How to Sample in Surveys

Arlene Fink

• Publisher: SAGE Publications, Inc
• Year: 2002
• Print ISBN: 9780761925774
• eBook ISBN: 9781483389295
• Edition: 2
• Topic: Social Sciences
• Subtopic: Social Science Research & Methodology
• Index: Social Sciences

"Good coverage of target populations and samples, and statistics of sampling. Logical progression through the material with good use of examples."

--Carol J. Lancaster, Medical University of South Carolina

How much is enough in your sample? What is the difference between a research question and a survey question? By exploring the answers to these questions, this book shows readers how to specify inclusion and exclusion criteria in a sample, select the appropriate probability and non-probability sampling methods, understand the sources of error in sampling, and calculate the response rate. New to this edition is coverage of sample size and power so that the reader can better understand the logic in determining the sample size to detect a difference if one exists. Filled with checklists and guidelines, this book will enable readers to select and use the most appropriate sampling methods for their survey.

"Fink provides detailed descriptions of inclusion/exclusion criteria, the logic in estimating standard errors and sample size determination, and the sources of sampling errors."

--Gerald Albaum, book review in Journal of Marketing Research

1	Target Populations and Samples

A sample is a portion or subset of a larger group called a population. The population is the universe to be sampled, such as all Americans, all residents of California during the 1994 earthquake, or all people over 85 years of age. Survey researchers often study samples rather than whole populations.

A good sample is a miniature version of the population of which it is a part—just like it, only smaller. The best sample is representative, or a model, of the population. A sample is representative if important characteristics (e.g., age, gender, health status) of those within the sample are distributed similarly to the way they are distributed in the larger population. Suppose the population of interest consists of 1,000 people, 50% of whom are male and 45% of whom are over 65 years of age. A representative sample will have fewer people (say, 500), but it also consists of 50% males and 45% over the age of 65.

Survey samples are not meaningful in themselves. The importance of a sample lies in the accuracy with which it represents or mirrors the target population, which consists of the institutions, persons, problems, and systems to which or to whom the survey’s findings are to be applied or generalized. Consider the two surveys described in Example 1.1. In Survey 1, 500 parents will be sampled, and their responses will be used to represent the views of the target population: all parents whose children are in the district’s elementary schools. In Survey 2, 200 persons will represent the target population of all library users who check out books.

EXAMPLE 1.1

Two Surveys: Target Populations and Samples

Survey 1

General Purpose: To examine the attitudes of parents regarding the introduction of new dietary and nutritional programs into elementary schools

Target Population: All parents of children in a school district’s elementary schools

Sample: 500 of the district’s 10,000 parents (100 chosen at random from each of the district’s five elementary schools)

Survey 2

General Purpose: To compare the reading habits of different users of the local library

Target Population: All persons who check out books from the library

Sample: Over the course of an allotted 6-month period, the first 200 of all persons who check out books and complete the survey

Why should you sample? Why not include all parents and all people who check out books in these two studies, for example? Sampling allows for efficiency and precision in a survey study. Samples can be studied more quickly than entire target populations, and they are also less expensive to assemble. Sampling is efficient in that resources that might go into collecting data on an unnecessarily large number of individuals or groups can be spent on other activities, such as monitoring the quality of data collection.

Sampling helps to focus a survey on precisely the characteristics of interest. For example, if you want to compare older and younger parents of differing ethnicities, sampling strategies are available (in this case, stratified sampling) to give you just what you need. In many cases, it is more suitable for survey researchers to examine a sample of the population with precisely defined characteristics than to try to look at the entire population.

When selecting a sample, you should make sure that it is a faithful representation of the target population. No sample is perfect, however; almost every sample has some degree of bias or error. Use the following checklist to help ensure that your sample has characteristics and a degree of representation that you can describe accurately.

---

Checklist for Obtaining a Sample That Represents the Target

✓ State your survey objectives precisely.

The objectives are the reasons for doing the survey. Surveys are done to describe, compare, and predict knowledge, attitudes, and behavior. For example, a company might conduct a survey of its employees to gather data that will describe employees’ educational backgrounds and preferences for work schedules and enable human resources personnel to compare employees on these characteristics. School administrators might have surveyors conduct interviews with students and use the data to help predict the courses that are likely to have the most influence on students’ future plans.

Researchers also use survey data to evaluate whether particular programs and policies have been effective. For example, a company’s management may be interested in investigating whether employee morale has improved 3 years after the firm’s reorganization, and school officials may want to know how students in a new ethics education program compare with those who have not been exposed to the program in terms of their goals and aspirations. If surveys involve self-administered questionnaires and/or interviews, they are being used for research purposes (research is used here in a very general way to include systematic inquiries or investigations).

Example 1.2 illustrates how the general purposes of the two surveys described in Example 1.1, of parents’ attitudes toward nutrition programs and of the reading habits of library users, might be further refined into specific objectives and research questions.

EXAMPLE 1.2

General Purposes, Specific Objectives, and Research Questions

Survey 1

General Purpose: To examine the attitudes of parents regarding the introduction of new dietary and nutritional programs into elementary schools

Specific Objective: To describe and compare the attitudes of parents of differing ages, ethnicities, and knowledge of nutrition toward the introduction of three new dietary and nutrition plans under consideration by the schools

Specific Research Questions:

1. What are the attitudes of parents of differing ages toward the introduction of three new dietary and nutrition plans?
2. What are the attitudes of parents of differing ethnicities toward the introduction of three new dietary and nutrition plans?
3. Do parents who know more about nutrition differ from other parents in their attitudes?

Survey 2

General Purpose: To compare the reading habits of different users of the local library

Specific Objective: To compare reading habits among local library users of differing ages, genders, and educational attainment

Specific Research Questions:

1. Do differences exist between older users and younger users in terms of reading habits?
2. Do differences exist between males and females in terms of reading habits?
3. Do differences exist among people of differing educational levels in terms of reading habits?

The specific research questions established are the guide to the specific questions or items that must be included in the survey. Given the research questions stated for Survey 1 in Example 1.2, the survey must include questions about the respondent’s age and ethnicity and questions that test the respondent’s knowledge of nutrition. Survey 2 must include questions about the respondent’s age, gender, education, and reading habits.

---

EXERCISE

Add at least one possible research question to Surveys 1 and 2 in Example 1.2.

POSSIBLE ANSWERS

For Survey 1: Do parents of older children and younger children differ in their attitudes?

For Survey 2: Which are the most important factors in predicting reading habits: age, gender, and/or educational attainment?

---

✓ Establish clear and definite eligibility criteria.

Eligibility or inclusion criteria are the characteristics a person must have in order to be eligible for participation in the survey; exclusion criteria are those characteristics that rule out certain people. You apply your inclusion and exclusion criteria to the target population, and once you have removed from that population all those who fail to meet the inclusion criteria and all those who succeed in meeting the exclusion criteria, you are left with a study population consisting of people who are eligible to participate. Consider the illustrations in Example 1.3.

EXAMPLE 1.3

Inclusion and Exclusion Criteria: Who Is Eligible?

Research Question: How effective is QUITNOW in helping smokers stop smoking?

Target Population: Smokers

Inclusion Criteria:

• Between the ages of 18 and 64 years
• Smoke one or more cigarettes daily
• Have an alveolar breath carbon monoxide determination of more than eight parts per million

Exclusion Criterion: Have any of the contraindications for the use of nicotine gum

Comment: The survey’s results will apply only to persons who are eligible to participate. The findings may not apply to any smokers under 18 years of age or 65 or over. Although the target population is smokers, the inclusion and exclusion criteria define their own world, or study population, as “people ages 18 through 64 who smoke one or more cigarettes a day, have an alveolar breath carbon monoxide determination of more than eight parts per million, and do not have any of the contraindications for the use of nicotine gum.”

Research Question: Are parents of elementary school children satisfied with the school district’s new reading curriculum?

Ta...

Table of contents

• Cover Page
• Title
• Copyright
• Contents
• How to Sample in Surveys: Learning Objectives
• 1. Target Populations and Samples
• 2. Statistics and Samples
• Exercises
• Suggested Readings
• Glossary
• About the Author