Understanding Political
Polling
Contents:
Polls and their Purpose
How Online Surveys Work
10 Myths of Political Polling
The development of a polling
plan is unique for every campaign. Is there going to be a
primary? Are you a challenger or an incumbent? Is there a unique
development in your district or state? What is the political
makeup of your state or district? These factors and more help
determine what types of polls you conduct and when they are done.
The following is a description of the types of polls and
purposes for conducting them. A specific plan for polling in your
campaign will be developed after your needs are assessed and
budgetary concerns are addressed.
Polls and their Purpose
Benchmark Poll: The benchmark poll is a
comprehensive survey that serves as your campaign's guide and
playbook. It should be a thorough review of public figures, the
candidate's strengths and weaknesses, your opponent's strengths and
weaknesses, issues, messages, potential messages, etc. The information
derived from the poll and the strategic analysis that accompanies it
serves as the basis for the campaign's strategic plan and its tactical
playbook.
Benchmark polls generally
range between 15 - 20 minutes in length and can be longer. Benchmark
polls should be as comprehensive as your campaign's budget allows.
Trend/Issue Poll: Trend polls are conducted
within the context of your active campaign. They serve as a
progress chart of your campaign. It shows if you are meeting your
goals and objectives. It can determine if strategic or tactical
changes are needed. Trend polls may also serve as an indicator of
changing dynamics in your campaign. Campaigns that catch the changes
first will have a strategic advantage.
Trend polls range between 10
- 15 minutes in length.
Tracking Poll: Tracking polls are short
polls designed to help you maneuver through the final weeks of the
campaign. They give you information vital to the refinement of
your strategy and tactics. Your field coordinators will use tracking
polls to adjust plans and assignments accordingly and your media team
uses them to alter TV and radio rotations or adjust mail
targeting. They are not intended for the development of
completely new plans, strategies and tactics.
Tracking polls range in
length from 5 - 10 minutes in length.
Vulnerability/Viability
Survey: These preliminary polls are
used to determine a candidates’ viability or opponents’
vulnerability. They are designed to give focus and discipline to
a campaign in its very early stages. These surveys assess the
campaign's starting point and may identify key issues, test potential
messages and determine where a campaign should focus its resources.
They may serve to prevent early mistakes as well. They are
conducted very early in the campaign, usually at a point that is
considered too early for a comprehensive benchmark poll.
Vulnerability/Viability polls
range between 10 - 15 minutes in length.
Focus Groups: Focus groups are
"qualitative" research tools used to refine messages and
strategies. They are used to evaluate and probe deeply into "how
and why" people feel about a particular issue or subject and how they
respond to factors related to the topic. Focus groups can be used
to test specific mail pieces, TV or radio commercial or other
communications. Ultimately, their purpose is to refine the
message you intend to communicate. Fako & Associates partners
with professional qualitative researchers to organize and conduct the
groups. Fako & Associates performs all strategic, analytical
and consulting components of focus group projects.
Online Surveys: Online Surveys are an
example of how the polling industry has evolved to meet real-world
realities. Some demographics present a challenge for traditional
phone driven polling. Online surveys can reach difficult to
contact demographics with statistical reliability that rivals
traditional phone interviewed surveys. Online surveys are,
however, not a replacement for traditional polling in all instances.
How Online Surveys Work
How
it Works: F&A
hosts the online survey on secure servers. E-mail invitations are
sent to appropriately targeted panelists, selected according to take
the survey.
Panel
Sourcing: Panel
recruitment uses a controlled mix of both online methods (e.g. solo
e-mail invitations and other targeted online modes) and offline methods
(e.g. physical post-card invitations, direct mail inserts, etc.), and
has resulted in a geographically and demographically diverse panel.
In order
to remain as balanced as possible, panel members have been recruited
from a diverse set of consumer and business-to-business sources using a
"by invitation only" approach. All panel establishment methodologies
employed are 100% opt-in and fully comply with CASRO guidelines.
By
employing the "by invitation only" approach, the panel can: 1) achieve
higher levels of panel normalization, and 2) effectively avoid the
"professional respondents" that affect some online research panels.
More
Normalized: Certain
panel companies cannot achieve the same level of normalization since
their "open" recruitment methods do not allow them to determine the
demographic make-up of their panel until after members have joined.
Less
"Self Selection" Bias: Panels
that use "open" recruitment techniques such as multi-panel
co-registration sites are allowing large-scale self-selection biases
into their panels, attracting "professional survey takers" and causing
a widespread amount of membership duplication across many other online
panels. To reduce the potential for self-selection bias, panelists
considered in this estimate may join only if they are first invited.
After an individual responds to an invitation and joins the panel, he
or she completes a 300+ item member profile, which collects information
about demographics, interests, life events, health ailments, various
product purchase intent data and more.
10 Myths of Political Polling
Campaigns & Elections, August, 1993 by Robert G.
Meadow, Heidi von Szeliski
1. "I know my constituents, so
I don't need a poll."
Legislative districts are large. No matter how involved you are with
the community, it is possible to know only a small fraction of the
constituents. Many candidates and public officials keep in touch
with their friends, neighbors, business associates, supporters and
community activists as well as the small number of people who make
their positions known through letters or phone calls. But the opinions
of these individuals are often quite different from those less involved
and less interested in politics -- in other words, the very people who
make up the bulk of the district. Failing to understand that the people
with whom you are in touch are often not representative of the district
at large can seriously jeopardize a campaign. A poll is conducted with
a random sample of likely voters, not just those activists or those
with who you are in contact.
2. "I don't need a poll to tell me
what to think or what to do."
In a representative democracy, candidates and
public officials are elected to reflect the views of the people they
represent. The best way to find out those views is to have a reliable
measure of public opinion -- a poll.
Pollsters do not tell a public official or
candidate what their positions should be on the issues, or how they
should respond to public policy questions. A pollster's job is not to
tell a candidate what to think. But a responsive and responsible public
official must know the range of opinions within his or her district. A
good poll can provide that information.
A pollster will not try to change your opinions
based on the results of a poll, but you want to modify your vocabulary
or emphasize different points as a result of a poll. A public opinion
survey can guide a campaign with necessary information, such as what
issues to emphasize, as well as provide the strategy to communicate the
reasons behind positions and decisions in such a way that more support
for your views might be generated.
3. "As soon as I hire a pollster,
I'll never see him or her again."
A pollster should be committed to the campaign,
and have experience in election and public policy polling, not market
research. The pollster should be attentive to the campaign, working on
only a limited number of campaigns at once. Think about it. If a
pollster is working on seven U.S. Senate races, three gubernatorial
campaigns, 16 congressional races and a dozen state legislative races
during the same election cycle, how much attention do you think your
campaign will receive? Expect the attention of the principals of the
firm selected, not junior analysts who have limited experience and no
involvement with this particular campaign.
While you need to benefit from the experience of
your pollster, expect the race to be viewed as distinct from others. A
poll should be custom-designed for your district, not some
off-the-shelf questionnaire used in countless other races.
4. "I already have a consultant -- I
don't need another one to play with numbers."
The pollster, as a member of your strategy team,
should be the keeper of the message. The pollster should play an
ongoing role to make sure the message is on track and on target at all
times. All voter communication should convey a consistent message, so
the pollster should be available to review speeches, direct mail,
television and radio advertisements, and other campaign messages. Of
course, the pollster is not the campaign manager, nor is the pollster
the speech writer, or the direct mail, or television specialist.
5. "A poll costs a lot of money. I'd
rather spend my money on other things."
Most campaigns budget about 5 percent to 10
percent of the overall campaign budget for research. The costs of a
research program are based on the size and scope of the campaign, and
there are two key variables -- the length of the questionnaire and the
number of people interviewed (sample size). A typical benchmark poll
contains about 75 response items, and usually takes about 20 minutes to
complete.
For a high-quality, professional public opinion
survey, expect to pay anywhere from $12,000 to $17,000 for a sample
size of 400. Shorter, follow-up polls, taken midway through a campaign,
are less expensive. Overnight tracking polls taken for fine-tuning near
the conclusion of a campaign, cost even less.
6. "All I want to know is what
portion of the vote I am getting."
Too many people think that polling is just about
the trial heat question: "If the election were held today, for whom
would you vote?" Everyone could wait until election day to find that
out. A poll should be designed to develop a strategy, to find the
winning arguments for the campaign, its strengths and weaknesses, as
well as the strengths and potential vulnerabilities of your opponent.
A poll should be conducted early, before the
campaign begins in earnest, so that the message, the announcement
speech, and early walking brochures are all guided by the poll results.
It is better to have the message designed right the first time, than to
spend the rest of the campaign correcting it. Perhaps most importantly,
a poll will be able to tell which messages are right for which voting
constituencies, such as women, seniors or those in a particular
geographic area.
7. "Let's use campaign volunteers to
take the poll."
Volunteers are a precious resource and they must
be used most effectively -- to persuade voters that the candidate
should be elected. Comprehensive interviewing requires trained
interviewers skilled in administering the questionnaires and who
maintain their neutrality. Supportive volunteers may tend to overstate
the case by wanting to give you the information they think you want to
hear. They may argue with those supporting the opponent because, as
volunteers, they want to convert all voters to your side. This is why
they should be used for persuasion calling, voter identification, and
phone banking.
The accuracy of the data collected by volunteers
cannot be, and should not be, guaranteed. Donors who must consider the
viability of your campaign simply do not trust volunteer polls; they
depend on the reputation of the polling firm. Some campaigns that
attempt to use volunteers find that the interviewing process (which
should take three to four days) can take up to six weeks because of
scheduling difficulties. Volunteers often "burn-out" when interviewing,
and are then not available to the campaign during the critical
get-out-the-vote phase of the campaign.
8. "Polling is a luxury I really
can't afford."
You can't afford not to poll. First, if the
message is wrong, or given to the wrong people, you limit the chances
of success. Second, because in large districts a candidate cannot
communicate with all voters, there is a need to know which voting
groups need special attention, and which voters do not need as much
attention, either because you will always have their support or never
have their support, no matter what. A comprehensive benchmark poll that
analyzes voting groups and the messages to which they respond will save
money in the long run by enabling the use of a voter communication
budget most effectively.
9. "All polling firms are alike."
No, they are not. In fact, they vary widely
along a number of dimensions. Some polling firms are commercial market
research firms that dabble in politics, do not have the experience, and
simply do not understand political polling. Other firms are political
polling specialists, but are retained by so many campaigns at once,
that the principals of the firm simply cannot give the attention to
each campaign that it requires. Some firms simply throw a book of
computer printouts at the campaign or give a one-shot oral briefing and
do not provide any written strategic analysis or campaign
recommendations. And some firms charge additional fees for a written
report, while some charge monthly retainers for ongoing consultation.
They move on to the next campaign and they are not heard from again.
The very best firms do several things. First,
they take on only the number of campaigns in which they can provide the
level of attention the campaign requires. Second, they give campaigns
full, comprehensive, written reports that serve as the basis for the
strategic plan. These reports look at each question and each
demographic group so that the messages for each group are clearly
identified. Third, they remain part of the campaign strategy team, and
are available to the campaign through election day and beyond.
10. "Pollsters just give candidates
the information they want to hear."
Good pollsters do not. The campaign team needs
to understand the political landscape, not stroke the candidate's ego.
If a pollster "cooks the numbers," or asks biased questions, a
candidate has the wrong pollster. Remember, the purpose of a poll is to
provide the candidate with accurate information to hone the message and
communicate with the voters. The one place a candidate cannot tolerate
"yes" people is when hiring a pollster. A pollster who is known for
fabricating or manipulating data will have no credibility with a fund
raising base.
Campaigns & Elections, August, 1993 by Robert G.
Meadow, Heidi von Szeliski
Robert G.
Meadow is president, and Heidi von
Szeliski is vice president, of Decision Research polling firm with
offices in San
Diego, Los Angeles, and Washington, D.C.
|
