Why sample?

Sampling is used when it is not possible or practical to include the entire research population in your study, which is usually the case. Sampling is the process of selecting a few from the many in order to carry out empirical research. It needs to be accepted from the outset that a sample represents a form of trade-off between the desirable and the attainable, but this is more often the case in statistical sampling than it is in descriptive sampling. In qualitative descriptive sampling the case is selected based on what we can learn from the case and the goal is rarely to make inferences about the wider population based on this discovery. There is discussion concerning the use of generalization in qualitative research but I would always tend to discourage this; transfer of findings is very different from generalization and if generalization is the purpose of the research then the applicability of qualitative research should be questioned. In most quantitative research the point is to take a sample and make inferences about the rest of the population based on that sample. With both approaches it may well be much more informative to study the entire population but this would almost always be impossible based on cost and time. For this reason we sample.

The method of sampling used plays a major role in any research investigation. Very often it is the characteristics, composition and scale of the sample that give weight to any findings that emerge from the investigation. You must take care when selecting a sampling technique; there are a number of different approaches to sampling and choice of approach should be influenced largely by the purpose of the investigation. You will need to demonstrate the appropriateness of the chosen sample to the nature and output of the research. It is totally inappropriate to engage in a small-scale, localized qualitative study then attempt to generalize from the findings. Likewise, it is inappropriate to engage in a large-scale, broad study and attempt to provide any real detail concerning individuals (unless it is a large-scale investigation involving many researchers investigating many people in a great deal of depth!).

Introduction

Historical research plays a vital role in the development of theory and practice; it has relevance as a research method in all subject disciplines and although very different from the other research methods discussed in this section, it is still a ‘scientific’ method, which must conform to standards of practice. The most notable difference between historical research and all other methods available to the researcher is the nature of the ‘evidence’ used to generate theory or test a hypothesis. Historical research relies on data that already exists in one form or another, unlike other methods, which are designed to create or generate data as a part of the research process. This is not to say that historical research cannot create or generate some data but this is likely to constitute a relatively small part of the dataset to be analysed and will depend on the research topic. History, after all, starts with the minute that has just passed; the sentence you have just read has now become a part of your history, however recent. Historical research is essentially qualitative because of the interpretation that is inevitably involved; there are uses for quantifiable data in some investigations but because this approach depends so much on interpretation, by definition it becomes qualitative. Most historical studies deal with natural behaviour in real-life situations and the interpretations the researcher brings to the evidence. Quantifying evidence can be provided where suitable data exists but it can be difficult to justify quantification of anything other than tangible evidence, such as reported statistics or records.

This research method is concerned with reconstructing the past, identifying pieces of a puzzle and putting them together to provide insight and understanding of a situation, event or process. In the field of information and communication studies historical research can be used to investigate processes, behaviour, individual events or patterns of use.

I'd like to start this preface with a huge thank you to the generous and kind individuals who reviewed the first edition. Stupidly, it hadn't occurred to me when writing it that it would be ‘out there’ and subject to the usual scrutiny all texts are exposed to; the realization dawned when a close friend and colleague told me to read the review in the online journal Information Research. It took me a while to click the link he had sent in the e-mail, I was genuinely terrified! To see the words ‘I have found it! Finally, I have found what I was looking for, since I started working with research students! A book that makes it clear’ (Maceviciute, 2007) made my heart jump and I decided that would be the only review I read. But of course it wasn't. I hope this second edition goes some way to address the things that people thought were missing from the first – thankfully there weren't many. In fact, I'm not entirely sure it was time for this but my publishers thought differently and I trust their knowledge and understanding. The most significant additions are the two guest chapters included here by close colleagues of mine: Professor Julie McLeod, Sue Childs and Elizabeth Lomas have provided an excellent chapter on Data Research Management, which is part of the output of their JISC-funded DATUM project. Dr Andrew Shenton has also contributed a chapter on the analysis of existing documents which adds another dimension to the text. My work has primarily been one of updating and restructuring in response to comments from colleagues and students, as well as my own desire to add or remove certain things. I hope I have addressed most if not all of the comments and I hope this second edition is useful to the current and next generation of LIS researchers.

Introduction

Case studies are not as simple to define as many other research methods, because of the nature and form of the method. A case study can be both the process engaged in to investigate a phenomenon and the written output of that investigation. The study is both the fieldwork and the report of that fieldwork. Leaving aside the written output for now, we shall focus on the method of investigation. The most commonly applied definition of case study research is provided by Yin, who defines it as: ‘an empirical inquiry that investigates a contemporary phenomenon within its real-life context; when the boundaries between phenomenon and context are not clearly evident; and in which multiple sources of evidence are used’ (Yin, 2008, 23).

Although you will find this definition is the one most often used by authors discussing case studies, you must remember that many case study researchers claim Yin takes a rather positivist approach to the case study method, and that a case study can be either qualitative or quantitative depending on what it is you are investigating and how you can acquire knowledge of the case.

Case study research is a method designed to study the particular within context and has a very specific purpose. The term has been used very broadly to cover a far wider remit than is appropriate: ‘The case study has unfortunately been used as a “catch-all” category for anything that does not fit into experimental, survey or historical methods. The term has also been used loosely as a synonym for ethnography, participant observation, naturalist inquiry and fieldwork’ (Burns, 2000, 458).

It is not surprising then that definitions of ‘case study’ have become rather blurred. A case study should be a study of ‘a functioning specific’ (Stake, 1994, 236). That is to say, a system that operates within well defined boundaries; the size and nature of that system are not the issue – what dictates the case is the purpose of the investigation.

Introduction

The purpose of survey research is to gather and analyse information by questioning individuals who are either representative of the research population or are the entire research population. The term ‘survey’ usually refers to a study that has used a representative sample; if the entire population is involved in the study it is a ‘census’. Questions must be asked using a standardized questioning procedure applied equally and consistently to all research participants.

The aim of survey research is to study relationships between specific variables, which are identified at the outset of the research and stated as either a hypothesis or a research question, or to describe certain characteristics of the population. The findings from the survey can then be generalized to the wider population. Survey research can include qualitative and quantitative research, but is usually quantitative with a limited qualitative element, which is more likely to be anecdotal than truly qualitative.

The term ‘survey’ is often used interchangeably with ‘questionnaire’; the two are not the same thing and it can lead to confusion if the distinction between the two is not made very obvious. A survey is a research method, the purpose and aims of which have already been stated; although data collection must be standardized, there are options for data collection within a survey. A questionnaire is a very specific data collection technique, which can be used within a variety of research methods. A survey, then, is the research method used to structure the collection and analysis of standardized information from a defined population using a representative sample of that population. Probability sampling is vital in order to make valid general izations about the wider population. When non-probability sampling is used you must take care with any statements you make that attempt to generalize to the wider population.

There are two types of survey: descriptive surveys and explanatory surveys. It is possible to apply both methods in the same study, as will become apparent when examining the nature of the two approaches.

In Part 1 I want to focus on the initial stages of any research investigation, the areas common to all research regardless of the character and purpose of that research. It is mostly very practical in nature, with the exception of the first chapter, which exam - ines the philosophical underpinnings of research. Although this philosophical debate is not for the faint-hearted, it is fundamental to understanding the entire activity we refer to as ‘research’. I suggest you take the time to engage in this debate if for no other reason than to decide which side of the paradigmatic fence you sit on, to determine your view of the world and the manner in which you feel we can explore that world. I secretly hope that reading this first chapter will capture your imagination and encour - age you to follow up the leads I offer; at the very least it should provide you with a broad overview of the fundamental differences between research paradigms and a framework for thinking about the research process. Following the completely new chapter contributed by Professor Julie McLeod, Sue Childs and Elizabeth Lomas on research data management, this section ends with a discussion of ethics in research.

Introduction

Action research is rapidly becoming one of the most popular research methods in information and communication research among practitioners. Increasingly they are recognizing the value of this approach in improving service provision, encouraging reflective practice and structuring and disseminating experience to the wider community. Action research has been used to investigate organizational functions such as the role of the library and library staff in the provision of distance learning programmes (Bailey et al., 2004), for educational research investigating individual solutions to differing learning needs, such as tackling student referencing errors (Kendall, 2005), and to improve students’ information literacy skills over a longitudinal series of interventions (Pickard, 2005). As Rowley points out:

Kurt Lewin is the sociologist who first used the term ‘action research’, to describe his work in the field of human dynamics with individuals affected by post-war social problems. He believed that in order to truly make a difference some form of change or action had to be embedded in the research design from the outset. An investigation leads to some form of action, which is followed by an evaluation of that action: ‘The research needed for social practice can best be characterized as research for social management or social engineering. It is a type of action-research, a comparative research on the conditions and effects of various forms of social action, and research leading to social action. Research that produces nothing but books will not suffice’ (Lewin, 1948, 202).

Introduction

As with ‘case study’, the term ethnography refers to the research method and the research output; ethnography is the process of engagement and the written account of that engagement (Agar, 1996). The most common use of ethnographic research is in anthropology, although social science research has increasingly adopted this method, which focuses on interpretations of behaviour or specific events in the everyday lives of individuals.

The primary data collection technique is participant observation, although other methods of collecting information concerning the context are often used. An ethnographic study begins with no a priori assumptions; the researcher seeks to explore and although the research will have a focus, the theoretical framework acts as cognitive signposts rather than controlling the direction of the investigation. Ethnography uses descriptive narrative to analyse and present the findings from the investigation; the purpose is to build theory by describing situations as they are and presenting patterns within a context.

Many of my students have asked me what the difference is between a case study and an ethnography, an understandable question, as these methods are similar. Conducting an ethnographic study may, on the surface, appear to be very similar to case study research but it is the extent to which the researcher is immersed in the context that is the real and most obvious difference. The focus of ethnography is to describe and interpret a cultural and social group, whereas the focus of case study is to develop an in-depth analysis of a single case. A case study site is usually visited at regular intervals to engage in data collection that can be largely predefined, whereas ethnography demands prolonged engagement within the context (Creswell, 1998).

Case studies are essentially ‘etic’ – outsider research – although the researcher attempts to gain as much ‘local’ knowledge as possible.

Introduction

Experimental research epitomizes the positivist paradigm discussed in Chapter 1. The opening quote in this chapter may appear to be very negative; the reason it is here is to highlight that experimental research has its place in library and information science but I would rarely advocate the application of this method in a study that involved human subjects. This is not to say it has not been done – it has – but I would argue that there are too many unstable elements involved ever to be in a position to construct a true experiment involving human beings and all the complex internal and external variables they are subject to. Experiments are possible on systems with stable functions, variables that can be systematically controlled; the human condition does not fall into this category. ‘Effects studies’ have been carried out in many branches of social science, but inference made from these studies is usually tenuous at best.

Before you decide to embark on experimental research you need to be very sure about what you are doing and the level of control you have over the situation. An experiment is a controlled research situation, which means ‘(a) that unwanted variables and external influences can be kept out of the experimental environment, and (b) that the researcher can establish the experimental conditions down to the smallest details’ (Corbetta, 2003, 94). When these conditions are not entirely met then the researcher has a responsibility to identify the extent of the error. ‘Experimental error occurs when a change in the DV (dependent variable) is produced by any variable other than the IV (independent variable)’ (Burns, 2000, 134).

Introduction

Research ethics is discussed in all research methods texts – or if it is not, then it should be. Although many of these texts will tell you the same, or similar, things, there is some conflict of opinion. A very public and often cited example of this conflict can be seen in the reaction to the research of Laud Humphreys (cited in Punch, 1994). Humphreys investigated the behaviour of homosexuals in semi-public places. His research was primarily concerned with the social background of the men who met for casual encounters in public toilets. From a vantage point in a city park he observed the behaviour of the men and secretly recorded their car number plates. Using this information he located the individuals, then visited them at their homes; this time he was in the guise of a medical survey researcher. In this way Humphreys acquired a considerable quantity of personal information about these men, which he later published. The reaction to his work demonstrates the conflict of opinion regarding research ethics. Humphreys was awarded one of the highest accolades available by an American research association for his work, but at the same time there was a campaign going on to have his PhD revoked for misconduct; he even received a thump in the jaw from another researcher (Punch, 1994).

Although this rather extreme example is in no way typical of the majority of social science research today, it raises a number of issues that, to a greater or lesser extent, face all researchers when engaging with other human beings in the research process. I would like to point out that although there is a school of thought that advocates that covert research is acceptable, the stance taken here is that research should be overt, meaning that all research participants have the right to know they are being studied and why they are being studied (Bell, 1999). Their knowledge of the research activity may well have an impact on their behaviour but that will be discussed in Part 2 as we look at individual research methods.

Introduction

I am not completely convinced that this is the right place in this book for a discussion of grounded theory; it is here because it is so often referred to as a ‘research method’. Grounded theory (see Example 16.1 on page 185) is an approach that uses ‘simultaneous data collection and analysis’ (Charmaz, 2006, 20) and as such can be applied in ethnography, case study, historical or action research. It is more about how data is collected and analysed than about the entire research design. Grounded theory is a process of qualitative analysis, not a research method; it is a ‘general method of comparative analysis’ (Glaser, 1978, 116). Although taking a grounded theory approach will influence the design of a research method – for example, in pure grounded theory the literature review very often comes after the analysis of empirical data (Charmaz, 2006) – it is not a research method in its own right.

Grounded theory is described as ‘a strategy for handling data in research, providing modes of conceptualisation for describing and explaining’ (Glaser and Strauss, 1967, 3). This is supported by Barney Glaser, who consistently refers to ‘grounded theory analysis’ and when making reference to the ways in which data can be gathered makes no reference to this as a research method (Glaser, 1998). The Discovery of Grounded Theory (Glaser and Strauss, 1967) offered a systematic approach to qualitative research practice, emphasizing that qualitative research had a logic of its own, and for the first time practical guidelines were offered in applying that logic.

Since 1967 the two originators have taken very different paths. Strauss (1987) and Strauss and Corbin (1990, 1994 and 1998) present a systematic and what some describe as ‘rigid’ approach to grounded theory analysis. Many claim their approach has moved too far from the original concept of grounded theory and suggest that the constant comparative method of Strauss and Corbin has changed the way data is used, from the concept of emerging theory, to theory being forced: ‘the technical tail is beginning to wag the theoretical dog’ (Melia, 1996, 371).

Introduction

The use of participant diaries as a data collection technique within LIS research has been largely overlooked, and where they are mentioned they are not looked on approvingly. Slater (1990) regards them as recorded self-observations and because of this places less value on their role in research in comparison with researcher observations. It is precisely the ‘systemization and editing by their creators, at the conscious or unconscious level’ (123) that Slater claims as a disadvantage, which provides the added data that may be valuable to the research. Ellis warns that diaries or logs rely ‘on the willingness and ability of the [research participants] to complete the diaries’ and that it is ‘questionable whether they would [be] able to [do] so comprehensively and accurately’ (Ellis, 1993, 475). Asking the participants in the research to keep logbooks or diaries of their behaviour can add to data collection in two ways: these records can provide confirmation of observations made by the researcher, and they can fill in gaps when the researcher cannot be present. Diaries can offer insight into the behaviour, feelings and thoughts of research participants. However, it is not easy to encourage participants to maintain dairies; you are asking for quite a commitment in terms of time and effort. Participants need to be convinced there is a good reason for giving this much of their time to your research.

Participant diaries are not the only form of diary in the research process; researchers’ diaries have long been used as a tool, although it is not strictly a data collection technique. I will discuss the use of researchers’ diaries in this chapter as I feel it is an essential part of the research process. It is rare to find researchers’ logs discussed in the methodology of any study but again this form of data collection has not received much attention in LIS research. Even in action research, which is heavily dependent on reflective thinking, we rarely find any real discussion about the use of a researcher's diary.

Introduction

Questionnaires are without doubt the single most popular data collection tools in any research involving human subjects. I usually begin any lecture on questionnaire design by saying that so much of what you are about to hear (or in this case read) would appear to be ‘common sense’. It's amazing how common sense can fly out of the window when researchers becomes so steeped in their subject and are eager to gather data that they lose their objective eye when it comes to critically analysing data collection instruments. A lot of the discussion in this chapter may appear to be ‘common sense’; if that is indeed the case then try to ensure that you use your common sense well.

There are a number of reasons for using questionnaires in your research: you can reach a large and geographically dispersed community at relatively low cost, you can harvest data from a larger sample than would be possible using any other technique, anonymity can be offered as well as confidentiality, and the data analysis can be determined from the outset, even as far as coding before the questionnaires have been distributed. Having said all of that, the most usual reason I am given by new researchers for the choice of questionnaire for data collection is ‘I won't have to talk to anybody’. This may well be a very good reason in the eyes of the neophyte researcher; it is not a good enough reason in terms of research credibility. Identify the purpose of your investigation and what you hope to do with your data once it has been analysed; base your choice on the aim of the research (although we all know personal preference will enter into the decision-making process).

Introduction

As I said in the introduction to Part 4, there is neither the space nor the necessity for me to provide a guide to each statistical process and test discussed in this chapter. There is a plethora of information available on these procedures, explaining how to perform them and indeed even explaining the software to perform them. Some internet sites will even perform certain tests on datasets without the need to purchase specific software. The purpose of this chapter is to present a discussion on the nature of numerical datasets, the appropriateness of statistical analysis and the fundamental principles involved in the procedures for this analysis. This chapter is concerned with techniques for processing and displaying data, ways of exploring datasets and analysis of relationships between datasets; it is not about how to do it. The discussion is concerned with when to do what to your data and why you do it. I would strongly urge that you use one or more of the recommended texts listed at the end of this chapter to become familiar with the processes involved.

It has become relatively simple to apply complicated tests to datasets using particular software; we discuss SPSS in more detail later in this chapter, and my concern is that you understand why you are performing these tests and what is actually happening to the data when you do it. I still teach statistical analysis in a ‘classroom’ setting with students, calculating many tests by hand before going on to use software for larger datasets, rather than in a computer laboratory using dedicated software; that comes later. I do this because I feel it is essential to understand what is being done if you are to provide a real description of what the results are showing us, so that you understand what has happened to your data and what your findings actually say about that data. You may now be pleased that I do not have the space here to go over these procedures but please take the time to understand the statistical procedures you apply to your data; even something as simple as a manual exercise using a very small dataset can clarify what is going on when you run a particular operation on your data in SPSS or some other dedicated software package.