Research awareness in nursing and midwifery: a workbook

journals are similar to indexes, but a short and often useful summary of each document (an abstract) is also given. Indexes and abstract journals may be in book form, although in the United Kingdom and some other countries they are now available on CD-ROM for computer use. In libraries where CD-ROM facilities are available, a librarian will usually show new users how to use it. You may also find relevant articles by scanning through copies of journals. Some journals are more likely to provide relevant articles than others. For example, the Journal of wound care might be expected to contain articles relevant to the topic identified above. page 6 Research awareness in nursing and midwifery: a workbook 2.4 Obtain selected articles, books and reports Abstracts will help you to identify which articles, books and reports are most relevant and worth obtaining. The documents obtained may provide further references that can be followed up. A librarian may be available to help you find your way around the library, to understand its classification system and to obtain other materials not kept in the library. Books and reports will usually have to be read in the library, although some libraries may allow you to borrow them. You may find it useful to make notes on them. The United Kingdom operates an inter-library loans system, which enables copies of books and reports and photocopies of articles to be obtained from other libraries. The librarians at major libraries in your country should also be able to obtain journal articles through this system. 2.5 Keep a record of each document Build up your own index of relevant documents. It is a good idea to write the details of each document on a separate card or sheet of paper and file them alphabetically (by author’s name) in a box, folder or other suitable holder. Use only one side of the card or sheet. The details could also be kept in a computer database. You need to include specific details that identify the document and its source. These details constitute a “reference”. For a journal article you need: • the name(s) and initial(s) of the author or authors; • the year of publication; • the title (and any subtitle) of the article; • the name of the journal in which it was published; • the volume and part numbers of the journal in which it was published; and • the inclusive page numbers. For a book you need: • the name(s) and initial(s) of the author or authors; • the year of publication; • the title (and any subtitle) of the book; • (for any book which is a second or subsequent edition) the number of the edition; • the place of publication; and • the name of the publisher. Reports, theses and specific chapters in books are referred to in a similar way. References are discussed in greater detail on page 14. Carrying out a literature search page 7 Examples and Exercise 1 Look at the reference list on page 71. Ref. 9 is an example of a reference to an article. Ref. 1 refers to a book, Ref. 4 to a chapter in a book and Ref. 16 to a report. Look at the other references in the list. Decide whether each reference refers to an article, book, chapter or report. Can you identify all the essential details of each reference, as given above? On your card or sheet of paper or in your database it may also be useful to record where you found the reference (e.g. in an abstract journal or in the reference list at the end of a specific article), where you can find a copy of the document (e.g. in which library, in which section of a particular library, or in your own collection of photocopies) and whether you have read it. For a book, it is also useful to include the International Standard Book Number (ISBN), a ten-digit number usually found on the back cover of a book and (along with other details) on the reverse of the title page. This may be particularly useful if you decide to order a book through a bookshop or by mail order. After you have read the document you could write details about its content and your critical evaluation on the reverse of the card or sheet, or add them to your database. Reading and evaluating research reports page 9 3. Reading and evaluating research reports Exercise 2 Using the information given in Section 2, obtain the three following journal articles. ALLEN, M. ET AL. Effectiveness of a preoperative teaching programme for cataract patients. Journal of advanced nursing, 17: 303–309 (1992). MCBRIDE, A. Health promotion in hospitals: the attitudes, practices and beliefs of hospital nurses. Journal of advanced nursing, 20: 92–100 (1994). KING, K.M. & JENSEN, L. Preserving the self: women having cardiac surgery. Heart and lung, 23(2): 99–105 (1994). Attempt to read these before you read the rest of Section 3, but do not worry if you find them difficult to understand at present. You will be asked to refer to these articles as you progress through Section 3. You will also be given some exercises relating to them. By the end of Section 3 you should have a better understanding of the three articles and of research. 3.1 Why critical evaluation is important As noted in Rundell’s article (9), not all documents that report research findings are of high quality; some have major flaws that weaken or invalidate the claims made by the authors. Some minor flaws may, however, be unavoidable and may not undermine the worth of otherwise rigorous research. An ability to identify both the strengths and weaknesses of research documents will enable you to: • identify good quality research documents relevant to your clinical situation; • recognize credible material that supports or challenges insights that you may have gained through rational thought, intuition, education or clinical experience; • make informed decisions about appropriate care, rather than relying on traditional practices or instructions from others, and implement these in consultation with patients; and • discuss the worth and relevance of research findings with nursing colleagues and other health workers in your clinical area. 3.2 Factors for critical consideration When reading documents critically you should pay particular attention to the following: • the biographical details of the researchers • the source(s) of research funding • details given in the abstract (if any) • the aims of the study • the literature review • the choice of research methods • pilot work • sampling procedures page 10 Research awareness in nursing and midwifery: a workbook • data collection methods • data analysis • presentation of data • discussion of data • conclusions and recommendations • references • ethical considerations • rigour • relevance to clinical practice. These factors are discussed individually below. Terms shown in bold italic are defined in the Glossary (Section 5, page 49). 3.3 Biographical details of the researchers Details about the researchers’ qualifications, current posts and/or academic background are sometimes provided. This is good research practice, as it facilitates critical consideration of the claims made in the document. For example, a team of doctors researching nurse education may employ different research methods and/or come to different conclusions than a team of nurses or a mixed team of doctors and nurses. Similarly, a member of the public and a nurse may be likely to take different research approaches and interpret data differently. 3.4 Source(s) of research funding Some small research projects conducted by practitioners do not have any outside source of funding. Much medical and nursing research, however, is funded by a variety of external agencies, including drug and other commercial companies, government departments and charities. Sources of funding should be explicitly stated in research documents so that readers can assess whether the nature of the funding agency might have influenced the choice of methods and/or the interpretation of the data. For example, it might be reasonable to be sceptical about research that claims that low-fat diets are unhealthy if it has been funded by a manufacturer of dairy products. Example If you look at the first page of each of the three articles you will see that all the authors have nursing qualifications, such as RN and RGN. Some also have academic qualifications such as PhD, MA and MN. The latter suggest that the authors have had training in research methods and/or experience of carrying out research. Some of the authors’ current posts are also given. Only article 2 by McBride gives enough detail to enable us to see how the current post of the author relates to the subject of the research. Only article 3 by King & Jensen acknowledges a source of research funding. We do not know whether the research projects described by Allen et al. (article 1) and McBride had any external funding and, if they did, which organizations provided it. Nevertheless, there are no obvious ways in which the source of funding might have affected the conduct or findings of the three research projects. If, for example, the project described by Allen et al. had tested a commercially available teaching package, we might have wondered whether the study had been funded by the company producing the package or by a commercial rival. Reading and evaluating research reports page 11 3.5 Details given in the abstract Not all research documents include an abstract (a short summary near the beginning). Abstracts are, however, always included in articles published in good quality health journals such as the Journal of advanced nursing and the British medical journal. University theses include an abstract, while some reports open with an executive summary (similar to an abstract). Books do not usually contain an abstract, although multi-author works sometimes have an introductory chapter by the editor(s) giving an overview of the material in the book. An abstract should provide an accurate summary of the content of the research document and should outline the main features of the research, such as the aims, methods, findings, conclusions and recommendations. Abstracts are useful in two main ways. 1. An abstract can often provide a quick assessment as to whether a document is relevant to your topic of search or your clinical practice. An abstract can therefore save time (and sometimes money), as it can help you to decide whether it is worth obtaining and/or reading a document. 2. An abstract may provide indications that a document has serious methodological flaws, or that the authors have reached far-fetched conclusions. Such suspicions can then be confirmed or rejected by careful reading of the main body of the document. Example All three articles reproduced at the end of this workbook have an abstract at the beginning. Read these three abstracts through. There are obvious differences in the style of presentation and the amount of detail given. The abstract of the article by Allen et al. is extremely short. It does not make the aims of the research or sampling procedures clear. Research design and data collection methods are not mentioned. It suggests that the findings are unequivocal (this will be discussed later). The abstract provided by McBride is much longer and this is reflected in the greater detail provided about aims, sampling, data collection, findings and recommendations. It does not, however, give a specific name to the research design. The abstract of the article by King & Jensen is the clearest of the three. This is partly because it is presented under separate headings (a presentational feature required by the journal publishing the article). We are given clear details under the headings Objective, Design, Setting, Participants, Results and Conclusion. You will better understand the issues involved in producing a good abstract as you read through the rest of Section 3. 3.6 The aims of the study The aims of the research should be made clear early in the document (usually in a section entitled Introduction or Background). The aims will probably have been influenced by the researchers’ previous experience and/or study of relevant literature and may be stated in terms of objectives, research questions and/or hypotheses. page 12 Research awareness in nursing and midwifery: a workbook Aims The aims of a piece of research are often expressed in a broad statement of intent, such as a description of the factors that may influence clinical outcome and satisfaction in patients treated for urinary incontinence. Objectives Objectives break aims down into specific elements to determine: • what treatments patients have received for their incontinence; • whether each treatment was given by a doctor, nurse, physiotherapist or other worker; • the perceived effectiveness of each treatment; • patients’ satisfaction with each treatment; • factors contributing to satisfaction or lack of satisfaction with each treatment; and • the clinical outcome of and patients’ satisfaction with their overall treatment. Research questions The term “research question” is sometimes used synonymously with “research topic” or “research problem”. More often, however, it is used to describe a specific question to be answered by the research. There may be a number of these questions for any one research project. The aims and objectives above could be expressed in terms of research questions of the following type. • What are the factors that influence clinical outcome and satisfaction in patients treated for urinary incontinence? • What treatments have patients received for their incontinence? • Was the treatment carried out by a doctor, nurse, physiotherapist or other worker? Hypotheses Certain questions may be phrased as research hypotheses. A hypothesis is a tentative statement about the relationship between two or more variables. The formulation of hypotheses is characteristic of quantitative rather than qualitative research (these types of research are defined in the discussion on choice of research methods below). A variable is any characteristic of a person, thing or situation that can vary. Age, weight, type of medication taken, smoking behaviour, life experiences and beliefs, for example, are all variables. In a hypothesis, a variable may be described as independent or dependent. The independent variable is the “presumed cause” and the dependent variable is the “presumed effect” (11). In the study on incontinence described above, factors related to treatment would be independent variables and clinical outcome and patient satisfaction would be dependent variables. In some studies, researchers manipulate the independent variable in order to produce an effect on the dependent variable. Examples of research hypotheses are that: • wounds heal more quickly by using lotion A than by using lotion B; and • cigarette smokers have higher rates of lung cancer than nonsmokers. In the above examples, type of lotion and smoking behaviour are the independent variables, while speed of wound healing and rates of lung cancer are the dependent variables. In Reading and evaluating research reports page 13 the first example, the researchers aim to manipulate the independent variable by using two different types of lotion that can then be compared. Certain variables may need to be defined. For example, when is a wound to be considered healed; and is a woman who has smoked one cigarette in her entire life to be classified as a smoker? If a hypothesis is to be subjected to statistical testing, a null hypothesis is sometimes formulated. This process is described in the discussion on data analysis below. Exercise 3 Read the Introduction to Allen et al., the subsection marked Aims in McBride and the opening (untitled) section of King & Jensen. How easily can you find the aims of each study and how clearly are the aims expressed? Are the aims stated in terms of objectives, research questions or hypotheses? Make some notes on your answers to these questions and then read the relevant discussion in Section 6. 3.7 The literature review Reviewing the literature allows researchers to set their research topic in the context of existing data and ideas. This enables them to demonstrate how their own work has added to knowledge on the subject. The literature review is usually presented in a specific section of the research document, but sometimes forms part of an introductory section or chapter. The review should discuss material relevant to the aims of the research. Occasionally, a literature search may reveal insufficient (or no) relevant nursing literature. In such cases the researchers should make this lack of nursing literature explicit, and may choose to review parallel literature from another discipline such as medicine or education. Similarly, international literature may be reviewed when there are no relevant documents from the researchers’ own country. Matters of context, such as differing work practices in medicine and nursing or the differences in health systems existing in different countries, should be acknowledged and their possible effects discussed. A good quality literature review should concentrate on reviewing original research documents (known as “primary sources”) rather than research described in other literature reviews (“secondary sources”). If all the references are more than five years old this may suggest that the researchers have failed to carry out a recent literature search or, more worryingly, that they have ignored recent documents because they provide new data that conflict with the researchers’ beliefs. It may, however, be difficult to determine the thoroughness of a review unless you are an expert on the research topic or you conduct your own search to check for comprehensiveness. A thorough literature review should attempt to present a balanced view by discussing the conflicting ideas of different researchers and authors. A literature review should not be just a list of documents published on a certain subject. As noted above, the ideas in different documents should be compared and contrasted. There may be gaps in certain areas of the literature that the authors should acknowledge. Sometimes the page 14 Research awareness in nursing and midwifery: a workbook researchers announce an intention to fill some of the gaps with their research findings. In addition, there should be a critique of the quality of the documents reviewed; significant methodological flaws, for example, may detract from the worth of data, conclusions and recommendations. The ability to critique research documents depends largely on skills developed through research awareness. Referencing systems There are two main systems in use for citing and referencing other work: the Harvard system and the numerical system. In the Harvard system, references are cited in the text by giving the name(s) of the author(s) followed by the year of publication. This may be done in one of two ways: • Polit & Hungler (1989) pointed out that ... • It has been observed (Briggs, 1972) that ... In the reference list, the references are set out in alphabetical order of the first author. If there are two or more references by the same author(s), these are given in chronological order. If there are two or more by the same author(s) in the same year, these are designated (and also in the text) as, for example, 1998a, 1998b, etc. In the numerical system (as used in this workbook), the references are numbered consecutively as they occur in the text. The names of the author(s) may or may not be given: • Polit & Hungler (11) pointed out that ... • It has been observed (5) that ... In the list, the references are set out in numerical order. Note that where there are several authors many publishers prefer to use “et al.” in the text to denote that there are more than a certain number, commonly two or three. Some publishers also use this system in the list of references. You may also come across the use of the term Vancouver system or Vancouver style. This is not a referencing system as such (although it uses the numerical system) but part of a set of standards, formulated by the International Committee of Medical Journal Editors, on how authors should prepare manuscripts for publication. Exercise 4 Can you locate the literature review in each of the three articles? Do the reviews discuss literature directly relevant to the aims of the research? Are the documents reviewed evaluated in terms of quality? Are the ideas in different documents compared and contrasted? Are most of the references recent (from the last five years)? Do the three articles use numerical or Harvard referencing systems? Write a few notes in response to these questions and then turn to the relevant discussion in Section 6. Reading and evaluating research reports page 15 3.8 The choice of research methods The aims, research questions and hypotheses should be restated at the start of the Methods section of a research document. This should help you to assess whether the methods chosen by the researchers were appropriate to meet the aims of the study. Quantitative or qualitative research? One of the choices that researchers may need to make is that between quantitative and qualitative methods. Put simply, quantitative research seeks to answer questions of “how much?” and “how many?” and is concerned with the relationships (especially causal relationships) between variables, whereas qualitative research is concerned with a matrix of “who, why, when and where?” questions and with the exploration of issues. Quantitative research often takes the form of experiments, quasi-experiments and surveys, which generate numerical data for tabulation and/or statistical testing. Qualitative research utilizes unstructured or semi-structured interviews, participant observation and/or the analysis of documents and conversations, often in combination. Qualitative data are in the form of words, such as interview transcripts and researchers’ field notes. Data are assigned to thematical categories, which are used as a basis for discussion. Quantitative research is associated with a philosophy known as positivism, which is based on ideas about the existence of objective reality and the ability of human beings to discover facts if they set aside their personal biases and take a scientific approach. Positivist philosophy and quantitative method underpin most research into the natural sciences (physical, chemical and biological experiments) and medical research (drug trials) and also some social science research (psychological experiments and social surveys). Qualitative research draws on a number of theories that can be loosely grouped under the term “interpretivism”. Interpretivists assert that, as human beings, we can never achieve complete objectivity and that we have to rely on a process of subjective interpretation in order to construct knowledge. Researchers working within an interpretivist framework believe that it is impossible for human beings to set aside their biases but that they can still be scientific (systematic and rigorous). Qualitative studies are sometimes described as ethnographic (exploring social/cultural interaction and meanings) or phenomenological (focusing on individuals’ experiences). Quantitative researchers are sometimes disparaging about qualitative research and vice versa. Some quantitative researchers describe qualitative research as unscientific, biased or anecdotal, and claim that it produces no “real knowledge”. Qualitative researchers, on the other hand, may believe that quantitative research seeks to reduce complex issues to simplistic explanations, fails to explore contextual matters and takes insufficient account of the effects of the researcher on the process and products of research. In addition, qualitative researchers might comment that, since all knowledge is based on subjective deliberations, there is no such thing as “real knowledge” but that quantitative researchers fail to acknowledge this. In general, doctors have tended to concentrate their efforts on quantitative research, while nurses have undertaken a mixture of quantitative and qualitative work. Many nursing researchers are in favour of multi-method research (sometimes known as triangulation), which has both page 16 Research awareness in nursing and midwifery: a workbook quantitative and qualitative elements. Some, notably Leininger (12), do not favour this because of the philosophical gulf between the two approaches. A qualitative exploration of a topic may generate hypotheses that can then be subjected to statistical testing in a quantitative study. Conversely, statistical relationships demonstrated in an experiment or a survey may suggest that a number of interrelated factors need qualitative explication. Some of the major differences between quantitative and qualitative research are presented in Table 1. You will learn more about these differences as you progress through Section 3. Table 1. Key differences between quantitative and qualitative research Quantitative research Qualitative research Underlying philosophy Positivism Interpretivism Underlying ideas about reality Reality is single and stable Realities are multiple, complex and everchanging Objective facts exist What is “real” is subjective Competing explanations for facts must be eliminated through control or acknowledged as limitations and flaws Complexity and contradictions cannot be explained away but should be explored Aims To make predictions and laws To explore phenomena and relationships To promote understanding To produce data that are generalizable to other situations To produce descriptions and theories that readers may find relevant to their own situations Research techniques Control Standardization Flexibility Consistency Research relationships Relationships with “subjects” must be limited and standardized to promote control and objectivity Relationships with “participants” are interactive The process and effects of interaction are often explored Sampling Statistical Random Purposive: to aid exploration Order of events Sampling Data collection Data analysis Sampling, data collection and data analysis are concurrent Data Numbers Statistics Words Ideas Some of the major forms of quantitative and qualitative research are described below. Experimental, quasi-experimental and survey work, all forms of quantitative research, are described first. Examples The articles by Allen et al. and McBride describe quantitative research studies, while that by King & Jensen describes a qualitative study. Reading and evaluating research reports page 17 Experimental research Experiments are a powerful way of testing cause and effect relationships between variables. Polit & Hungler (11) describe the three characteristics of a true experiment. (Note: people participating in experiments are sometimes described as “subjects”.) 1. Manipulation. The experimenter must do something to at least some of the subjects in the study. 2. Control. The experimenter must introduce one or more controls over the experimental situation, including the use of a control group. 3. Randomisation. The experimenter must assign subjects to a control or experimental group on a random basis. True experiments are sometimes known as randomized controlled trials. In the discussion on the aims of the study (page 11), you read about independent and dependent variables. In experimental research the researchers manipulate the independent variable (the presumed cause) in order to observe whether this has an effect on the dependent variable. The independent variable may involve a treatment, an information leaflet, a training course or any other device (sometimes termed the intervention) that might produce an effect on the dependent variable. For example, a treatment is given to some patients (known as the experimental group) and withheld from others (known as the control group) in order to see whether those who receive the treatment demonstrate a greater improvement than those who do not. Relevant data are collected from all the patients both before and after any treatment is given (pre-test and post-test). Since improvement may occur spontaneously, without any treatment being given, an experiment must involve some form of comparison. The ability to compare is provided by the control group (11): “The term control group ... refers to a group of subjects whose performance on a dependent variable is used as a basis for evaluating the performance of the experimental group (the group that receives the treatment of interest to the researcher) on the same dependent variable”. The control group may be given the conventional treatment for the condition or a placebo (such as tablets with no active ingredient or a half-hour visit but no therapy) or no treatment at all. Some study designs involve more than one control group such as: • experimental group vs. placebo group vs. no treatment group; or • experimental group vs. conventional treatment group vs. placebo group. The latter design has the ability to demonstrate that the experimental treatment is not only more effective than the placebo but is also more effective than the conventional treatment. Other factors that could affect the dependent variable need to be controlled. For example, the temperature in a laboratory or room might affect biochemical, physical or psychological test results. For this reason, the temperature needs to be constantly controlled throughout the experiment. In a drug trial, patients might take other medicines that could affect the medical condition under study. Such patients must either refrain from taking these medicines or be excluded from the study. A study that has body temperature as a dependent variable must take account of the fact that this varies with the time of day. Body temperature should, therefore, be page 18 Research awareness in nursing and midwifery: a workbook measured at the same time (or times) of day. These additional factors that can affect the dependent variable are known as extraneous variables. Randomization, the third essential element of an experiment, is performed to prevent the researcher(s) from distorting the results through “systematic bias”. The random assignment of individuals to groups aims to make the experimental group and the control group comparable in terms of characteristics such as age, gender, educational status and other factors that could affect the dependent variable. Randomization procedures can be performed by simple means, such as drawing names out of a hat or generating random numbers by computer. There are several types of experimental design. Two common ones are: • the comparison of independent groups; and • the comparison of repeated measures carried out on the same group. In the latter case the experimental group acts as its own control group. The “cross-over” study is a common form of the repeated measures design. In this, participants are divided into two groups. Group I receives treatment A followed by treatment B and group II receives treatment B followed by treatment A. The advantages of repeated measures studies are that they require fewer participants (as each individual is used twice) and that the characteristics of the experimental group and the control group are the same (as they are composed of the same people). The major disadvantage is that it can be claimed that data collected during the second treatment period have been influenced by the fact that there was a previous treatment period. It is possible to carry out some experiments, such as drug trials under double-blind conditions. This means that neither the researchers nor the participants know which is the experimental group or treatment and which is the control group or treatment. Thus, in a doubleblind cross-over drug trial, neither the researchers nor the participants know whether the experimental medication has been administered during the first or the second treatment period (placebo medication is given in the other treatment period). The study is only “unblinded” once data collection and analysis are complete. Quasi-experimental research A quasi-experiment does not possess all three essential attributes of a true experiment, as it lacks either randomization or a control group. There are numerous questions worth researching where randomization is not possible. For example, a researcher may wish to determine the physical and psychological effects of a new shift-work system on nurses working in a particular hospital. Since the new shifts are being introduced for all nurses throughout the hospital, randomization is not possible. It may, however, still be possible to collect pre-test and post-test data and to use the nurses at a similar hospital, which is keeping the old shift system, as a comparison group. In some studies, however, it may not be possible to find an appropriate comparison group. Quasi-experimental designs do not lend themselves to the testing of hypotheses concerned with cause and effect because of the presence of too many uncontrolled factors. It is, however, possible to test whether variables are correlated, i.e. related to each other in a statistically significant way. Survey research Surveys can be used to collect information about variables and, sometimes, to test relationships between variables. There is no experimental intervention in survey work; rather information is Reading and evaluating research reports page 19 collected about (for example) people’s health status, living conditions, health beliefs or satisfaction with services. The information is turned into numerical data, which can be subjected to statistical analysis. Survey data are obtained by means of questionnaires or interviews, using a highly structured interview schedule. Interviews may be conducted face-to-face or by telephone. You can read more about questionnaires and structured interviews in the section on data collection methods on page 24. Some surveys are designed to gather information from the entire population of (for example) a country or town or all the patients who attend a particular health clinic. A survey of an entire population is known as a census. Censuses may be costly and time-consuming to conduct, however, especially if the population under study is large or widely spread. For this reason, it is more usual for part of the population to be surveyed (known as sampling). Assumptions are then made about the health, living conditions, etc. of the entire population on the basis of the data provided by the sample. Issues relating to the size and representativeness of survey samples are discussed in the section on sampling procedures (page 21). Exercise 5 Read the section entitled The study in the article by Allen et al. and that entitled The research project in the article by McBride. Do these describe experiments, quasiexperiments or surveys? Write down your answers and then turn to Section 6 for discussion. In the article by Allen et al., what were the dependent variables? Can you identify any uncontrolled extraneous variables? Was it a double-blind study? In the article by McBride, what was the study population? Write down your answers and then turn to Section 6. Qualitative research It is fairly easy to divide quantitative research designs into methodological types, such as surveys and experiments, but this is not the case with qualitative research. Qualitative research is associated with a number of overlapping philosophies and methodologies such as naturalism, interpretivism, ethnography and phenomenology. It is not uncommon for qualitative researchers to draw on all of these and it may not be easy, for instance, to determine why a study has been described as ethnographic rather than phenomenological. Hammersley & Atkinson (13) give a definition of naturalism: Naturalism proposes that, as far as possible, the social world should be studied in its ‘natural’ state, undisturbed by the researcher. Hence, ‘natural’, not ‘artificial’ settings like experiments or formal interviews, should be the primary source of data. Furthermore, the research must be carried out in ways that are sensitive to the nature of the setting. Interpretivism incorporates the above ideas but also stresses that, even when the social world is studied in its “natural” state, researchers can only understand it through their own five senses and a process of subjective interpretation. Because of this, researchers never have direct page 20 Research awareness in nursing and midwifery: a workbook access to the experiences, understandings and meanings of the people under study. Thus, research documents can represent the views of the people under study but only as interpreted by the researchers. As noted towards the beginning of this section, ethnography (a term with roots in sociology and social anthropology) is primarily concerned with social interaction and meaning, whereas the main focus of phenomenology are the experiences of individuals. Some researchers, however, use the terms “ethnographic” and “phenomenological” as synonymous with “qualitative”. A qualitative approach can enable researchers to explore issues in depth. An experiment may show that a medication is clinically effective, but it is less suited to exploring the experiences, opinions and feelings of patients who have had to decide between having an operation or taking long-term medication for a chronic illness. Data from a quantitative survey may suggest that there is a statistical relationship between urinary incontinence and low selfesteem, but an in-depth qualitative approach may be required to draw out how these two factors relate to each other and how the lives of incontinent people might be improved in both physical and emotional terms. Relatively unstructured data collection methods, such as participant observation and semi-structured interviews, are used in qualitative research so that everyday experiences can be studied and research participants can put forward their own ideas in their own words. Qualitative researchers (particularly ethnographers) often draw on grounded theory (14) when undertaking sampling procedures, data collection and data analysis. This ensures, as far as possible, that the theories that the researchers construct are grounded in the data obtained (rather than in the personal interests or the previous knowledge of the researchers). Grounded theory is discussed further below in the sections dealing with sampling procedures, data collection methods and data analysis. Exercise 6 Read the Method section of the article by King & Jensen. Is the study described as qualitative, ethnographic or phenomenological? Turn to Section 6 for discussion. Other forms of research Research can be described and categorized in various other ways, some of which cross the quantitative/qualitative divide. Two of these types of research will be discussed briefly: feminist research and action research. Feminist research is particularly relevant to nursing because the vast majority of nurses are women. In addition, because women often take care of children and dependent relatives they come into contact with the health care system on behalf of others as well as themselves. Feminist research can be either quantitative or qualitative. It frequently focuses on gender relations between doctors and nurses, gender issues in the care and treatment of women patients, and issues related to caring. Reading and evaluating research reports page 21 Action research is an approach in which researchers and research participants work together to solve problems and effect change. Data are collected at the start of the study and used as a basis for planning change. The process of change is monitored and plans modified if necessary. At the end of the study a final assessment is made. The process (as well as the outcome) of change is important in action research: during it researchers enable participants to learn about research and to gain control of their work situation or their care. Action researchers often take a multi-method approach combining quantitative and qualitative methods. Example The article by King & Jensen provides an example of feminist research, although this is not explicitly stated in the article. 3.9 Pilot work It is good practice in quantitative research (and sometimes in qualitative research) to carry out a pilot study. This can help to eliminate problems related to access, sampling and data collection instruments that might otherwise arise in the main study. Polit & Hungler (11) describe a pilot study as “a small-scale version, or trial run, of the major study”. Sometimes, if the problems arising are minor, data from piloting work can be included in the main study data. This frequently happens in qualitative studies where, in any case, data collection instruments tend to evolve as the main study progresses. Research documents should state whether a pilot study was performed and give some idea of any problems encountered and subsequent modifications made to plans for the main study. Exercise 7 Refer to the three articles. Did the researchers involved in the three studies conduct any form of pilot work? If so, what modifications were made for the main studies? Note down your answers and then turn to the discussion in Section 6. 3.10 Sampling procedures Sampling procedures tend to be quite different in quantitative and qualitative research. Whereas the aim of sampling in quantitative research is to select a sample that is representative of the study population, the aim in qualitative research is to select participants who will allow a detailed and in-depth exploration of the issues under study. Sampling in quantitative research Probability sampling is the most effective way of selecting a sample that is representative of the population from which it is drawn. In probability sampling every element (person or thing) in the population under study has an equal chance of being selected for the sample. There are three main types of probability sampling. page 22 Research awareness in nursing and midwifery: a workbook 1. Simple random sampling uses a computer, a random number table or numbers drawn at random to select a sample from the sampling frame (a numbered list of the entire population under study). 2. Stratified random sampling divides the population into two or more strata, such as male/female or nursing assistant/student nurse/qualified nurse, and a random sample is then drawn from each stratum. Stratified sampling can be proportional or disproportional to the population under study. Proportional stratified sampling is used to make the sample more representative, whereas disproportional sampling can ensure that minority groups are represented in sufficient numbers to allow meaningful statistical analysis. 3. Cluster sampling can be used in large-scale studies where compiling a sampling frame of the entire population might prove costly, difficult or impossible. Compiling a sampling frame of all the diabetic patients in the United Kingdom, for example, would be difficult and expensive. It would be easier and cheaper to select a random sample of health clinics, thus giving access to “clusters” of diabetic patients, all of whom (or a random sample of whom) could be selected depending on the sample size required. It is important, when using a sampling frame, to be sure that it indeed represents the population that the researchers wish to study. For example, a register of nurses registered to practise is not a register of nurses in practice. As implied above, there are no existing sampling frames for some populations. Compiling a nationwide database of diabetic patients or nursing assistants is a major piece of work in itself. Because of the expense, time and difficulty that may be involved in probability sampling, non-probability sampling is often employed. This has the disadvantage of being less representative. There are three main types of non-probability sampling. 1. Convenience sampling adopts the most readily accessible people or things as the sample. Thus, a nurse performing a survey may decide to distribute questionnaires to the first 200 patients with leg ulcers seen at home or in a clinic. Some experiments, such as drug trials, may utilize convenience samples: all available patients meeting the inclusion criteria and agreeing to participate are recruited. Randomization occurs only at the point of assignment to the treatment or control group. 2. Snowball sampling employs a questionnaire or a request to participate. This is sent to known contacts, who are asked to provide names of other people who will agree to participate. These, in turn, are asked to provide further names and so on. 3. Quota sampling involves the researchers identifying strata in the population and selecting a convenience sample that reflects the proportions of those strata. Thus, if a particular population is composed of 400 males and 600 females, researchers might aim to survey 40 males and 60 females. It might also be composed of 500 nurses, 250 physiotherapists and 250 doctors, in which case 50 of those chosen would be nurses, 25 physiotherapists and 25 doctors. The aim of quota sampling is to make more representative a sample that would otherwise be a convenience sample. When critiquing non-probability sampling designs in research documents, you should consider whether the researchers might have been able to select a more representative sample by using an existing sampling frame (such as doctors’ lists of patients or a national register of nurses) and employing random sampling techniques. Reading and evaluating research reports page 23 Determining a suitable sample size can be difficult. The larger the sample the more likely it is to be representative of the population from which it is drawn. However, the larger the sample the greater the resources needed for data collection and analysis. Where limited resources have necessitated the adoption of a small sample (and/or of non-probability sampling) this should be acknowledged in the research document. If a small, non-probability sample has been used, researchers should be extremely cautious about claiming that their findings can be extrapolated to the entire population. Whenever possible, and prior to sample selection, researchers should take expert advice about sample size from a statistician (this too should be noted in the research document). Exercise 8 What types of sample were used in the studies described by Allen et al. and McBride? Do you think that the sizes of the samples were adequate? Do the authors say whether they took statistical advice on sample size? Make some notes and then turn to the discussion in Section 6. Sampling in qualitative research Qualitative researchers rarely employ random sampling techniques, but neither do they claim that their data and interpretations are generalizable to larger populations. They do, however, attempt to highlight and explore issues that may be interesting and relevant to people in settings and contexts other than that in which the research was carried out. Thus, a study of the morale of staff at one hospital may be of interest to staff at a second hospital; an exploration of doctors’ communication skills may throw up interesting points about communication in nursing; cancer patients may identify with some of the experiences, thoughts and feelings presented in a study of AIDS patients. In qualitative studies, sampling is often described as purposive or judgement sampling. This means that sampling has been conducted with a particular purpose in mind or on the basis of the researchers’ judgement. In other words, both terms mean that researchers choose a sample that should enable a thorough exploration of issues relevant to the aims of the study. Sampling decisions in qualitative research include the selection of a setting or settings that will allow exploration of the relevant issues. For example, researchers might be interested in how various aspects of policy, practice and staff relationships affect the recovery of patients who have undergone surgery for cancer. They may decide to study just one hospital ward or two or more wards at different hospitals, which can then be compared and contrasted in respect of the issues under study. Once a setting or settings have been chosen, there are decisions to make about who to interview and who or what to observe within the setting(s). This may involve decisions about, for example, which grades of nursing staff to interview, which doctors to interview, whether patients with particular types of cancer might have different things to say, whether to observe all ward rounds, or whether to observe staff in conversation during meal breaks. “Theoretical” sampling is a specialized form of purposive/judgement sampling adopted by researchers who take a grounded theory approach to their work. In theoretical sampling, once a setting has been chosen, researchers make most of their sampling decisions as the research page 24 Research awareness in nursing and midwifery: a workbook progresses. This allows them to select interviewees, for example, who may facilitate exploration of theoretical concerns identified in the data already collected and analysed. Sampling can cease when no new concepts are being generated from the data or, to put it in grounded theory terms, when theoretical categories are “saturated”. This is one way in which sampling, data collection and analysis are linked in grounded theory. When writing research documents, qualitative researchers should make the thinking (including any theoretical influences) behind their sampling decisions explicit. Exercise 9 How are sampling procedures described by King & Jensen? Do you think there are any flaws in the sampling procedures described? Make some notes and then turn to the discussion in Section 6. Data collection methods Quantitative and qualitative researchers also tend to utilize different data collection methods to pursue their aims. Quantitative methods are designed to promote standardization and control, whereas qualitative methods aim for a balance between flexibility and consistency (15). Quantitative data collection methods Data collection techniques (such as tests) and devices (such as questionnaires and interview schedules) are sometimes known as instruments. The most commonly used instruments in the collection of quantitative data are: • tests and measurements (often used in experimental studies); • questionnaires and structured interview schedules (often used in surveys); and • structured observation schedules. In well designed quantitative studies, data are recorded in tightly structured formats, which may be in the form of questionnaires or interview/observation schedules. In the case of questionnaires, data forms are completed by research participants (sometimes known as respondents), while in other cases the researchers complete the forms. It is good practice to include a copy of the survey questionnaire, observation schedule or other data form in the research document although, in the case of research articles, the need for brevity may prevent this. Inclusion of the data form may aid comprehension of the research document and will allow readers to evaluate the quality of data collection procedures and the validity and reliability of the study (see the section on rigour, page 42). Experiments and quasi-experiments may involve the collection of several different types of data. For example, data collected during a clinical drug trial might include objective biophysiological measurements such as blood pressure levels and body weight, patients’ subjective reports of their symptoms, and the presence or absence of any confounding variables such as other medication taken. Patients may be weighed and have their blood pressure taken each time they visit the person managing the trial. They may record their symptoms daily on a structured diary card (using, for example, the terms “mild”, “moderate” or “severe” to describe pain), which they hand in at each visit. They may be verbally questioned about their use of other Reading and evaluating research reports page 25 medication, concurrent illness and other potentially confounding variables. In other words, the data collected during the course of a clinical trial may involve biophysiological measurement, questionnaire and interview techniques. Data will be recorded on a data form at each visit. In a quasi-experiment designed to assess the effect of an educational package, data on nurses’ knowledge may be collected preand post-test using questionnaires or structured interviews. When critiquing experimental and quasi-experimental data collection techniques, you should attempt to assess whether data have been collected on potentially confounding variables. It is also important to consider ethical issues (discussed on page 40). This may be particularly relevant where biophysiological measurements are concerned, as some tests may be painful, distressing or even dangerous. Exercise 10 What research instruments were used in the study described by Allen et al.? Are copies of these included in the article? Turn to Section 6 for discussion. Questionnaires may be distributed to a survey sample by post or, for smaller studies, by hand. In either case, attempts should be made to preserve respondents’ anonymity. Individual questionnaires are usually identified by a code number rather than a name and, even where distribution has been by hand, questionnaires are often returned in a prepaid envelope. Questionnaires should be designed with a high proportion of closed (“yes/no” or “tick box”) rather than open (“give details below”) questions. Questions should be arranged in a logical order and should avoid complexity and ambiguity. Pilot work (as discussed above) can help researchers to identify and eliminate questionnaire design flaws. Structured interviews are similar to questionnaire surveys, except that researchers (sometimes called “data collectors”) obtain the data from participants by reading questions from and recording answers on an interview schedule. This can be done face-to-face or over the telephone. To preserve a degree of anonymity, completed schedules usually bear a code number rather than a name. The issues involved in designing an interview schedule are very similar to those involved in questionnaire design. Although data collectors are able to clarify questions that participants do not understand, they must be careful that they do not bias the data when doing so. Each participant should be asked the questions in a standardized way. This is extremely difficult to achieve in practice, particularly when there are two or more data collectors. Training may help to minimize some of these problems. Some research topics might be better dealt with by questionnaire, some by telephone and some face-to-face. For example, people might be willing to complete a questionnaire about a potentially embarrassing subject, such as incontinence or impotence, but might not wish to talk about it. As factors such as this may affect the response rate (see under Data analysis, page 27) as well as the quality of the data, they should be taken into consideration when critiquing research documents. page 26 Research awareness in nursing and midwifery: a workbook Exercise 11 Is the research instrument included by McBride? What conclusions can we draw about the quality of the instrument? Turn to Section 6 for discussion. Observational research takes two main forms: participant observation, which is usually employed in qualitative studies (discussed below), and non-participant observation, which is more common in quantitative work and may be known as “structured observation”. Structured observation can take several forms, all of which involve the assignment of behaviour to categories recorded on an observation schedule. Activity sampling is a common form of structured observation: for example, every 15 minutes the observer records what each member of the ward staff is doing. When selecting a “time frame” for activity sampling, researchers must avoid choosing an interval that is too short to allow the observer to record all the necessary data, or is too long so that important events are missed during non-observation periods. Other difficulties with observation include the problem of reactivity: people may alter their behaviour if they know they are being observed (reactivity is also an issue in experiments and interviews and is sometimes known as the Hawthorne effect). On the other hand, it may be unethical to observe people without their consent, because they have a right to decide whether they wish to participate (see the section on ethical considerations, page 40). All these factors should be taken into consideration when critiquing observational studies. Exercise 12 Did the study by Allen et al. involve an observational component in the assessment of skills? Turn to Section 6 for discussion. Qualitative data collection methods Qualitative researchers, particularly ethnographers, often “draw on a wide range of sources of information” and collect “whatever data are available to throw light on the issues” (13). Some qualitative studies may rely primarily on participant observation or semi-structured interviews, but researchers often use a combination of the two and may also draw on conversations, written documents (such as nursing care plans or policy statements) and other sources. Participant observation is, perhaps, the most important data collection technique in ethnographic research. Observers may participate in the work, social activity or culture under study (for example, nursing work) to a variety of degrees. In addition, the degree to which they are “insiders” or “outsiders” to the research setting may vary. Complete insiders study their own work or social settings; complete outsiders are strangers to the types of setting they study. Researchers who study settings similar to their own can be considered as relative insiders/outsiders. There are numerous difficult issues that can arise in relation to participant observation. One such issue is whether it is ethical to do covert research (pretending to be a participant and not a researcher). There is also a problem in uncritically identifying with and even adopting the culture Reading and evaluating research reports page 27 of a group of people under study. It may also be very difficult to manage relationships: some people, for example, can be so unhelpful or threatening that researchers choose to avoid them. The research setting (sometimes known as the field) will have been selected as an appropriate place for exploring the issues of interest and, thus, fulfilling the aims of the research. Quite commonly, however, the issues under study (which are initially determined by researchers’ previous experiences and/or their knowledge of the literature) undergo some degree of change as field work progresses. This is particularly the case with researchers who take a grounded theory approach: theories emerging from the data already collected determine who and what should be observed next. Participant observers record their experiences in the field, and their reflections on them, in field notes. In ethnographic studies, data gained through observation may influence decisions about who to interview and the themes to be pursued at interview. Qualitative interviews are commonly described as unstructured, semi-structured, conversational or focused. They differ from structured interviews in that interviewers work from an interview schedule (or interview guide), which has topics for discussion or a few open questions that will get interviewees talking about the relevant issues. Interviewers aim for a balance between flexibility and consistency (15), allowing interviewees to introduce their own ideas and concerns while trying to pursue themes central to the aims of the research. The interview guide usually evolves during the course of a series of interviews. Lines of inquiry that have proved unproductive are dropped, while key themes introduced by previous interviewees are added. This, again, accords with a grounded theory approach. Qualitative interviews are often recorded to save interviewers from attempts to listen and write down large amounts of data simultaneously (each interview may last an hour or more). In order to demonstrate rigour, a document describing a qualitative study should make the decisions involved in data collection procedures explicit. Issues such as failure to manage relationships in the field may affect the rigour of the work: this should be acknowledged by the researchers. In addition, there may be ethical issues to consider. Exercise 13 What data collection methods were used in the study described by King & Jensen? Are there any flaws in the data collection methods or the reporting of them? Make some notes and then turn to the discussion in Section 6.