There are two main types of research; quantitative and qualitative. Quantitative research involves the analysis of numerical data in order to explain a pre-defined hypothesis. The researcher aims to remain objective and control the conditions of the research. Qualitative research involves the analysis of words, pictures or objects in order to explore and describe a situation, process or experience. The researcher interprets the data as it unfolds during the collection process and the subjective nature of the process is embraced.

Behind these types of research are different philosophical assumptions and different approaches to examining the research question. It is beyond the scope of this book to examine these differences in detail but I would recommend that you explore them, and the Resources section will guide you to useful text books. Each type of research has its own advantages and disadvantages and will answer different types of question, which I will describe in the next sections. When making your decision, think about what your chosen method will allow you to find out and also what it will not allow you to explain.

The main approaches within quantitative research are experimental and observational studies. These approaches test theories and produce numerical data which are statistically analysed. The intention is always to use a sample of a population and then make inferences about the whole population from the results of the smaller samples. Data can be collected prospectively or retrospectively; in other words from the start of the study forwards or searching back through previously collected information. Data can be explored within the study sample longitudinally, to study changes over time, or crosssectionally, to observe an occurrence at one point in time only.

In experimental studies we are trying to establish cause and effect. We want to show that a particular intervention causes particular effects. This is the basis of the evaluation of new therapies or procedures in health care to demonstrate effectiveness. To do this we first need to compare the new therapy (applied to the intervention group) to something else, this could be no therapy at all or the therapy that is in normal use. This is applied to the control group in the experiment.

We next need to decide what piece of information will tell us the therapy is effective and decide exactly how to measure it. This is known as the outcome, and I have discussed this elsewhere in Chapter 7, Section 7.2.6.5.

In order to ensure that the changes in the outcome measurement are due to the experimental therapy we need to consider bias – the main purpose of the experimental design is to minimise bias. Firstly, the control group should be comparable to the experimental group in all ways other than the experimental treatment. This is ideally achieved through randomisation, which ensures that the unmeasured and unmeasurable characteristics are equally distributed between the groups. Randomisation means that each person entering the study has an equal chance of being assigned to either the control or the experimental group. Randomisation is achieved through using random number tables or randomisation schedule generators on the internet (see Resources section). Despite randomisation it is still important to check that both groups are comparable, so we need to collect information about any factors that could potentially alter the response to the experimental treatment.

Another source of bias is through unconscious favouring of the intervention group over the control by both the researcher and the participants. To get round this problem, blinding is used. Single blinding means that either the researcher measuring the outcome (the assessor) does not know which group each participant is in or the participant is unaware of the group they are in. Double blinding means that neither the participant nor the researcher knows which group the participant is in. A placebo is commonly used in drug trials and sometimes in other types of therapy, and is designed so that the control group think they are receiving the same treatment as the experimental group. This allows the psychological effect of simply receiving a treatment to be taken into account. In many therapy trials it is not possible to use a placebo or achieve double blinding and in this circumstance it is crucial that every effort is made to blind the assessor to the treatments, so that the measurement of the outcome that demonstrates the effectiveness of the treatment is subject to the least bias possible.

Bias can also occur if the method of measuring the outcome is difficult and prone to variation, for example blood pressure or ultrasound measurements. This variation can be minimised by taking several measurements and using the mean of the two or three measurements; this is known as replication.

The preceding description describes the components of the randomised controlled trial, which, as I have discussed in Section 6.1, is regarded as the best form of experiment. The absence of any of these essential elements reduces the strength of the evidence produced by the research. For example, the absence of randomisation reduces the design to what is termed a quasiexperimental design, where there is a control group but the treatment is not assigned randomly. In this case it is difficult to argue that the control is similar to the experimental group and there may be differences between the two groups that could account for any effect observed. It is always better to plan a proper experimental trial and if this is not possible you need to be able to justify why a quasi-experimental approach is to be used.

I came to qualitative research later in my research career. Qualitative research does suffer an image problem with researchers strongly oriented to the quantitative approach; however, this is slowly changing and the value of qualitative work is being increasingly recognised. Quantitative methods fail to provide information on the context of the studied situation, and is limited by the highly structured format and a high level of control put on the situation of study.

Qualitative research can provide a much more holistic view of the problem under investigation, including information that cannot be reduced to numbers. The approach to the data collection and design of the study is more flexible, as is the subsequent analysis and interpretation of collected information, allowing the researcher to evolve their view of the problem under investigation and adapt to the changing view. The qualitative approach also allows the researcher to interact with the research subjects in their own language and on their own terms. This approach gives a deeper understanding of the problems faced in health care and can help evaluate treatment in terms of appropriateness and feasibility discussed in Section 6.1.

Phenomenology is an approach that aims to explore the ways in which people conceive of and interpret the world. In this view, reality is relative and subjective (Stokes 2003). Phenomenology informs and helps the understanding of a particular phenomenon from a particular view point. An example is an investigation of the ways in which patients report their satisfaction with health care (the phenomenon). The researcher used in-depth interviews to explore the process of reflection that patients went through when asked to appraise their satisfaction with received health care (Edwards, Staniszweska & Crichton 2004). The interviews provided descriptions of a range of experiences from different patients, from which an essence of the phenomenon and a generalised statement about it could be derived. A critical aspect of phenomenology is the need for the researcher to suspend their own personal views or preconceived ideas about the phenomenon (known as bracketing), and seek to understand it through the views of the participants (Cresswell 1998). The aim of this type of inquiry is to show that although the phenomenon is experienced uniquely by each person, there exists an underlying unifying meaning of the experience that is essential and invariant for all people.

Grounded theory is the systematic inquiry into a problem aiming to develop an overall theory based on personal experiences. The emergence of new data can alter or destroy the theory, so extensive data collection is required. As the theory emerges the researcher must constantly ensure it fits new data. An example is the exploration of the barriers to evidence-based nursing in order to develop a model of what barriers exist and how they relate to each other (Hannes et al. 2007). Hannes et al. used focus groups to collect data from nurses and reduced down the responses to five main themes that helped to formulate, what they called a ‘problem tree’, describing the interrelating issues.

Discourse analysis is a research approach that focuses on how talk and text can indicate how the world is understood. Studying conversations or written communications within a particular area can aid the understanding of people’s beliefs, views and attitudes towards something. An example is the analysis of conversations between pharmacists and older patients during consultations for medication review (Salter et al. 2007). The researchers analysed in detail the transcribed conversations to identify how advice was offered and how patients received it, to establish the value of such consultations.

The ethnographic researcher will observe and collect data on a group of people within their usual setting over a prolonged period of time. Ethnography is a form of inquiry that relies particularly on participant observation, focusing on the meanings of individual’s actions in a particular context. The context is important to enable a true understanding of events. For example, studying the decision-making process prior to withholding artificial administration of food and fluids involved two researchers becoming part of the health care teams at two nursing homes for seven and fourteen months (The et al. 2002). They made comprehensive notes on their observations and informal conversations; recorded formal interviews and meetings; studied medical records of the patients; and kept a diary of their own behaviour and attitudes. The latter process is vital in qualitative research and allows the researcher’s own preconceptions and experiences, as well as the effect of their presence, to be taken into account in the interpretation and analysis of events, and is known as reflexivity.

Narrative research is an approach which relies on the stories that people tell. Narratives are structured by the individual telling the story and will be organised in the way they see fit. How someone tells a story reflects how they have chosen to organise their thoughts and could be seen as the basis for how they acted. The order of the narrative will reflect the meaning that the storyteller wants to get across to the researcher. In a study to explore the patients’ experience of living with back pain, participants were asked to tell their own back pain story from when they first encountered such pain up to the current day (Corbett, Foster & Ong 2007). These narratives were analysed to produce themes which represented commonalities between the stories, and used to try to better understand the patients’ experience of this condition.

Matching the research approach and specific design to your defined research question is the next step in the research process. The preceding sections have provided a broad overview of the available methods you could use and you now need to consider which will provide the answers you are looking for. This is easier when your research question is clear and you may need to think through your question again at this point.

Think about the following: