg Bernero 32.52%
(Certainly for Rick Snyder): 43.97%
(Leaning toward Rick Snyder): 9.91%
Total Rick Snyder 53.88%
(Another Candidate or undecided): 13.60%
Snyder’s margin in the FMW & Baydoun Consulting poll was 21.36 points.
The General election was held on November 3, 2010. The final results are listed below:
Virg Bernero 39.90%
Rick Snyder 58.11%
Other Candidates: 1.99%
Snyder’s margin of victory in the November General election was 18.21 points. Bernero
gained 7.38 points from the undecided voting universe (54.26% of the undecided
respondents). Snyder gained 4.23 points from the undecided voting universe (31.10% of
the undecided respondents).

______________________________________________________________________________________
8
Geographical and Political Voting Regions
Cross Tab Grouping Rick
Snyder
(October
11 Poll)
Virg
Bernero
(October
11 Poll)
+/-
Margin
for
Snyder
Rick
Snyder
(Nov. 3
Final)
Virg
Bernero
(Nov. 3
Final)
+/-
Margin
for
Snyder
Southeastern Michigan –
Snyder Home Region
50.48% 38.00% +12.48 51.24% 47.10% +4.14
Southwestern Michigan 58.35% 26.87% +31.48 67.49% 30.52% +36.98
Thumb Region of Michigan 49.99% 34.80% +15.19 56.79% 41.08% +15.71
Central Region of Michigan
- Bernero Home region
59.25% 30.62% +28.63 63.46% 34.47% +28.99
Northern Lower Peninsula
Region of Michigan
55.15% 26.34% +28.81 66.45% 30.71% +35.74
Upper Peninsula Region of
Michigan
42.64% 29.41% +13.23 56.85% 39.95% +16.90
Major 17 County Cluster 52.72% 35.21% +17.51 55.41% 42.82% +12.59
Major Democratic 6 CC 43.27% 45.15% (-1.88) 43.64% 54.67% (-11.03)
Major Republican 7 CC 66.82% 22.35% +44.33 69.57% 28.55% +41.03%
Swing 4 County Cluster 54.84% 33.39% +21.45 60.42% 37.81% +22.61
Summation
The data suggest that while the Governor’s campaign was tightening pre-debate, it was
reversed by the debate and follow up coverage of it. The party themes of corporate
outsider who will take jobs from Michigan aren’t closing the knowledge gap inside of
Democratic voters, Major Democratic counties and urban communities. What Bernero
needed was a impactful and responding personal messaging event, to introduce him to all
key voter groups and start the community impact discussion of the campaign. Bernero
and Brenda Lawrence are both well positioned to have the discussion on community
impact, due to their current positions. Bernero’s campaign was either unwilling or unable
to create that type of messaging. FMW & Baydoun Consulting suggested the following
campaign strategy changes based on our polling data and political consulting experience:
 Improve his support ratings among Democratic voters and decease Snyder’s
incursion into the Democratic voter community
 Recapture his home voting region and win that region by a minimum 10 point
cushion
 Find a way to raise the policy and philosophical differences between Snyder and
the right of center organizations and Tea Party groups within the Republican
Party, to weaken Snyder’s base of support.
 Use his position as Mayor and running mate Brenda Lawrence’s experience as a

______________________________________________________________________________________
9
Mayor to reshape his message to more specific community impact focus and their
personal commitment to help citizens. The traditional partisan messaging isn’t
strengthening his support with Democratic voters or making him competitive with
independents. He is better positioned to argue for those voters due to the daily
impact mayor’s have on voters lives. That needs to be clearly communicated.
Snyder stayed on the right course for victory. Bernero’s campaign was never able to take
advantage of Snyder has to be the inability to consistently climb about the 45% threshold
in most voter groups. The data suggested that soft partisan and undecided voting blocs
are enough to throw this race into a competitive nightmare. However Snyder was able to
stay the course and the lack of clear positive alternative narrative from Bernero’s
campaign allowed Snyder to expand his base and solidify him at the 45% and above
threshold of hard core voter support. Overall, Snyder’s was able to do the following items
that FMW & Baydoun Consulting suggested from our polling data:
 Continue the messaging plan. He is making inroads into Democratic areas and
dominating Independent voting blocks. He must maintain this advantage.
 Press more campaign activity into the Central, Southeastern and Thumb Regions
of Michigan, to press the advantage and force Bernero to fight in Democratic and
his home base
 Continue and expand outreach efforts into Minority voting communities. These
communities are significant (12% and higher) in 5 of the major 7 Republican
counties and 3 of the 4 major swing counties along with all of the Major 6
Democratic counties.
Another Candidate or undecided: 13.60%
The General election was held on November 3, 2010. When all of the ballots were
counted, the final results are listed below:
Snyder’s margin of victory in the November General election was 18.21 points.

______________________________________________________________________________________
10
Our data model suggested that the candidates would split the undecided voters evenly and
Snyder won 31% and Bernero won 54% of the undecided voters. Our polling findings for
Rick Snyder’s lead among the aggregate voting universe and geographical cross tabs
were statistically accurate. We correctly projected that Virg Bernero would only win the
Major Democratic 6 county cluster and that Snyder would win 43% of the vote in this
heavy Democratic voting counties (43.27% poll results and 43.64% actual voter totals).
We correctly projected Snyder would win all six geographical regions of Michigan, the
Major 17 counties cluster group, Major Republican 7 counties cluster and the Major
swing 4 counties cluster which are the bellwether indicator for every Michigan general
election. Our polling results were within less than 1 point of the margin in the Thumb
region of Michigan, Central region of Michigan, Republican 7 counties cluster and swing
4 counties cluster in Michigan.

______________________________________________________________________________________
11
Michigan November General Election Polling Study for
Attorney General
Secretary of State
Michigan Ballot Proposal 2
October 7, 2010
FMW & Baydoun Consulting conducted a state-wide poll on the Michigan Attorney
General, Secretary of State and Ballot Proposal 2 for the November 2010 General
election. A survey of registered and likely voters, equally pre-weighted towards
registered Democratic and Republican voter percentages and Independent voters who
have already voted or are certain or likely to vote in the November 3rd
general election
was conducted on October 7th
, 2010. This study produced included 2,282 respondents.
The margin of error for this polling sample is 2.05%. We have made no weighting
adjustments to the aggregate baseline responses. Our respondent demographics matched
our PBVA demographic models with a variance ratio of 5%.
Baseline question – Attorney General
The general election for Attorney General is 27 days away. If the election for Attorney
General was held today, who would you vote for, solidly for Bill Schuette, leaning
toward Bill Schuette, solidly for David Leyton, leaning towards David Leyton, another
candidate or undecided?
(Solidly for Bill Schuette): 37.71%
(Leaning toward Bill Schuette): 11.59%
Total Bill Schuette 49.30%
(Solidly for David Leyton): 20.11%
(Leaning towards David Leyton): 5.68%
Total David Leyton 25.79%
Schuette’s margin in the FMW & Baydoun Consulting poll was 23.51 points.
Bill Schuette 43.48%
David Leyton 52.59%
Other Candidate: 3.94%
Schuette’s margin of victory in the November General election was 9.11 points. Leyton
gained 17.69 points from the undecided voting universe (71.02% of the undecided
respondents). Schuette gained 3.29 points from the undecided voting universe (13.20% of
the undecided respondents).

______________________________________________________________________________________
12
Baseline question – Secretary of State
The general election for Secretary of State is 27 days away. If the election for Secretary
of State was held today, who would you vote for, solidly for Jocelyn Benson, leaning
toward Jocelyn Benson, solidly for Ruth Johnson, leaning towards Ruth Johnson, another
candidate or undecided?
(Solidly for Jocelyn Benson): 35.53%
(Leaning toward Jocelyn Benson): 11.13%
Total Jocelyn Benson 46.66%
(Solidly for Ruth Johnson): 41.47%
(Leaning towards Ruth Johnson): 11.42%
Total Ruth Johnson 52.89%
Johnson’s margin in the FMW & Baydoun Consulting poll was 5.94 points.
Johnson’s margin of victory in the November General election was 5.46 points. Benson
lost 1.44 points from the polling respondent voting universe. Johnson lost 2.21 points
from the polling respondent voting universe.
Baseline question – Constitutional Amendment Ballot Proposal 2
The November General Election will have two ballot proposals. Proposal 2 will AMEND
THE STATE CONSTITUTION TO PROHIBIT CERTAIN FELONS FROM HOLDING
ELECTIVE OFFICE AND SPECIFIED TYPES OF PUBLIC EMPLOYMENT
POSITIONS if approved by voters. If the general election was held today, who would
you vote for, vote YES for Proposal 2, vote NO for proposal 2, undecided or will not vote
for this ballot proposal
(Vote YES for Proposal 2): 62.11%
(Vote NO for proposal 2): 14.18%
(Undecided): 23.30%
(Will not vote for this ballot proposal): 0.42%

______________________________________________________________________________________
13
Total Vote YES for Proposal 2 74.91%
Total Vote NO for proposal 2 25.09%
(Another Candidate or undecided): 0%
The Yes side’s margin of victory in the November General election was 49.82 points.
The No side’s gained 10.91 points from the undecided voting universe (46.82% of the
undecided respondents). The Yes side’s gained 12.8 points from the undecided voting
universe (54.18% of the undecided respondents).
2010 Attorney General Geographical Voting Communities
Cross Tab Grouping Bill
Schuette
(October
7 Poll)
David
Leyton
(October
7 Poll)
+/-
Margin
for
Schuette
Bill
Schuette
(Nov. 3
Final)
David
Leyton
(Nov. 3
Final)
+/-
Margin
for
Schuette
Southeastern Michigan 47.29% 27.29% +20.00 45.17% 51.41% (-6.26)
Thumb Region of Michigan
– Leyton Home Region
45.50% 29.03% +16.47 50.55% 45.06% +5.48
Central Region of Michigan
- Schuette Home region
49.12% 28.07% +21.05 59.06% 36.59% +22.46
Region of Michigan
50.73% 19.81% +30.92 62.24% 32.51% +29.65
Michigan
48.58% 28.58% +20.00 53.84% 41.50% +12.34
Summation
The Attorney General’s campaign was over at the time of our polling study. Bill
Schuette’s aggregate advantage of 23.51 points and the cross tab advantage of an average
of 24.13 point margin over Leyton were too much to overcome.

______________________________________________________________________________________
14
Total Bill Schuette 49.30%
Total David Leyton 25.79%
Schuette’s margin in the FMW & Baydoun Consulting poll was 23.51 points.
Bill Schuette 43.48%
David Leyton 52.59%
Schuette’s margin of victory in the November General election was 9.11 points.
We correctly projected that Bill Schuette would win all five of the six geographical
regions of Michigan and the Major 17 counties cluster group. Our polling results were
within less than 1.5 point of the margin in the Northern Lower Peninsula region of
Michigan and Central region of Michigan.

______________________________________________________________________________________
15
2010 Secretary of State Geographical Voting Communities
Cross Tab Grouping Jocelyn
Benson
(October
7 Poll)
Ruth
Johnson
(October
7 Poll)
+/-
Margin
for
Benson
Jocelyn
Benson
(Nov. 3
Final)
Ruth
Johnson
(Nov. 3
Final)
+/-
Margin
for
Benson
Southeastern Michigan –
Benson & Johnson’s Home
Region
50.00% 49.78%% +0.22 53.09% 43.48% +9.62
Southwestern Michigan 41.30% 58.71% (-17.41) 35.30% 60.55% (-25.25)
Thumb Region of Michigan 46.55% 52.88% (-6.33) 46.32% 49.15% (-2.83)
Central Region of Michigan 49.53% 50.00% (-0.47) 39.43% 55.98% (-16.55)
Region of Michigan
41.26% 56.64% (-15.38) 35.72% 58.74% (-23.02)
Michigan
50.01% 50.00% 0.00 41.78% 52.64% (-10.78)
Major 17 County Cluster 47.98% 51.80% (-3.82) 48.32% 48.00% +0.32
Summation
Benson was the one Democratic candidate that had a reasonable roadmap to victory.
Johnson underperformed significantly when considering the brand advantage that
Republicans enjoyed during the 2010 election season. Johnson was hamstrung by two
factors; voter mood against institutionally experienced candidates and voter interest in
fresh eyes to view and fix government. Benson offered both factors, which made it
difficult for Johnson to advance her message. The election results proved our polling
analysis that the Secretary of State was the closest race of the three topline campaigns.
Benson was able to win Southeastern Michigan by 9.62 points in part to her ability to
regain support among Macomb County Democrats and the major 17 counties cluster.
Johnson was able to hold off Benson by increasing her victory margin in the following
regions:
 Central region of Michigan, winning 16.55 points.
 Northern Lower Peninsula region, winning by 23.02 points.
Benson also was unable to win the Thumb region of Michigan which was critical to her
ability to win statewide. She out performed both topline Democratic candidates (Bernero
and Leyton) but was unable to win this important corridor.

______________________________________________________________________________________
16
Johnson’s margin in the FMW & Baydoun Consulting poll was 5.94 points.
Johnson’s margin of victory in the November General election was 5.46 points.
Our polling findings for Ruth Johnson’s lead over Jocelyn Benson and their individual
percentages of voter support were all within the overall margin of error. We correctly
projected that Benson would win the Southeastern Michigan region. We correctly
projected Johnson would win the other five geographical regions of Michigan.

______________________________________________________________________________________
17
2010 Proposal 2 Ballot Campaign
Cross Tab Grouping Yes on 2
(October
7 Poll)
No on 2
(October
7 Poll)
+/-
Margin
for Yes
on 2
Yes on 2
(Nov. 3
Final)
No on 2
(Nov. 3
Final)
+/-
Margin
for Yes
on 2
Southeastern Michigan 59.85% 14.78% +45.07 74.44% 25.56% +48.87
Thumb Region of Michigan 60.53% 14.47% +46.06 73.30% 26.70% +46.61
Central Region of Michigan 60.91% 17.26% +43.65 75.01% 24.99% +50.02
Region of Michigan
70.00% 13.33% +56.67 77.48% 22.52% +54.96
Michigan
60.00% 17.50% +42.5 76.54% 23.46% +53.07
Summation
The data suggested that State ballot Proposal 2 was trending towards a statistically
significant victory. Proposal 2 had wide margins of support among most voter groups and
was on track to perform above 60% among voter categories across the State. Voter
sentiment seemed to reflect a mood to make all persons in elected and appointed office
accountable for in office actions that violate the public trust and misuse public funds for
personal benefit in all levels and areas of government.
(Vote YES for Proposal 2): 62.11%
(Vote NO for proposal 2): 14.18%
(Undecided): 23.30%
(Will not vote for this ballot proposal): 0.42%
The “Yes on 2” margin in the FMW & Baydoun Consulting poll was 47.93 points.

______________________________________________________________________________________
18
Total Vote YES for Proposal 2 74.91%
Total Vote NO for proposal 2 25.09%
(Another Candidate or undecided): 0%
The Yes side’s margin of victory in the November General election was 49.82 points.
Our polling findings were within a two point variance for the overall victory margin and
three of the six geographical margins. We correctly projected the “Yes on 2” side would
win all six geographical regions of Michigan and the Major 17 counties cluster group.

______________________________________________________________________________________
19
Michigan 14th Congressional District Democratic
Primary Election Exclusive Polling Study for
Fox 2 News Detroit
July 23, 2012
, a national public opinion polling and voter
analytics consulting firm based in Michigan and representing the combined resources of
Foster McCollum White & Associates (Troy Michigan) and Baydoun Consulting
(Dearborn Michigan) conducted a telephone-automated polling random survey of
Michigan registered and most likely August 2012 primary election voters in the 14th
congressional district to determine their voting and issue preferences on the Democratic
Congressional nomination. 685 respondents fully participated in the survey. The margin
of error for this polling sample is 3.74% with a confidence level of 95%. We have made
no weighting adjustments to the aggregate baseline responses. Our respondent
demographics matched our PBVA demographic models with a variance ratio of 5%.
(FMW)B
polling analysis for Fox 2 News Detroit on the Michigan 13th
and 14th
congressional district campaign and U.S Senate primary was extremely accurate and
identified the key trending factors that allowed Congressman Peters, Conyers and former
Congressman Hoekstra to defeat well financed serious competitors in their districts. Our
analysis starts with the 14th
Congressional District.
We found that Congressman Peters had a significant lead over all five candidates, who
are or have been elected officials themselves. The aggregate results for all five candidates
are listed below:
The 2012 Michigan 14th Congressional District Democratic Primary election will be held
in August. Who are you most likely to vote for in the primary for US House of
Representatives?
District - Wide Aggregate Results – 685 Respondents MOE +/- 3.74%
Current Democratic Congressman Hansen Clarke: 26.72%
Current Democratic Congressman Gary Peters: 45.11%
Southfield Mayor Brenda Lawrence: 9.64%
Former State Rep. Mary Waters: 1.17%
Retired Magistrate Bob Costello: 0.29%
Undecided: 17.08%

______________________________________________________________________________________
20
Undecided: 0%
27%
45%
10%
1%
0%
17%
Hansen Clarke
Gary Peters
Brenda Lawrence
Mary Waters
Bob Costello
Undecided
35%
47%
13%
4%
1% 0%
Hansen Clarke
Gary Peters
Brenda Lawrence
Mary Waters
Bob Costello
Undecided

______________________________________________________________________________________
21
As reflected in the final results, Congressman Peters support was consistent across the
new 14th
Congressional district. Despite last minute negative attacks from the Clarke and
Lawrence camps, Peters was able to add to our polling data level of support by picking
up 1.92 points from the undecided voters (45.11% to 47.03%). Congressman Clarke and
Mayor Lawrence were able to grow their individual bases of support with their more
aggressive posture against Peters, but not enough to put the race in jeopardy. Clarke
picked up the biggest share of undecided voters, gaining 8.46 points or 49.53% of the
undecided voters. Mayor Lawrence gained 3.65 points or 21.37% of the undecided voters
from our July 23rd
poll. Mary Waters and Bob Costello were non factors in the aggregate
polling universe and with the Election Day voters.
A. Congressman Hansen Clarke positive cross tabular results.
Our findings reflect a limited range of support for Congressman Clarke across the cross
tabular groups. As we review the Election Day findings, we will assess Congressman
Clarke’s actual performance against his geographical bases of support. Our voters cross
tabular comparisons are as follows:
July 23, 2012 poll findings
Cross Tab
Grouping
Hansen
Clarke
Gary
Peters
Brenda
Lawrence
Mary
Waters
Bob
Costello
Undecided
Detroit region
Respondents
39.65% 28.42% 7.37% 1.05% 0% 23.51%
Wayne County
Cities Cluster
38.33% 33.06% 6.11% 0.83% 0% 21.67%
Actual August 7, 2012 primary election results
Cross Tab
Grouping
Hansen
Clarke
Gary
Peters
Brenda
Lawrence
Mary
Waters
Bob
Costello
Undecided
Detroit region
Respondents
54.18% 29.19% 10.77% 4.88% 0.84%
Undecided
gain
+14.53 +0.77 +3.40 +4.83 +0.84
Wayne County
Cities Cluster
51.29% 33.53% 9.43% 4.44% 1.30%
Undecided
gain
+12.96 +0.47 +3.32 +3.61 +1.30

______________________________________________________________________________________
22
In our poll findings, Congressman Clarke’s biggest margin of support was with
city of Detroit voters. He led Congressman Peters by 11.23 points. The data
suggest that Clarke is underperforming with Detroit voters. 91.52% of Detroiters
believe it is important to continue Detroit representation in Congress, yet only
39.65% of Detroiters would vote for the Detroit Congressman. Clarke was able
to improve his performance in Detroit winning by 24.99 points over
Congressman Peters and taking 61.80% (14.53 points) of Detroit’s undecided
voters.
Clarke also was able to improve greatly in the Wayne County region of the 14th
Congressional district. He gained 59.81% (12.96 points) of the undecided voters
in Wayne County. The results show that Congressman Clarke’s campaign
messaging change after our Fox 2 News polling did help him with his primary
base and secondary base voters. The data also shows that his messaging change
was too late to reduce the base that Peters built up in the Detroit and Wayne
County. Peters had 28.42% of the Detroit vote and 33.06% of the Wayne County
vote via our poll findings and he finished with 29.19% of the Detroit vote and
33.53% of the Wayne County vote. This reflects his efforts to build and retain a
solid beachhead of voter support in opposition regions worked. In spite of the
dual plank of negative attacks, neither Clarke nor Lawrence was able to move
voters away from Peters.
Additionally, the data suggest that Southfield Mayor Brenda Lawrence and
former State Rep. Mary Waters did not impact Congressman Clarke campaign in
Detroit or the Wayne County portion of the district. Lawrence and Waters have a
combined 8.42% of the Detroit vote and 6.94% of the Wayne County vote in our
July 23rd
polling study. They finished with a combined 15.65% of the Detroit
votes and 13.87% of the Wayne County votes. The results confirm the data
findings that Clarke’s challenge is not vote splitting by Lawrence or Waters; it
may be messaging and resource deployment which would have provided a wider
margin and reduced Peters’ beachhead of support.
B. Congressman Gary Peters positive cross tabular groupings:
Our voters cross tabular comparisons are as follows:

______________________________________________________________________________________
23
July 23, 2012 poll findings
Cross Tab
Grouping
Hansen
Clarke
Gary
Peters
Brenda
Lawrence
Mary
Waters
Bob
Costello
Undecided
Near Detroit
Suburbs
Region-Wayne
Respondents
31.94% 52.78% 1.39% 0% 0% 13.89%
Southwestern
Oakland
County Region
Respondents
10.57% 73.98% 4.07% 0.81% 0% 10.57%
Southeastern
Oakland
County Region
Respondents
16.37% 46.78% 22.22% 2.34% 1.17% 11.11%
Central
Region of
Oakland
County
Respondents
12.90% 61.29% 3.23% 0% 0% 22.58%
Major
Oakland
County Cities
Respondents
13.43% 58.58% 12.31% 1.87% 0.75% 13.06%
Oakland
County Region
Respondents
13.85% 58.46% 13.54% 1.54% 0.62% 12.00%
Cross Tab
Grouping
Hansen
Clarke
Gary
Peters
Brenda
Lawrence
Mary
Waters
Bob
Costello
Undecided
Near Detroit
Suburbs
Region-Wayne
Respondents
39.46% 50.82% 3.70% 2.46% 3.32%
Undecided
vote gain
+7.52 -1.96 +2.31 +2.46 +3.32
Southwestern
Oakland
County Region
11.12% 77.77% 8.02% 1.55% 1.47%

______________________________________________________________________________________
24
Respondents
Undecided
vote gain
+0.55 +3.79 +3.95 +0.74 +1.47
Southeastern
Oakland
County Region
Respondents
21.78% 46.47% 28.87% 2.09% 0.71%
Undecided
vote gain
+5.41 -0.31 +6.65 -0.25 -0.46
Central
Region of
Oakland
County
Respondents
11.04% 75.30% 9.65% 3.21% 0.64%
Undecided
vote gain
-1.86 +14.01 +6.42 +3.21 +0.64
Major
Oakland
County Cities
Respondents
15.22% 64.09% 17.78% 1.86% 0.98%
Undecided
vote gain
+1.79 +5.51 +5.47 -0.01 +0.23
Oakland
County Region
Respondents
16.04% 63.04% 17.87% 2.01% 1.02%
Undecided
vote gain
+2.19 +4.58 +4.33 +0.47 +0.40
Congressman Peters had significant number of cross tabular groups where his
support is above 49% of the most likely voter universe, what we define as a
“Super Majority”. The cross tabular groups that Congressman Peters has a
FMWB defined “Super Majority” are as follows:
 Near Detroit Suburbs region of Wayne County (Detroit, Hamtramck,
Harper Woods and the 5 Grosse Pointes) voters.
 Southwestern Oakland Region communities (Farmington Hills, West
Bloomfield, Orchard Lake City, Keego Harbor and Sylvan Lake) voters.
 Central Region of Oakland County (Pontiac) voters
 Major Oakland County Communities (Southfield, West Bloomfield,
Farmington Hills, Pontiac) voters
 Oakland County Region voters

______________________________________________________________________________________
25
Congressman Peters took Oakland County out of play for Congressman Clarke and
Mayor Lawrence campaigns. (FMW)B
predictive voter behavioral analysis model
projected Oakland County communities would make up 44.92% of the net voter turnout
for the August 7th
primary. Oakland County communities actually comprised 45.70% of
the net voter turnout for the August 7th
primary. Congressman Peters won 63.04% voter
support to Congressman Clarke’s 16.04% and Mayor Lawrence’s 17.87%, a margin of 47
and 45.17 points respectively.
(FMW)B
Polling Accuracy
The (FMW)B
polling model was extremely accurate when comparing the aggregate and
geographical cross tabular groups.
(Current Democratic Congressman Hansen Clarke): 26.72%
(Current Democratic Congressman Gary Peters): 45.11%
(Southfield Mayor Brenda Lawrence): 9.64%
(Former State Rep. Mary Waters): 1.17%
(Retired Magistrate Bob Costello): 0.29%
(Undecided): 17.08%
Congressman Peters’ margin in the (FMW)B
poll was 18.39 points.
The General election was held on August 7, 2012. The final results are listed below:
Congressman Peters’ margin of victory in the August Primary election was 11.85 points.
Our polling findings were very exact for Congressman Peters’ aggregate support
percentage and key geographical cross tab groups. Our poll findings found that
Congressman Peters’ had the following support:
 Southwestern Oakland County region – 73.98% poll support to 77.77% actual
election voter support
 Southeastern Oakland County region – 46.78% poll support to 46.47% actual
 Near Detroit Suburbs communities – 52.78% poll support to 50.82% actual
 Wayne County region – 33.06% poll support to 33.53% actual election voter
support

______________________________________________________________________________________
26
 Detroit region – 28.42% poll support to 29.19% actual election voter support
We also correctly projected that Congressman Peters would win all Oakland County
voting regions and the Near Detroit Suburbs voting region of Wayne County and that
Congressman Clarke would win the Detroit region and Wayne County region of the 13th
Congressional District.

______________________________________________________________________________________
27
Michigan 13th
Congressional District Democratic
Primary Election Exclusive Polling Study for
Fox 2 News Detroit.
July 21 and July 22, 2012
Michigan registered and most likely August 2012 primary election voters in the 13th
congressional district to determine their voting and issue preferences on the Democratic
Congressional nomination. 643 respondents fully participated in the survey. The margin
of error for this polling sample is 3.84% with a confidence level of 95%. We have made
no weighting adjustments to the aggregate baseline responses. Our respondent
demographics matched our PBVA demographic models with a variance ratio of 5%.
Our polling study reflected a strong likelihood for current Detroit congressman John
Conyers to retain his congressional seat. The overall results for all five candidates are
listed below:
The 2012 Michigan 13th
Congressional District Democratic Primary election will be held
in August. Who are you most likely to vote for in the primary for US House of
Representatives?
District - Wide Aggregate Results – 643 Respondents MOE +/- 3.84%
Current Democratic Congressman John Conyers: 48.21%
State Senator Glenn Anderson: 20.68%
State Senator Bert Johnson: 2.33%
State Representative Shanelle Jackson: 4.51%
Westland School Board member John Goci: 2.02%
Undecided: 22.24%

______________________________________________________________________________________
28
48%
21%
2%
5%
2% 22%
John Conyers
Glenn Anderson
Bert Johnson
Shanelle Jackson
John Goci
Undecided
55%
21%
10%
13%
4% 0%
John Conyers
Glenn Anderson
Bert Johnson
Shanelle Jackson
John Goci
Undecided

______________________________________________________________________________________
29
As reflected in the final results, Congressman Conyers support was consistent across the
new 13th
Congressional district. Despite representing a district with a majority of voters
being non-Detroiters, Conyers was able to add to our polling data level of support by
picking up 7.07 points (31.79% of the undecided universe) from the undecided voters
(48.21% to 55.28%). Senator Anderson actually lost ground in his base, dropping from
20.68% in our poll to 18.14% of the actual vote (12.28% of Anderson’s base).
Representative Jackson benefits the most in the shift of undecided voters, picking up 8.02
points or 36.06% of the universe. Representative Johnson gained 7.65 points (34.40% of
the undecided universe) and School Board member Goci gained 1.82 points (8.18% of
the undecided universe) from the undecided voters.
C. Congressman John Conyers’ positive geographical cross tabular results.
Congressman Conyers’ positive voter tabular variances are as follows:
July 21 and 22, 2012 poll findings
Cross Tab
Grouping
John
Conyers
Glenn
Anderson
Undecided Shanelle
Jackson
Bert
Johnson
John
Goci
Detroit region
Respondents
65.00% 1.88% 21.88% 7.81% 2.81% 0.63%
Cross Tab
Grouping
John
Conyers
Glenn
Anderson
Shanelle
Jackson
Bert
Johnson
John Goci Undeci
ded
Detroit region
Respondents
69.14% 3.64% 15.68% 9.73% 1.60%
Undecided
gain
+5.14 +1.76 +7.87 +6.92 +0.97
The polling data suggested that Representative Jackson and Senator Johnson
were not able to connect with voters that were their base in past elections or
communicated a clear message to sway these voter bases away from
Congressman Conyers. The actual results proved the polling data to be accurate.
Congressman Conyers was able to win Detroit by 53.46 points over
Representative Jackson basically matching his 57.19 point margin lead in our
polling.

______________________________________________________________________________________
30
What is additionally surprising is the lack of impact that Senator Anderson is
having on the City of Detroit. Twenty and eleven hundredths percent (20.11%)
of the Detroit’s most likely voter population for the Congressional district
primary is non African American and Anderson was only receiving 1.88% of the
Detroit vote. After our poll was released through Fox 2 News, Senator Anderson
bombarded Detroit voters with 4 mailings and 7 automated calls. He also
deployed suburban African American leaders to help validate him with Detroit’s
African American voters. The last minute rush of cash and validators only
moved an additional 1.76 points for Anderson. His campaign has missed a strong
opportunity to capitalize on the diversity of the electorate in the City of Detroit.
D. Senator Glenn Anderson’s positive geographical cross tabular
groupings:
Anderson’s positive voter tabular variances are as follows:
July 21 and 22, 2012 poll
Cross Tab
Grouping
John
Conyers
Glenn
Anderson
Undecided Shanelle
Jackson
Bert
Johnson
John
Goci
Western
Wayne
Region
Respondents
28.52% 43.64% 21.99% 1.03% 1.37% 3.44%
Major 14
Wayne
County Cities
Respondents
24.12% 49.75% 20.10% 1.01% 0.50% 4.52%
Cross Tab
Grouping
John
Conyers
Glenn
Anderson
Shanelle
Jackson
Bert
Johnson
John
Goci
Undecided
Western
Wayne
Region
Respondents
35.59% 39.42% 8.43% 7.66% 8.68%
Undecided
vote gain
+7.07 -4.22 +7.40 +6.29 +5.24
Major 14
Wayne
County Cities
Respondents
29.29% 47.21% 7.44% 8.13% 7.59%

______________________________________________________________________________________
31
Undecided
vote gain
+5.17 -2.54 +6.43 +7.63 +3.07
Even through Senator Anderson is a Westland native, he was not able to build a
political firewall to counter his failure inside of the city of Detroit. Anderson lost ground
in both the Western Wayne region of the district and the major 14 cities (Westland,
Redford and Dearborn Heights). Anderson lost 4.22 points from our poll with actual
voters in the Western Wayne County cities. He also lost 2.54 points among voters in the
major 14 cities, even though he won 63.06% of the vote in Westland.
(FMW)B
Polling Accuracy
The (FMW)B
Undecided: 22.24%
Congressman Conyers’s margin in the (FMW)B
Congressman Conyers’s margin of victory in the August Primary election was 37.14
points.
Our polling findings for Congressman Conyers’s lead over Senator Anderson and his
opponents and their individual percentages of voter support were all consistent with the
actual election results. We correctly projected that Congressman Conyers would win the
Detroit region and Senator Anderson would win the Western Wayne and Major 14 cities
in the district. We also correctly projected the order of finish of all five candidates in the
Democratic primary.

______________________________________________________________________________________
32
Michigan’s U.S. Senate General Election matchup,
Republican Primary, and State House of
Representatives Exclusive Polling Study for
Fox 2 News Detroit.
July 28, 2012
Michigan registered and most likely November 2012 General election voters to determine
their voting and issue preferences on the United States Senate potential general election
match-ups, the US Senate Republican nomination. 1,046 respondents fully participated in
the survey. 656 respondents filtered to the Republican US Senate Nomination ballot test
question. The margin of error for this total polling sample is 3.03% with a confidence
level of 95%. The margin of error for the respondents who filtered to the Republican
Nominating Ballot Test section is 3.83% with a confidence level of 95%.
We have made weighting adjustments to the aggregate baseline responses based on the
following four groups who were underrepresented in our aggregate polling respondents:
 Male respondents – 41.18% of respondent universe versus 46% of (FMW)B
PVBA model projections for 2012 November general election.
 African American respondents – 10.08% of respondent universe versus 17.5%
of (FMW)B
 Voters ages 18 to 30 years old – 2.77% of respondent universe versus 16% of
(FMW)B
 Voters ages 31 to 50 years old – 13.10% of respondent universe versus 25%
of (FMW)B
The 2012 Republican Primary for United States Senate will be held on August 7, 2012. If
you plan on voting or have already voted the Republican Primary, who is your choice in
this election? (Only for respondents to question 3, selection 1 & 2)
State-wide Republican Primary Respondents Aggregate Results (Weighted to
projected age, gender & ethnicity of electorate) – 656 Respondents MOE +/- 3.83%
(Former U.S. Congressman Pete Hoekstra): 39.76%
(Republican Cornerstone Schools Founder Clark Durant): 23.59%
(Another candidate): 15.31%
(Undecided): 24.04%

______________________________________________________________________________________
33
Another Candidate/Undecided
Former U.S. Congressman Pete Hoekstra: 52.06%
Republican Cornerstone Schools Founder Clark Durant: 32.31%
Gary Glenn: 9.36%
Randy Hekman: 6.37%
Another Candidate/Undecided
38%
23%
15%
24% Pete Hoekstra
Clark Durant
Another Candidate
Undecided
52%
32%
9% 7%
Pete Hoekstra
Clark Durant
Gary Glenn
Randy Hekman

______________________________________________________________________________________
34
As reflected in the final results, former Congressman Hoekstra’s support was consistent
across the State. Even though Clark Durant did start to move closer towards Hoekstra, he
would have needed at least an additional month to build enough momentum to finish
closing the gap. Hoekstra was able to add to our polling data level of support by picking
up 12.3 points (51.16% of the undecided universe) from the undecided voters (39.76% to
52.06%). Clark Durant gained ground at a slower rate than Hoekstra, moving from
32.21% in our poll to 23.59% of the actual vote. Durant gained 8.72 points or 36.72% of
the undecided voter universe. Gary Glenn and Randy Hekman essentially split the other
candidate selection base from our July 28th
poll. They combined to win 15.73% of the
vote, which is an increase of 0.42 points from the 15.31% of poll respondents who
wanted another candidate.
Cross Tab Grouping Pete
Hoekstra
(July 28
Poll)
Clark
Durant
(July 28
Poll)
+/-
Margin
for
Hoekstra
Pete
Hoekstra
(August 7
election)
Clark
Durant
(August 7
election)
+/-
Margin
for
Hoekstra
Southeastern Michigan
Region Respondents
29.22% 18.72% +10.50 45.69% 34.54% +11.15
Central Region of
Michigan
36.54% 25.00% +11.54 52.16% 34.11% +18.05
Thumb Region of
Michigan
31.94% 23.61% +8.33 49.47% 37.67% +11.80
Region
41.46% 26.83% +14.63 51.78% 37.14% +14.64
Upper Peninsula Region 42.11% 15.79% +26.32 58.89% 24.92% +33.97
Major Democratic 6
County Cluster
31.29% 22.09% +9.20 40.76% 31.33% +9.43
Major Republican 7 CC 50.00% 26.03% +23.97 60.20% 25.55% +34.65
Major Swing 4 County
Cluster
31.08% 18.92% +12.16 53.31% 36.27% +17.04
Next 7 County Cluster 50.00% 16.67% +33.3% 52.57% 33.89% +18.68
Other 59 County Cluster 40.80% 24.14% +16.66 53.22% 33.48% +19.74
Wayne County 23.19% 17.39% +5.80 32.42% 28.52% +3.90
Oakland County 27.03% 21.62% +5.41 53.20% 37.28% +15.92
Macomb County 32.69% 17.31% +15.38 50.88% 38.20% +12.68

______________________________________________________________________________________
35
(FMW)B
Consulting Polling Accuracy
The (FMW)B
(Former U.S. Congressman Pete Hoekstra): 39.76%
(Republican Cornerstone Schools Founder Clark Durant): 23.59%
(Another candidate): 15.31%
(Undecided): 24.04%
Hoekstra’s margin in the (FMW)B
The General election was held on August 7, 2012. When all of the ballots were counted,
the final results are listed below:
Former U.S. Congressman Pete Hoekstra: 52.06%
Republican Cornerstone Schools Founder Clark Durant: 32.31%
Gary Glenn: 9.36%
Randy Hekman: 6.37%
Hoekstra’s margin of victory in the November General election was 19.85 points.
We correctly projected that Hoekstra would win all of the geographical regions of
Michigan and the political region clusters. We also projected within the margin of error
Hoekstra’s margin of victory statewide and across most of the geographical and political
region clusters of Michigan.

______________________________________________________________________________________
36
Michigan February Presidential Primary Election Polling
Study for
Republican Nominee
Automated Poll Methodology and Statistics
Aggregate Results
Cross Tab Results
February 27, 2012
Foster McCollum White & Associates, a Political and Governmental Affairs and
Organizational Development consulting firm based in Troy Michigan and Baydoun
Consulting, a political communications consulting firm based in Dearborn, Michigan
conducted a telephone-automated polling random survey of Michigan registered and
Most Likely Republican Primary voters to determine their voting preferences for the
2012 February Presidential Primary Election. This 5-question automated poll survey was
conducted on the evening of February 27, 2012. Of our 1,496 Respondents who
participated, 1,359 self-identified that they were certain or likely to participate in the
Primary election and 137 self-identified that they were not likely to participate in the
Primary election. We disqualified the not likely respondents from our reporting pool.
The margin of error for this polling sample is 2.66% with a confidence level of 95%.
Our polling study reflected a very close race between Mitt Romney and Rick
Santorum. The overall results for all four candidates are listed below:
Question 1: If the 2012 Republican Presidential Primary was held today, who would you
vote for, Mitt Romney, Newt Gingrich, Rick Santorum, Ron Paul or are you undecided?
State - Wide Aggregate Results – 1,359 Respondents MOE +/- 2.66%
(Mitt Romney): 37.90%
(Newt Gingrich): 8.31%
(Rick Santorum): 35.76%
(Ron Paul): 9.12%
(Undecided): 8.90%

______________________________________________________________________________________
37
The primary was held on February 28, 2012. The final results are listed below:
Mitt Romney: 41.10%
Newt Gingrich: 6.53%
Rick Santorum: 37.87%
Ron Paul: 11.63%
Other candidates: 2.88%
Mitt Romney
38%
Newt Gingrich
8%Rick Santorum
36%
Ron Paul
9%
Undecided
9%
Mitt Romney
Newt Gingrich
Rick Santorum
Ron Paul
Undecided

______________________________________________________________________________________
38
As reflected in the final results, Mitt Romney and Rick Santorum’s support was
consistent across the state. Their voter support from our polling was consistent with their
final tallies, along with the margin between the two candidates. Ron Paul’s last minute
push into Michigan provided him with a slight bump in support (9.12% in our final poll
to 11.63% election voter support). Newt Gingrich completely disappeared from a State
that he had been competitive in as late as December 2011 polling on Michigan.
Rick Santorum benefited from the wide geographical reach of the evangelical Christian
and Conservative voter base across Michigan. Santorum was in position to win Michigan
if he could maintain his geographical advantages. On Election Day however, we found
that Rick Santorum was not able to maintain his leads in the Northern Lower Peninsula
and Central region’s of Michigan. These defeats along with Santorum’s narrow victory in
the Thumb region of Michigan were significant factors to his narrow lost. The key cross
tab findings are listed below:
Cross Tab Grouping Mitt
Romney
(February
27 Poll)
Rick
Santorum
(February
27 Poll)
+/-
Margin
for
Romney
Mitt
Romney
(February
28
election)
Rick
Santorum
(February
28
election)
+/-
Margin
for
Romney
Southeastern
Michigan
46.30% 30.74% +15.56 45.05% 32.58% +12.47
Southwestern
Michigan
30.31% 40.77% (-10.46) 37.28% 44.40% (-7.12)
Mitt Romney
41%
Newt Gingrich
6%
Rick Santorum
36%
Ron Paul
12% Undecided
3% Mitt Romney
Newt Gingrich
Rick Santorum
Ron Paul
Undecided

______________________________________________________________________________________
39
Thumb Region of
Michigan
34.04% 46.10% (-12.06) 38.38% 40.13% (-1.76)
Central Region of
Michigan
33.95% 36.28% (-2.33) 39.95% 38.84% +1.11
Northern Lower
Peninsula Region
34.53% 37.41% (-2.88) 40.88% 38.73% +2.15
Upper Peninsula
Region
31.15% 26.23% +4.92 32.57% 44.71% (-12.14)
Major 17 County
Cluster
41.00% 33.70% +7.30 42.66% 36.15% +6.51
Major Democratic 6
CC
40.69% 34.14% +6.55 41.02% 35.80% +5.22
Major Republican 7
CC
33.73% 34.92% (-1.19) 39.28% 41.83% (-2.55)
Swing 4 County
Cluster
46.13% 32.53% +13.60 46.54% 32.28% +14.26
Next 7 County
Cluster
37.74% 47.17% (-9.43) 37.78% 41.86% (-4.08)
Other 59 County
Cluster
30.34% 37.64% (-7.30) 37.04% 42.22% (-5.18)
Wayne County 42.62% 24.81% +17.81 41.45% 33.24% +8.21
Oakland County 46.26% 29.44% +16.82 50.39% 29.01% +21.38
Macomb County 43.12% 34.86% +8.26 43.28% 34.56% +8.72
Michigan Congressional Districts
Michigan’s Republican Primary will award its delegates through a combination of
individual congressional district winners and overall State winner. This hybrid model will
allow a second place or third place state-wide contestant to pick up delegates by winning
one of Michigan’s congressional districts.
Cross Tab
Grouping
Mitt
Romney
(February
27 Poll)
Rick
Santorum
(February
27 Poll)
+/-
Margin
for
Romney
Mitt
Romney
(February
28
election)
Rick
Santorum
(February
28
election)
+/-
Margin
for
Romney
1st
Congressional
District
33.80% 30.99% 2.81 38.85% 39.79% (-0.94)
2nd
Congressional
District
30.25% 36.39% (-6.14) 35.49% 47.24% (-11.75)

______________________________________________________________________________________
40
3rd
Congressional
District
33.03% 35.78% (-2.74) 39.85% 41.83% (-1.98)
4th
Congressional
District
31.13% 40.57% (-9.44) 38.49% 41.91% (-3.42)
5th
Congressional
District
41.77% 41.77% 0 39.79% 38.98% +0.81
6th
Congressional
District
23.53% 46.08% (-22.55) 36.90% 43.14% (-6.24)
7th
Congressional
District
37.07% 37.93% (-0.86) 38.93% 40.10% (-1.17)
8th
Congressional
District
42.20% 33.03% 9.17 44.44% 33.13% +11.31
9th
Congressional
District
48.39% 33.33% 15.06 43.83% 32.26% +11.57
10th
Congressional
District
33.62% 42.24% (-8.62) 40.26% 37.78% +2.48
11th
Congressional
District
50.00% 29.03% 20.97 50.84% 29.44% +21.40
12th
Congressional
District
42.37% 33.90% 8.47 40.24% 36.65% +3.59
13th
Congressional
District
36.67% 30.00% 6.67 30.21% 36.04% (-5.83)
14th
Congressional
District
57.14% 16.07% 41.07 46.87% 27.37% +19.50
The last week of the Republican primary showed the continual fluidity in the Republican
field. In our initial poll on February 16th
, Rick Santorum was in the lead in 8
congressional districts and in a statistically relevant position to win 12 of Michigan’s 14
Congressional districts. Our next poll on February 23rd
found that Mitt Romney has
regained the advantage in 10 of Michigan’s 14 congressional districts. Based on our final
poll data, Romney was leading in 7 congressional districts. Santorum is leading in 5 and
two are in a statistical tie. We projected that the following districts could have change
hands on Election Day:
1st
district
2nd
district
3rd
district
13th
district
Our polling model correctly identified two of the four Congressional districts that swung
based on Election Day turnout, the 1st
and 13th
Congressional Districts. Both of those
districts swung from Mitt Romney to Rick Santorum. We also correctly identified the
battleground important of the 5th
Congressional district. Our final poll found that the

______________________________________________________________________________________
41
district was tied at 41.77% for Romney and Santorum. Romney won the 5th
Congressional District by 452 votes. That margin ultimately prevented Rick Santorum
from winning the majority of delegates in Michigan by having won 8 of the 14
Congressional districts.
(Mitt Romney): 37.90%
(Newt Gingrich): 8.31%
(Rick Santorum): 35.76%
(Ron Paul): 9.12%
(Undecided): 8.90%
Romney’s margin in the FMW & Baydoun Consulting poll was 2.14 points.
The primary was held on February 28, 2012. The final results are listed below:
Mitt Romney: 41.10%
Newt Gingrich: 6.53%
Rick Santorum: 37.87%
Ron Paul: 11.63%
Other candidates: 2.88%
Romney’s margin of victory in the February Primary was 3.23 points.

______________________________________________________________________________________
42
We correctly projected the following election occurrences:
 Eleven (11) of the 14 Congressional Districts.
 Southeastern Michigan, Southwestern Michigan and Thumb Region of Michigan.
 The major 17 county cluster and each sub group (Major Democratic 6,
Republican 7 and Swing 4 counties).
 The Next 7 County and Other 59 County cluster.
 Wayne, Oakland and Macomb Counties
Our polling findings for also identified the gap of support among Catholic voters for Rick
Santorum. The Republican Catholic voting base is heavily centered in Southeastern
Michigan. This gap proved to be fatal to the Santorum campaign. Our analysis below
highlights the missed opportunity for Rick Santorum, a Roman Catholic, with catholic
voters. If Santorum had been able to move 8 points of support among catholic voters in
Wayne, Oakland and Macomb Counties, he would have been able to potentially win
Michigan by 1,142 votes.
Counties
Santorum
(loses
Catholic
Vote 43%
to 37& as
consistent
with
State
Average)
Romney
(Wins
Catholic
Vote
43% to
37& as
consistent
with
State
Average)
Margin
of defeat
for
Santorum Counties
Santorum
(Wins
Catholic
Vote 45%
to 35% in
these 3
counties
only)
Romney
(Loses
Catholic
Vote
35% to
45% in
these 3
counties
only)
Margin
of defeat
for
Santorum
Shift
of
votes
Wayne 38,890 48,498 -9,608 Wayne 44,500 42,888 1,612 5,610
Oakland 42,465 74,030 -31,565 Oakland 49,340 67,155 -17,815 6,875
Macomb 30,218 37,838 -7,620 Macomb 34,293 33,763 530 4,075
Net total 111,573 160,366 -48,793 Net total 128,133 143,806 -15,673 16,560
Other 80
Counties 265,580 248,765 16,815
Other 80
Counties 265,580 248,765 16,815
Statewide
Total 377,153 409,131 -31,978
Statewide
Total 393,713 392,571 1,142

______________________________________________________________________________________
43
Statistical modeling and weighting methodology
Our polling call lists are weighted to the historical weights for age, gender, race, region and
congressional district area. Our list is also comprised of voters with previous voting histories in
Presidential, state and local elections. We include the moderate and low performance voters, but
the call files do contain a significant portion of voters who have a likely history to participate. We
do not call voters who have never participated in elections but are registered.
Our PVBA model reviews election statistics for age, gender, voting participation pattern, gender
and socio-economic factors to determine the likely voting universe for an upcoming election. Our
turnout models are based on state based historical turnout statistics provided by the municipal and
county clerks and secretaries of state’s office of a state for age, gender, party, ethnicity and voting
method (early, absentee, poll location) instead of exit polls. We trust the reliability of the election
statistics from the clerks’ offices to give us value data reads on future elections. For example, our
PVBA model for the Primary election in Wayne County Michigan (the largest voting county in
Michigan.) was within 0.316% of the actual August 7, 2012 primary. We projected a total county
turnout of 246,299 voters for all 43 communities including Detroit and actual turnout was
245,450 (after spoiled ballots were discounted for partisan contest).
The reason we take the historical data for a state is to give us a baseline for each precinct within
the state and then build models up from there. We work to identify solid trends of turnout over a
series of primary and general election contest so that we can remove outliers within turnout, age,
gender, partisan (if collected) and ethnicity and determine the true participation base for that
precinct. We can then project out for the variable election conditions (type, advertising impact,
voter mobilization, outlier ballot issue impact, etc.) that allow us to determine our high moderate
and low performing turnout and voter models.
An additional example for using historical election statistics is as follows:
Michigan has a historical Presidential participation variance of 18.4% from the baseline voter
model and has an -18.08% historical gubernatorial participation variance. The swing is equal to
2.3 million moderate and low performance voters in Michigan for every given Presidential
election who primarily leave the participation rolls for the gubernatorial election. The difference
between a Governor Snyder and Governor Bernero was the complete absence of the low
performance voters and a 15% participation rate among moderate participation voters. If Bernero
gets the participation rate of Granholm’s re-election in 2006 (85% moderate performing voters
and 25% low participation voters) He defeats Snyder by 200,000 votes and wins 40 counties. This
model allows us to help our political clients understand their election audience more clearly than
exit polling. We then use it in assessing our polling models to help us gauge data quality and
participation models.
When we call through the list, we report the demographics of the respondents without weight. If
our demographics match the likely voter demographics for the polling study. If there are
underrepresented groups within our aggregate respondent universe, we use our weighting model
to adjust for their representative weight and the groups reflected polling preference for the
baseline questions. We still will report the un-weighted demographics of our respondents because
they reflect the prevailing interest level of the voting groups at the time of our polling survey.

______________________________________________________________________________________
44
Based on the respondent universes, we will determine if weighting adjustment need to be made
for underrepresented demographic groups. If our respondent demographics match our PVBA
model, we will not make any weighting adjustments and will report our findings unweighted. If
respondent demographics do not match our PVBA model, we will make weighting adjustments
for the topline baseline questions and adjusted aggregate data findings.
Data Analysis Statement -
The data has been separated analytically into cross tabulation results that are statistically
significant with respect to Michigan General Election cycle statewide Any sectional analysis
within the aforementioned categories can be useful when inferring strengths and weaknesses and
possible strategy.
For the assessment of individual cross tabulation categories, we use a correlation
coefficient model based on the Pearson r correlation, also called linear or product- moment
correlation. Pearson correlation (hereafter called correlation), assumes that the two variables are
measured on at least interval scales and it determines the extent to which values of the two
variables are "proportional" to each other. The value of correlation (i.e., correlation coefficient)
does not depend on the specific measurement units used. Our proportional model for correlating
the statistical relevance of a geographical region, age grouping or congressional district is based
on the random proportionality of our respondent pool to the specific proportionality of the
group’s weight to the aggregate model. The correlation coefficient (r) represents the linear
relationship between these two variables (aggregate and cross tabular category).
We project that any review of the polling report can allow for the statistical relationship
between the aggregate and cross tabulation margin of error for the reported clusters. All polling
samples are pre-weighted for gender and ethnicity based upon Foster McCollum White Baydoun
Predictive Voter Behavior Analysis Model for historic General Election demographics throughout
Michigan. We strive to adhere to the principles and standards of the National Council on Public
Polls and the American Association for Public Opinion Research in the gathering and reporting of
polling data.

______________________________________________________________________________________
45
Methodology
, a national public opinion polling and voter analytics
consulting firm based in Michigan and representing the combined resources of Foster McCollum
White & Associates (Troy Michigan) and Baydoun Consulting (Dearborn Michigan) conducts
telephone-automated and live telemarketing call polling random survey of Michigan and Florida
registered and most likely voters for primary and General elections to determine their voting and
issue preferences on specific and unbiased nature.
The population surveyed consisted of a sample of traditional Michigan and Florida high
participation registered voters and voters that fit Michigan and Florida Primary and General
Election voting patterns. The majority of these voters have participated in a significant majority
of the available primary and general election and odd year municipal and county elections in
Michigan since their registration. Additionally, our call file does allow for random moderate and
low participation voters to be included in the sample. Our call file was randomized to allow for
the maximum range of participation and randomization.

(FMW)B Polling Accuracy Analysis Report

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Andere mochten auch

Andere mochten auch (12)

Ähnlich wie (FMW)B Polling Accuracy Analysis Report

Ähnlich wie (FMW)B Polling Accuracy Analysis Report (20)

(FMW)B Polling Accuracy Analysis Report