Survey Research

The Basics

Keith F Punch

• Editorial: SAGE Publications Ltd
• Año: 2003
• ISBN:
  9781446234921
• Edición: 1
• Categoría: Scienze sociali
• Categoría: Metodologia e ricerca nelle scienze sociali

Survey Research can be used as an independent guide or as a workbook to accompany Keith F Punch?s bestselling Introduction to Social Research (SAGE, 1998). It represents a short, practical `how-to? book on a central methodology technique aimed at the beginning researcher.

The focus of this book is on small-scale quantitative surveys studying the relationships between variables. After showing the central place of the quantitative survey in social science research methodology, it then takes a simple model of the survey, describes its elements and gives a set of steps and guidelines for implementing each element. The book then shows how the simple model of the quantitative survey generalizes easily to more complex models. It includes a detailed example of both simple and complex models, which readers should find very helpful.

It is directed primarily at beginning researchers - upper-level undergraduate and graduate students in any area of social science, who often have to do small scale surveys in projects and dissertations. Beyond this, it will be of interest to anybody interested in learning about survey research. It is written in non-technical language, aiming to be as accessible as possible to a wide audience.

1

Introduction and Purpose

---

CONTENTS

1.1 Focus and purpose of this book

1.2 Plan of the book and chapter outline

1.3 Important conceptual tools – independent and dependent variables

1.4 Model of research

1.5 Review concepts

Note

---

1.1 FOCUS AND PURPOSE OF THIS BOOK

The survey has long been a central strategy in social research. However, the term ‘survey’ itself is very broad, covering many different types, and it is used in many different ways and in many different contexts. Because of this, there already exists a great deal of literature on survey methods, some of it general and some of it focused on surveys in particular types of situation. A new book on surveys therefore needs to have a distinct focus based on a clear rationale.

This book focuses on small-scale quantitative surveys which study the relationships between variables. The reason for the focus on quantitative-surveys-relating-variables is that this method is central to a very wide range of social science research, but is not often dealt with directly, and as a whole, in the literature. The reason for the focus on small-scale surveys is that the book is aimed at graduate students in a research training environment where very often only small-scale surveys are possible.

Of course, not all surveys are quantitative. Qualitative surveys, usually asking open-ended questions, do not normally produce quantitative or numerical data. People respond with answers to these open-ended questions in words, and researchers often proceed to analyse such responses without somehow transforming the words into numbers. Such surveys are not included in this book, which deals only with surveys designed around quantitative data.

The essence of quantitative research is the study of relationships between variables. For the quantitative researcher, reality is conceptualised as variables which are measured, and the primary objectives are to find how the variables are distributed, and especially how they are related to each other, and why. As is described in Chapter 2, a feature of the methodological history of quantitative social science has been the move away from the experimental method, where the researcher manipulates one or more variables in order to study their effects on other variables, to a broader non-experimental method and design, where the researcher uses naturally occurring variation in variables to study relationships between them. In other words, the strategy for studying the relationship between variables has broadened from a narrow concentration on experimental method to a more widely applicable approach using non-experimental methods.

The basic reason for this move has been the limited scope of the research questions which can be answered using the true experimental method, especially as social science areas expanded beyond psychology and education. For the great majority of research areas and topics, in a large number of social science areas, the researcher cannot manipulate variables for research purposes in order to study their effects. Research methodologists therefore developed non-experimental methods, by applying the logic of the experiment to the non-experimental research situation. This means that non-experimental quantitative methods are now an essential part of the methodological expertise of the social science researcher. It becomes correspondingly important for student researchers to learn these methods. The phrase ‘non-experimental quantitative methods’ really means quantitative survey methods which focus on the relationships between variables.

The purpose of this book is to teach these methods. To do that, it takes a simple model of the quantitative-survey-relating-variables, dismantles it to show how it works, and illustrates it with a detailed example. It then shows the flexibility of these methods, and how this model extends and generalises to more complex situations. I see mastery of this model, and of these methods, as an important objective in the training of social science researchers. In this, I agree with Babbie (1990: 40): ‘If you fully understand the logic and skills of survey research, you will be excellently equipped to learn and to use other social research methods.’

Thus this book focuses on surveys which are:

• quantitative;
• concerned with relationships between variables.

Quantitative means that the survey is designed to produce numerical data, and proceeds by measuring variables. As noted, not all surveys are quantitative. But the focus here is on quantitative surveys which use numerical data produced by the measurement of variables.

Concerned with the relationship between variables means that the point of the survey is not simply to describe variables and how they are distributed. Its main aim is to study how variables are related to each other. This distinction is important, and the point is elaborated in Chapter 2.

In addition, the focus here is on surveys which are:

• small-scale;
• cross-sectional;
• based on the individual person as the unit of analysis;
• built around a self-administered questionnaire.

Small-scale means that I want to focus on situations where researchers have limited resources, and are therefore restricted in the size and scope of the survey they can conduct. A typical example of this is the dissertation or project of a student working in an area of social science.

Cross-sectional means that the survey collects data from people at one point in time, rather than at two or more points in time. This latter type of survey is longitudinal. While much of what is said here applies also to a longitudinal survey, the focus is on the cross-sectional survey since it is a more common method, especially for the graduate student.

The individual person as the unit of analysis means that the logic of the research is to investigate how individual people vary on the different variables, and how that individual-person-variance is related across the different variables. As noted later in this chapter (p. 5), surveys can be designed with other possible units of analysis (in educational research, for example, the school class, the school itself, or the school system could be the unit of analysis), but surveys with the individual as the unit of analysis are generally more common.

Built around the self-administered questionnaire is a simplification concerning the way the survey data are collected. Again, there are other possibilities, but the self-administered questionnaire (or paper-and-pencil self-report questionnaire, as it is sometimes called) is the most common method of data collection in the quantitative survey. This is elaborated in Chapters 3 and 4.¹

The purpose of this book is therefore to describe the logic and strategy behind this type of survey, to analyse the different elements-which make up this type of survey, to outline the steps by which such surveys are done, to demonstrate these steps through an example, and to show how the logic and strategy extend from simple to more complex investigations.

1.2 PLAN OF THE BOOK AND CHAPTER OUTLINE

The strategy for this book has two parts. The first part shows the central place of the quantitative survey in social research, identifies and analyses its elements, and shows how each element is operationalised (see Chapters 2-5). The second part of the strategy is to show how the simple model of the survey generalises to more complex versions (see Chapters 6 and 7).

The chapter plan follows this strategy. Thus, after this introductory chapter, Chapter 2 gives some methodological background to show the central place of the quantitative survey, Chapter 3 describes each of the main elements of the survey, and Chapter 4 shows how each element is operationalised through a set of steps. Chapter 5 then discusses the survey report and Chapter 6 contains the example, which is presented first in simple form and then extended. To finish the book, Chapter 7 shows how the logic of the simple model of the survey generalises to more complex examples.

The book concentrates initially on a simple model of the quantitative survey. It does this for ease of communication and to show both the logic of the quantitative survey and the steps involved in implementing that logic as clearly as possible. But later in the book, I show how the simplified version extends easily to more complex versions, and how the same logic supports the more complex versions. This means that the early chapters are written in simple independent-dependent variable terms, and the simple form of the example in Chapter 6 has two independent variables and one dependent variable. The more complex form of the example extends this, and the theme of extensions is taken further in Chapter 7.

A great strength of the quantitative survey, as a research strategy, is this breadth and flexibility. The type of quantitative survey described here is suitable for many different contexts, making it of great value in many different areas of social research. It is also easily adapted to different configurations of variables and different numbers of variables. This means that a wide variety of research problems can be investigated using these methods. Chapter 7 illustrates some of that variety.

1.3 IMPORTANT CONCEPTUAL TOOLS – INDEPENDENT AND DEPENDENT VARIABLES

As noted, the quantitative researcher sees the world as made up of variables, modelling what we do constantly in everyday life. Thus, in the physical world, we think in terms of such variables as length, mass or weight, temperature, and so on. In the economic world, we have such variables as inflation, unemployment and productivity, among many others. We can conceptualise aspects of the social world similarly. For example, people may differ in the variables age, gender, self-esteem, level of job satisfaction or motivation. Groups or organisations may differ in the variables size, cohesiveness, clarity of purpose or level of goal achievement.

We should notice two things about these commonplace examples of thinking in terms of variables. Both have important technical consequences. First, a variable is seen as the property (or characteristic) of some entity. Most often, in social research, the entity is the individual person – it is the individual’s gender or level of self-esteem or motivation which is of interest. But the entity can also be the family, the group, the organisation, the city, the industry or the economy. This is the technical issue of the ‘unit of analysis’, something we need to get clear when planning a survey. For simplicity this book describes surveys where the individual person is the unit of analysis. But the logic and methods described are the same for other possible units of analysis. Secondly, the property – the variable – may vary either in categories or along a continuum. Thus gender varies in categories (male or female), whereas job satisfaction ranges along a continuum from low to high. This is the important technical issue of categorical versus ...