Social Research

An Introduction

Matthew David, Carole D Sutton

• Publisher: SAGE Publications Ltd
• Year: 2010
• Print ISBN: 9781847870131, 9781847870124
• eBook ISBN: 9781446245040
• Edition: 2
• Topic: Social Sciences
• Subtopic: Social Science Research & Methodology
• Index: Social Sciences

The perfect book for any student taking a research methods course for the first time!

The new edition of David and Sutton?s text provides those new to social research with a comprehensive introduction to the theory, logic and practical methods of qualitative, quantitative and mixed methods research. Covering all aspects of research design, data collection, data analysis and writing up, Social Research: An Introduction is the essential companion for all undergraduate and postgraduate students embarking on a methods course or social research project.

The second edition features:

- Brand new chapters on visual methods, case study methods, internet research, mixed methods and grounded theory

- Updated chapters on interviews, questionnaire design, surveys, and focus groups

- Improved coverage of qualitative and quantitative methods of data analysis, including practical instruction on the latest versions of software packages NiVivo 8 and SPSS 18

- An attractive new layout which aids navigability and enhances the book?s student learning features

- A companion website (www.uk.sagepub.com/david) with PowerPoint slides and links to useful websites

- Many more practical examples helping bring theory to life!

Designed for social science students with no previous experience, this book provides a balanced foundation in the principles and practices of social research.

PART ONE

STARTING YOUR RESEARCH

ONE

GETTING STARTED: THEORY, RESEARCH QUESTION AND RESEARCH DESIGN

Chapter Contents

Getting into the water

What is social research? Science, theory and data collection

Identifying a research question

Social problems, political issues, personal motives?

Sources of a research question

What makes a good research question?

The value of a good research question

Refining your research question: from research question to research design

Testing or exploring?

Primary and secondary sources

Causes, meanings and probabilities? Logic, relationships and people

Data: asking, looking, reading and recording

What, and how much, is good enough? Validity, reliability and generalizability

Evaluation, participation and action research

Writing a research proposal

What do they want?

Presenting yourself

How long have you got?

Sequence

Summary

Questions

Further reading

---

Aims

By the end of this chapter you will be able to:

• Understand key elements in formulating a research question.
• Distinguish between deductive and inductive research designs, and be able to address the questions that social researchers need to ask when choosing which approach to adopt.
• Distinguish primary and secondary data, and be able to critically evaluate their relative advantages and disadvantages.
• Identify logical fallacies that recur in research and in everyday life, and which distort understanding.
• Comprehend the basic elements of sampling.
• Understand key elements of data validity.

---

Getting into the water

This chapter, along with Chapters 2, 3 and 4, sets out issues that all social researchers need to think about, decide upon and carry out in order to begin an empirical research project. In some respects this is like getting into the water when going swimming. There is an advantage to just jumping in: the best way to learn is to have a go. But just jumping in without having learnt some basic things might lead to drowning, so here we offer some basic guidelines on what to do. These chapters are about ‘getting into the water’ safely and with confidence. Doing social research requires that you do a large number of things, seemingly at the same time. This, at first, seems confusing. Those who have been doing research for a while tend to take it for granted, and so they are not always aware of every aspect of what they are doing when they do it. This can confuse the beginner even more. Getting into the loop is about picking up the taken-for-granted routines of the more experienced researcher, and practising them. Once you are familiar with the steps, processes and short-cuts you will no doubt develop your own style, your own routines and your own agenda. The way things are set out here is to help you get started. You will always have to bear them in mind, but after a while you may do things in your own way. This chapter will start with a discussion of how you can generate a research question. It will then discuss how to begin converting this research question into a research design, which involves decisions about testing or exploring; using primary or secondary data; causal or descriptive approaches; interviews, surveys, archival data or observations; issues of validity, reliability and generalization; and evaluation, participatory or action designs. It will finally address the question of writing a research proposal that will allow you to actually do your research. Box 1.1 outlines a few key distinctions that you will need to familiarize yourself with before you read the rest of this book.

---

BOX 1.1 HINTS AND TIPS

Setting out in the minefield of terminology

You will need to check the glossary of terms in the back of this book to start learning the meaning of the words that commonly litter this and any other book about social research methods. However, just to start you off, it is useful to remember that so much of social research hinges on distinctions between research that seeks numerical values and research that collects words. The former is called quantitative and the latter is called qualitative. Much hinges on this division as those that collect numbers need to carry out different kinds of data collection, such as a questionnaire, where a qualitative researcher would be more likely to use an interview to collect more open-ended verbal responses. Issues then arise over whether you want a large sample with short answers (such as can be gained from quick closed numerical questions) or a smaller sample of longer answers (such as can be gained from open-ended interviews). Do you want highly prestructured data collection or space to explore a topic? Will you try to measure how one thing influences another or only whether some things go together with others? These questions all go to the heart of how social research should be carried out, and choices should be informed by what you want to find out. Words like inductive, deductive, causal, exploratory, qualitative and quantitative, naturalistic, controlled and structured should all be looked up in the glossary if you are unfamiliar. Do this now or when the words come up in the text.

---

What is social research? Science, theory and data collection

Social research has many forms, and this book seeks to introduce the reader to their basic forms and logics. Part I sets out the key questions that the social researcher must address when starting a research project. In so doing, Part I will outline the basic properties of social research. While practical in nature, the first four chapters will highlight more abstract issues and debates, in particular the relationship between theory and the research process, both in conduct and in choice of method. These debates hinge around two questions:

• Is social research a science?
• Can humans be studied usefully in a scientific manner?

These two questions will be briefly addressed now, before moving on to the practical questions of how to (1) formulate a research question; (2) select a research design; and (3) present a research proposal.

Science is, in the popular imagination, the experimental method. The experiment is the stereotypical image of scientific method. Some social researchers use the experimental method. However, most social researchers do not! Experimental method is the establishment of controlled conditions in which the effect of variables on other variables can be measured. Regulation of inputs allows accurate estimation of causes in the variation of outputs. Experimental method requires an initial prediction about how variation of inputs will affect outputs such that this prediction can then be tested. This prediction is a provisional theory (or thesis). This is called a hypothesis. A variable is anything whose amount can vary, and which is defined in such a way that its variation can be measured (and in an experiment also controlled in this variation). A number of variables may be specified. In the classic experiment (Box 1.2) all identifiable variables are held constant (controlled conditions), bar two variables. These are the independent and dependent variables set out in the prediction/hypothesis. The hypothesis predicts that variation in the independent variable causes variation in the dependent variable. With all other things held constant, the experiment is designed to allow this hypothesis to be tested. The hypothesis is drawn from prior examination of research on the subject. As such the experiment is theory driven (in other words, the data collection is designed to fulfil a need for the information required to answer a theoretical question). This approach to the relationship between theory and research is called deduction. Hence the experimental method is called hypothetico-deductive.

---

BOX 1.2 CONSIDER THIS

A classic experimental design

It is commonly suggested that students underperform in assessments due to lack of sleep before exams. They stay up all night revising. It would be possible to select a group of students with otherwise similar characteristics (age, gender and previous exam performance, for example). These students could all be given the same amount of preparation in the week before an exam. Then, on the night before the exam, all the students would be kept together in a controlled hall of residence. The group could be divided into subgroups. One group would be required to go to bed 10 hours before the exam, the next 9 hours, the next 8 hours etc., with the time not spent sleeping being given over to exam revision. Who would do best? Those who had 10 hours sleep and did no additional revision, those who had 8 hours sleep and 2 hours revision, those who had no sleep and 10 hours revision? Or would the best performance come from somewhere in between?

---

It is important to note that much research in the physical sciences is not strictly speaking experimental. Much of geology, astronomy and biology deploy methods of data collection beyond the laboratory. Geological and evolutionary time, galaxies and ecosystems cannot be replicated in controlled conditions. This is also the case for many aspects of social life (Box 1.3). Science cannot be defined exclusively in terms of the classical experiment.

---

BOX 1.3 CONSIDER THIS

The arrival of television: naturalistic experiments

Does television damage children, cause violent behaviour, encourage sexism, consumerism, political apathy, anorexia, body dissatisfaction and/or laziness? Even if those who watch violent TV are more likely to commit violent acts, it could just be that those with a liking for violence choose to watch such images, rather than being caused to be violent by such images. How can we tease out what is causing what? One way would be to compare a location where television is not watched with a location where it is. This may be hard to do now, as television is almost everywhere, but historically and geographically this has been done often. It may be possible to compare, for example, one prison where television is allowed with one where it is not. The ideal naturalistic experiment is where the same place can be studied before and after the introduction of the variable you think is going to make a difference (such as TV). Of course, it has to be assumed that the differences between places or in a place before and after the introduction of a particular variable are only differences in the variable itself, and not in other significant factors. The fact that a place gains television may be linked to increased income, which may impact on diet, and it might be the increased food intake that caused changes in body image, not television. Those seeking to carry out naturalistic experiments are always having to reflect on such possibilities.

However, despite not always using the experimental method, much of the remaining physical science research does deploy another form of hypothetico-deductive research (that is, research where a prediction is tested through the variation of observed conditions). This may be through comparison over time or between locations. If different levels of a particular variable exist in different locations or at different times, it may be possible to measure the levels of other variables in those times and places to see if patterns (or correlations) exist. Whilst an element of control is lost, it may still be possible to show that variations in one factor go along with variation in other factors (even if what is causing what is harder to pin down). A hypothesis can be stated. The researcher can then go looking for the conditions necessary to test that hypothesis. Data can be collected and results analysed which will then support or challenge the hypothesis. This is still therefore hypothetico-deductive research.

A far greater amount of social research adopts this approach. Proponents of this type of social research tend to see themselves as scientists. However, some forms of social research are not hypothetico-deductive and pursue an exploration-based approach. Sometimes this is to identify what is going on when existing knowledge is insufficient to generate hypotheses. There are plenty of such examples in the physical sciences. Sometimes, this exploratory approach is adopted as a rejection of the hypothetico-deductive method, its predictive process and causal assumptions. This raises the question of whether scientific methods are appropriate to study humans, or whether humans possess qualitatively different characteristics to physical objects (most particularly consciousness and choice) that invalidate predictive forms of research and the predictive model of explanation which hypothetico-deductive research is based upon. These issues will be discussed later (see section ‘Testing or exploring?’ later in this chapter, and see Chapter 4). These are questions that will recur throughout the research process, but which cannot be resolved at this stage.

Identifying a research question

While a hypothesis is a proposition to be tested, rather than a question to be answered, hypotheses are designed to focus attention within broader research areas or questions. While some research questions are very specific and others far looser in definition, the process of identifying a research question is always an essential first step in any project.

Social problems, political issues, personal motives?

A researcher may enter the process of identifying the research question at a number of different stages. Ironically, the student conducting a research project for their studies and the well-established research professor may have more in common here with each other than either may have with the majority of researchers in the middle. The privilege of starting from first principles, rather than being brought in at the middle or towards the end of the research problem identification process, is most often denied to those neither well established nor researching for study.

Identification of the research question may have many levels, only some of which will be within the researcher’s power to alter, at least in the first instance. Issues may become ‘ripe’ for research in the minds of those able to fund such activities for a number of reasons. Bodies engaged in education, health care, law and order, social work, economics, urban planning, commercial and governmental administration and so on will, for various reasons, come to the view that research may help them address or more clearly identify problems. Research may be funded by charities on issues of concern to them so that findings stimulate awareness and debate about those issues. These bodies will form an opinion about what needs to be researched and such motivations play a crucial role in the identification of research questions.

However, even while such factors play a crucial role in directing research, the question of how such research is to be carried out requires the researcher to develop the identification process from an idea to a practical activity. It is here that the researcher’s own interpretations of the ‘problem’, and the best way to research it, come into play. When the researcher can claim a degree of expertise both in the subject to be studied and in the methods by which such a subject can best be studied, they are in a position to introduce their own definitions of themes and interpretations of ‘problems’. To this extent the more de...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• CONTENTS
• Introduction and Structure of the Book
• Acknowledgements
• Guided Tour
• Part One Starting Your Research
• Part Two Data Collection Strategies
• Part Three Data Analysis
• Part Four Presenting Research
• Glossary
• References
• Index