Skip to main content
main-content
Top

About this book

Quantitative research makes a very important contribution to both understanding and responding effectively to the problems that social work service users face. In this unique and authoritative text, a group of expert authors explore the key areas of data collection, analysis and evaluation and outline in detail how they can be applied to practice.

Table of Contents

1. Introduction

Abstract
Social work counts – every day social workers make a positive and lasting difference in the lives of many people. Whether it is supporting adolescents with life-limiting conditions, enabling an older person to remain living independently in their own home with a range of coordinated supports or providing therapy to a young child living with domestic violence, social workers practise at the very forefront of the challenges facing all of us in society. In spite of the many different forms that social work can take, whether working with an individual, a group or a community, social workers draw upon a common set of values to guide their practice. At the heart of these values is a commitment to promoting social justice. Social justice can be expressed in many ways, but a key feature is the imperative for social workers and the organisations for whom they work to bring about positive change. Yet, as Sheldon and Macdonald (2008, p. 66) note: It is perfectly possible for the good-hearted, well-meaning, reasonably clever, appropriately qualified, hard-working staff, employing the most promising contemporary approaches available to them, to make no difference at all to (or even on occasion to worsen) the condition of those whom they seek to assist.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

2. Why Numbers Matter in Social Work

Abstract
Numbers are a part of everyday life, and our ability to engage with them is second nature to us. This also extends to the world of social work, whereby numbers are a necessary part of practice and policy, and our ability to engage with and understand numerical concepts and data should enable us to work more effectively with service users and to address social issues. In this chapter, we will explore in more detail why being able to engage with numerical concepts and data is important, and introduce you to some key numerical and statistical concepts (see Box 2.1) that will be developed throughout the remainder of the book. Ofcom, the UK government’s independent regulator of telecommunications, reports that 93% of adults in the UK own a mobile phone, with 61% using a smartphone that has the capacity to connect to the internet (Ofcom, 2013). As such it is very likely that you, the reader, own a mobile phone. The mobile phone market is highly competitive, with a range of different companies making the mobile phone handset, and another range of companies offering the service for making calls, sending texts and mobile browsing.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

3. Who is Being Studied?

Abstract
Perhaps the most important question to ask of any study is – who is being studied? It is rarely possible to study everyone in the group you are interested in (e.g. all children in care; everyone with a particular disability), therefore researchers need to find a smaller group (known as a sample) who are representative of the larger group (the population). This gives rise to two very important questions Who are researchers trying to study? (sample type) Who did they actually study? (response rate) The key concern with sampling is whether it is possible to make a statistical inference from the sample to the population. In research terms we should differentiate between the definition of population in the general sense and the research sense. In a general sense the population is often understood as referring to everyone in society, whereas in research the population is more typically an object (such as individuals, households, communities) which shares the same characteristic as the issue we are interested in; for example, all young people in care.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

4. What is Being Studied?

Abstract
In the previous chapter we considered who is being studied – the ‘sample’. This chapter examines what is being studied. In quantitative research the ‘whats’ are known as variables. A variable is simply something that can be measured, and that varies. Quantitative research involves examining the relationship between variables. For instance, gender can be assigned a number (say 1 for male and 2 for female) and so can height (say 170 centimetres). So it is relatively straightforward to test in any given sample whether there is a difference in the height of men and women. It is also possible to say what the size of this difference is on average and to describe it in a variety of other ways (which are explored in Chapters 8, 9 and 10). Of course, it is not usually of interest to social workers to explore the differences in height between men and women. Yet variables allow us to explore things that are likely to be important to social workers. For instance, instead of the relationship between gender and height we might want to explore the relationship between gender and other variables. This would allow us to investigate questions such as Are women more likely to be depressed?
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

5. How to Describe Issues Using Numbers

Abstract
This chapter builds on the description of key concepts such as reliability and validity to consider how quantitative data can be used to describe things in practice. This might include research questions such as how common a particular issue is or how serious a particular problem tends to be. The chapter covers in particular questionnaires and routine data as sources for quantitative research. Existing data sets are covered in Chapter 12. The chapter begins by considering routine information gathered by social work and social care agencies and its strengths and limitations for research purposes. This discussion is illustrated with examples from government sources to illustrate the political importance of a critical understanding of routine data collection in social work. The chapter then turns to an exploration of the contribution that questionnaires can make, and an understanding of some basic issues in questionnaire design with examples from recent social work research studies. The information is intended to help students who may wish to use questionnaires in small-scale primary research as well as those wishing to critique them.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

6. How to know if a Service Makes a Difference

Abstract
Research suggests that social workers want research to answer questions around what works (Stevens et al., 2009). This chapter explores the complexity involved in answering whether services or ways of working make a difference. It does this through a particular focus on three common evaluative designs. These designs provide a way of introducing broader debates about the nature of evidence around what works. The chapter explores the key issue of how to establish whether services provided for people make a difference, including the debates around this topic within social work and the social sciences. The most popular evaluative designs are introduced, with a particular focus on before-after, quasiexperimental and experimental designs. The broader evaluative research tradition is considered, including logic models and realist approaches. Throughout, key critiques are considered in relation to each type of evaluative design. The overall argument is that each method has strengths and weaknesses, and is likely to be appropriate for particular purposes in specific contexts.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

7. How to Use Numbers to Describe a Sample

Abstract
In this chapter, we discuss how numbers can be used to describe a sample. In Chapter 3, we defined a sample as a subset of a population. For example, a population could be all 50 adults who live in a residential care home, and a sample could be 25 of those adults selected based on a specific sampling method (i.e. simple random, stratified, convenience). We would anticipate that an understanding of those 25 adults could provide an understanding and insight into the population as a whole; although, as Chapter 3 highlighted, the extent to which we can transfer knowledge of a sample to a population is contingent on the choice and rigour of our sampling methods. Through this chapter, we aim to demonstrate how a description of a sample, through the use of quantitative methods, can better assist in understanding the characteristics and needs of the sample and the population which the sample represents. This knowledge is relevant to your social work practice as such an understanding can enable you to better understand the characteristics of a particular group of service users, tailor social work services to service users’ needs and begin to determine whether such services are effective.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

8. How to make a Decision with Confidence

Abstract
Social work research often relies on gathering and analysing data from a sample of subjects that represent a larger population. The aim of such research is to then generalise the findings from the subjects to the population. Yet how can we be confident that the findings from the research on a sample accurately reflects the true situation or condition of the population and that the findings are not due to chance? The use of specific statistical methods, referred to as inferential statistics, enable us to determine how probable it is that our findings from the research on a sample are not due to chance and, thus, are likely to reflect the situation among the population. In order to do so, we must understand the concept of probability. In this chapter, we will explore probability and a probability distribution, otherwise referred to as the standard normal distribution or normal curve. We will then move to discuss how we can use hypothesis testing to determine the probability that our research findings are not due to chance (or sampling error) and, thus, whether any research findings of differences between variables are ‘statistically significant’.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

9. How to know if Two Variables are Related

Abstract
Quantitative social work research requires researchers to gather and analyse data in order to answer research questions or test hypotheses. The findings allow for researchers to tell a story about a sample, determine if social work interventions are effective or not and determine the extent to which the findings can be generalised to the larger population or across different situations. But, once we have collected our data, how do we know what statistical test to use to answer our research questions or test our hypotheses? How are we able to determine if variables in our data set are related, influence one another or cause one another? In Chapter 7, we discussed a ‘variable’, the different levels of measurement of variables (nominal; ordinal; interval; ratio) and how to calculate descriptive statistics (frequency; percentage) and the measure of central tendency of variables (mean; median; mode) based on their level of measurement. Then, in Chapter 8, we looked at the theory of probability, probability distribution and the normal curve and how to determine if there is a ‘statistically significant’ relationship or difference between two variables based on the p-value from a statistical test and the set α level.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

10. What is the Effect of One or More Variables on Another Variable?

Abstract
Quantitative research and statistics enable us to explore the relationship between several variables and determine the extent to which they are related or how one or more variables might influence another variable. Given this information, we are then able to see if a change in one or more variables is associated with a change in another variable and, thus, use this knowledge of those variables and their relationship to predict future situations. For example, through quantitative research and statistics we could determine the variables (or factors) that lead to receiving higher scores on a research methods test. This information will enable us to predict a student’s research methods test results when we have information on the variables found to predict test results. Likewise, we may want to predict the number of days a service user will remain sober after completion of a substance misuse treatment programme when we have information about her/his demographic characteristics, substance misuse history and characteristics of the treatment programme. The statistical test we would use in both situations is called linear regression analysis.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

11. What are the Key Elements of Ethical Quantitative Research?

Abstract
Ethical principles are not limited to particular research designs or methods. Responsible research practice should be rooted in core principles which apply to all types of studies. The core concepts behind this chapter are therefore not only relevant to quantitative research. The chapter introduces the four principles of respect for autonomy, non-maleficence, beneficence and justice (Beauchamp and Childress, 2013) and outlines their application in social work research. In keeping with the rest of the book, the practical examples used will be taken from quantitative studies and there will be a focus on the particular ethical issues that arise for quantitative research. So, among other things, there will be discussion of the ethics of secondary analysis of survey data, using social services databases for research, following up participants in longitudinal studies and the ethics and politics of effectiveness research. Ethical conduct should of course be kept in mind from the very start of a research study and not tagged on at the end.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

12. How to do Quantitative Research without Collecting New Data

Abstract
Generating original quantitative data can sometimes be very challenging for many different reasons. For many students, time is limited. For example, they may need to do some of their dissertation research while also working full time on placement. It may be difficult or impossible in practice to identify a robust sample. You may have a brilliant idea but a gatekeeper refuses you access to approach potential respondents. Or you may get as far as carrying out a quantitative study – for example a survey – but you get a terrible response rate which risks invalidating your study. You may find it impossible to contact service users for follow-up interviews because they have moved or changed their phone numbers. Life can get in the way of good quantitative research. However, you should not be pessimistic after this downbeat chapter opening, because the good news is that students and other researchers do not necessarily need to collect their own original data. There are some excellent existing data sets out there and you can take these data sets and carry out quantitative secondary analysis. Some of these data sets are in fact much more detailed and have been conducted with far better standards of validity than you as a lone researcher could possibly achieve on your own.
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter

13. Using Mixed Quantitative and Qualitative Methods in Social Work Research

Abstract
In a book which advocates the value and importance of quantitative research for social work, it may seem strange to begin a chapter with some comments about the limitations of these methods. Certainly, quantitative social work research can tell us a great deal about such matters as the extent of social problems, the outcomes of services and the effectiveness of interventions. But sometimes the numbers do not tell us enough. For example A survey of a ‘new’ population (e.g. young people with a life-limiting condition living unexpectedly into adulthood because of advances in medicine and surgery) may not ask the most pertinent questions; existing instruments may miss the mark. Statistical analysis based on average scores may fail to attend to, or actually mask, the ‘outliers’, those cases whose results are markedly different from the rest of the study population (e.g. children who are exceptionally resilient in the face of extraordinary life circumstances such as migration alone across continents).
Barbra Teater, John Devaney, Donald Forrester, Jonathan Scourfield, John Carpenter
Additional information