Read an Excerpt
Public Opinion Polling
A Handbook for Public Interest and Citizen Advocacy Groups
By Celinda C. Lake, Pat Callbeck Harper
ISLAND PRESSCopyright © 1987 Montana Alliance for Progressive Policy
All rights reserved.
Before We Begin ...
In 1984, polling became a $4-billion-a-year business. Polls help us determine everything, from which dog food people buy to which political candidate they will vote for.
Jimmy Carter probably knew on election eve in 1980 that he would not be reelected because his pollsters had done a good job of predicting.
The editors of Literary Digest magazine in 1936, foretelling their demise because of their prediction on the FDR-Landon race, probably wished they had had such good pollsters working for them — they had predicted that Landon would win with 60% of the vote!
Polls are a tremendously valuable tool in today's political world. Our goal with this guidebook and associated software package is to make this tool fully accessible to public interest groups. You need to know how to use this important tool. But first, you need to know what's in this book and how to use it.
THE USES OF THIS BOOK
There are two major uses for this guidebook:
1. To help you plan and complete a professional poll.
2. To help you become a wise consumer of polls for your own and your organization's best interests.
This book presents hands-on information on how to plan, administer, and analyze a poll. And because polls can be used against you, this book also helps you to analyze the sampling, interpretation, and question-wording of polls conducted by other orgnizations.
The text includes examples, checklists, warnings of possible pitfalls, and lists of additional material and human resources to help you in your polling efforts. Samples of most of the materials you will need to produce a poll are also included.
The software package POLLSTART, its manual, and this book give you a complete package for polling that takes you from sampling through analysis. More information on POLLSTART is found in Appendix F. This book also is designed to be used on its own without the software package.
Some polling manuals tell you just enough to conduct your poll but don't explain the theory or logic behind this tool. This often means that any adjustments you may make can unwittingly endanger the quality of your poll. Other textbooks are too complex, detailed, and lengthy to be used well and easily by public and citizen interest groups. This book strikes a middle ground.
HOW TO USE THIS BOOK
You might ask, "Why not just pick a chapter and start polling?" Since the results of this important scientific tool could be very important to your work and your organization, we suggest that you read the book through completely. A good understanding of polling and the choices you must make before you begin will reap benefits for you later.
The chapters will walk you through the rationale and the steps in polling. Most chapters end with a checklist of questions that you need to answer before going on to the next stage. Suggestions of community resources for additional help are also noted.
If you are using this book primarily as a consumer of polls, you will want to concentrate on the chapters on sampling, questionnaire construction, and analysis. By using the suggestions and checklists provided in this book, you are certain to become a more skilled user and consumer of polls to your benefit and that of your organization.CHAPTER 2
Introduction to Polling
LET'S START AT THE BEGINNING — WHAT IS A POLL?
A poll is a systematic, scientific, and impartial way of collecting information from a subset, or sample, of people that is used to generalize to a greater group, or population, from which the sample was drawn. A poll is not designed to persuade or identify individuals — there are cheaper and more efficient ways of doing that (telephone canvass, for example). Confusing these goals with those of a poll can seriously bias the information you receive. A poll also is not intended to describe any one individual in depth. Again, a case study is a cheaper and more efficient way to do that. A poll is a measurement at one point in time that reveals attitudes, behaviors, beliefs, attributes, and the interrelationship of all of these parameters. These generalizations can then be extended to the larger society.
In a poll, information is obtained in a scientific, controlled way from a selected subset of people.
A properly selected subset enables you to generalize your findings reliably to a greater population (without talking to everyone) after attributing a known margin of error to the sampling. Careful interviewing, questionnaire construction, and analysis also minimize other forms of error that are difficult to measure.
Because a poll is not designed to influence or persuade people, you should never identify your organization or goals in such a way as to influence your respondents' answers. The interviewing should be kept as neutral as possible.
WHAT ARE THE VALUES OF POLLING?
To give you an idea of how valuable a poll can be to your organization, we have listed some of the kinds of information a poll can help you obtain. Polls can help you determine:
 What people are thinking — what they see as important problems, what their opinions on policy and issue questions are, what tradeoffs they are willing to make in policy and budgetary decisions, what they think are appropriate arenas for public involvement, and how they want to see resources allocated.
 What people know — what political figures and groups they are aware of, which issues and arguments about issues are important to them, and what factual information they have.
 How people perceive issues and political objects — how they evaluate their political figures' and institutions' performances, what emotional attachment they have for groups and individuals, for whom they would vote, and what reactions they have to certain slogans or information about political figures and issues.
 Characteristics of people — what their social and political characteristics are, how interested they are in a topic or event, and where they get their information on different topics.
The most valuable aspect of polling, however, is not just looking at these parameters individually, but rather linking them — seeing who feels what, where they live, how they can be reached, what points are important to them, and what issues are linked for them.
Examining the covariance, or linkage of opinions with other attitudes and social and political characteristics, reveals why certain beliefs are held, as well as what beliefs are held. This information can be used to develop strategies for influencing public opinion, political events, and political figures. You can target groups by issue and demographic characteristics. You can discuss or confront issues with an awareness of how those issues are actually perceived by different groups.
Polls have tremendous internal value for your group in developing strategies and assessing the impact of strategy and events. They also can be used externally. Polls can be released as the basis for news items; they also can be used to obtain money, political support, and media attention by demonstrating the viability of your ideas or candidacy. Polls can be used to influence the behavior of public officials or rally the support of your volunteers. They can be a service you exchange with allied groups or political candidates ... and much, much more.
There are many situations in which a public interest group might want to conduct a survey using volunteers. For example, you might want to contribute this as an in-kind service to a candidate who supports your issues. You might use a poll to target a bad incumbent, using the results to develop strategies for recruiting a challenger and defeating the incumbent. You might want to bring an issue to people's attention, but before you do so you may need to measure what people believe and why they hold those beliefs.
After a press or grassroots strategy, you might want to measure whether and how much attitudes have changed to plan your next move and assess your success. Survey data may make news and persuade officials, even when the survey is conducted by volunteers.
This book will help you not only to conduct polls, but also to analyze the results of polls conducted by other groups. Other groups may use their poll results to undermine your efforts. It is important that you be able to evaluate the validity of their results to determine your own strategy and counter their efforts. Because different results frequently are released by different groups, you should be able to evaluate the wording of the questions, methodology, and sampling.
Only with a basic understanding of polling methods can you be an informed reader and purchaser of polls. Today, everyone is affected by polls, and this book is an important resource regardless of whether we're conducting or being influenced by them.
WHAT TYPES OF POLLS ARE THERE?
There are four basic types of polls: in-depth surveys, short polls, tracking polls, and panels.
In-depth surveys, the most common type of poll, are 20-60 minute surveys that assess public opinion on one or more topics in depth. This type of survey sometimes can serve as a benchmark when followed by short polls, 10-15 minute surveys that assess change over time and the impact of events and strategy.
A tracking poll is used to assess a rapidly changing trend occurring over a short period of time. This type of poll, which often is used at the end of a political campaign, for example, asks a few key questions of a small sample (100-200 people) in 5-10 minutes at short intervals (for example, every other night).
If you are interested in understanding change and why it occurred, you need to conduct a panel poll, in which you interview the same people at two different points in time. Asking the same individuals similar questions at two different times enables you to make some assessments regarding changes in opinion, as well as determine some of the reasons for the change. Panel polls often are used to assess the effectiveness of a public education campaign to influence opinion or knowledge on an issue.
DEBUNKING SOME MYTHS ABOUT POLLS
Myth No. 1: "Volunteers cannot conduct good polls."
Good polls can be conducted by volunteers at almost every stage. As in any volunteer task, however, the volunteers must be well trained, understand what they are doing, and be well supervised.
It is essential that someone with a good understanding of polling be in charge of the poll. It's also helpful to identify some polling experts in the community who can answer the questions that invariably arise.
The most successful polling efforts frequently are achieved when the volunteers have a sense of ownership for the task. Volunteers should be recruited for the task in the early stages, see the poll through to the end, be delegated distinct responsibilities, and be supplemented by other volunteers as needed.
Myth No. 2: "Polling is too complex and costly for public interest groups."
In fact, with donated services and local calls, a poll can require little actual cash outlay. Money will be needed to pay for photocopying questionnaires and lists of registered voters, getting telephone directories, or using a computer to obtain telephone numbers. The major (and potentially costly) tasks of typing, sampling, interviewing, questionnaire writing, planning, analysis, data management, and training, however, theoretically can all be done by existing staff or volunteers. These tasks will be time-consuming, of course, especially if done by volunteers. A key factor in doing an in-house poll is to allow ample time, particularly for the first stages.
If you are planning to conduct anything lengthier than a 10-minute poll, you also may have to pay for data analysis.
The POLLSTART software package that can be used with this book is designed for direct data entry from your questionnaires and has safeguards against "invalid punches." This software is designed for easy, accurate data entry by volunteers and can be used on most personal computers. Given the widespread availability of personal computers, someone in your group probably will have access to a system with the equipment needed to print all of your poll analyses.
As discussed in Chapter 5 on Interviewing, interviewing from a central location is infinitely preferable to having volunteers operate from their own homes. Law offices, banks, construction firms, union halls, or even your own office are good places to start looking for a telephone bank. Keep in mind while searching that you will be calling after working hours. To avoid long-distance calling, you may want to establish telephone centers in several cities with supervisors for each.
To conduct a quality poll, you need to plan ahead. You will need someone to supervise the poll, someone to coordinate volunteers, and someone to take care of data entry. A volunteer poll usually requires a core of three to five people, at least 20 people willing to do the interviewing, and two to three volunteers for data entry. Although advance planning is needed, these volunteer requirements are usually well within the reach of most public interest groups.
Myth No. 3: "Anyone can do a 'quick and dirty' poll."
Polling is not complex, but it is a precise exercise, with many steps leading to the finished product. It requires detailed planning, careful supervision, attention to detail, and a great deal of time. During some stages, it also can demand substantial volunteer time.
No one can do a perfect study, but this book is designed to help you conduct the best poll possible. This book helps you identify your goals and resources and recognize the tradeoffs inherent in the decisions you must make. This book should give you a thorough enough understanding of the polling process to: (1) implement the necessary steps efficiently; (2) conduct a survey at minimal cost; and (3) be fully aware of the nature of what you are beginning. Well-organized polls are enormously useful to an organization, as well as interesting for the people involved. Polls are also a great way of using the talents of the people who support your organization.CHAPTER 3
WHAT'S IT ALL ABOUT?
The first step in any poll is to determine the study objective — what information you want and how you plan to use it. Your objectives should be clear, and every question should contribute to them. Think about what information you want, how you plan to use it, and then consider what questions you need to ask to both understand and utilize your findings. Study objectives may include questions that help you:
 Separate people into political, social, and issue target groups.
 Determine how best to use the media to reach your groups.
 Identify "intervening attitudes," that is, factors that influence the relationship you are studying, so you know something about how people form their opinions.
As soon as you know what you want to ask, it will become clear who you want to ask — that is, your target population. Your target population may be all registered voters, all people in a geographic area, all likely voters, all members of an organization, and so on.
Planning the who, what, where, when, and how of the survey is called laying out the study design . A detailed plan will help you determine what people you need at what times and what tasks have to be completed by what time so that the next task can begin. The case study presented in Appendix E will help you plan your study. This book also goes into more detail about each of the steps in subsequent chapters. Figure 3-1 is a flowchart showing the steps in a poll.
Once you know who you want to interview and what you want to ask them, you are ready to determine what type of poll you want to use. There are three possible methods of polling: by telephone, in person, and by mail. All three methods of polling have tradeoffs in terms of cost, types of resources needed, coverage of the population, response rate (the number of eligible sampled people who generate completed interviews), types of questions that can be asked, control over the interview, ease of administration, types of training needed, and type and amount of error incurred.
Excerpted from Public Opinion Polling by Celinda C. Lake, Pat Callbeck Harper. Copyright © 1987 Montana Alliance for Progressive Policy. Excerpted by permission of ISLAND PRESS.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.