Actually, there can have been very few clinical problems in nursing that have been solved by research; nevertheless, it was precisely how one of us (Watson) came to be involved in nursing research. In the area of urinary incontinence it was clear that some older men with dementia reached a stage where preventing the incontinence became unrealistic and, indeed, interventions aimed at alleviating incontinence became more distressing for the older men than managing the urinary incontinence (Watson 1989 and Watson 1990). Clinical experience with urinary sheaths was varied and there were also several products on the market being offered at varying prices. We were able to demonstrate the superiority of one product and also that, despite its greater unit cost, it was better and proved to be cheaper to use as fewer sheaths were required and patient comfort and safety were increased (Watson 1989 and Watson 1990).

The feeding difficulty of older people with dementia proved to be a more intractable problem to ‘solve’ by research. The literature on feeding difficulty in older people with dementia was reviewed and an assessment instrument to measure feeding difficulty was designed. This body of work is well published and the instrument is known as the EdFED Scale (Watson 1996).

In a different vein from research aimed at providing solutions to specific problems, nurses also carry out research to evaluate their practice. This may be done to see if some change in clinical practice that has been implemented is working or to find out, based on current practice, if any changes are required. Clearly, such research may help to identify problems in clinical practice that need to be solved by further research. In fact, this is not uncommon – research rarely provides answers; it usually provides more problems.

Closely related to research that evaluates practice is the use of audit, a key feature of UK NHS clinical governance (Knight & Hostick 2004), and many are confused about the distinction between audit and research. One way of looking at audit is to consider it as a means of evaluating, for the purposes of management, agreed and standard aspects of clinical practice. On the other hand, if data are being gathered to evaluate clinical practice based on someone’s hypothesis that a particular result may be found and that the result may merit further investigation to seek improved practice, then this is definitely research. It should be noted that audit and research are not distinguished by the methods employed to carry them out – audit should be as rigorous as research.

Policy and practice are not entirely distinct as a great deal of practice stems from policy. However, policy often arises for purely political reasons and it is common for policy changes to be implemented prior to any research being conducted to inform that policy change. Thus, policies need to be evaluated after their implementation. Whereas investigating changes in clinical practice may involve a very limited range of colleagues, policy evaluation commonly involves a wider range because practice implications, economic implications and social implications may all need to be included in the evaluation.

A word of caution is required to those who would become involved in policy evaluation. Due to the political nature of most policy, there is a preconception by those who have implemented the policy and also commissioned the evaluation, that the policy is good and that the changes implemented as a result of it are beneficial. This leads to the possibility of bias in the research and also to difficulty in publishing negative results. Cadet nurses were reintroduced on a national scale in England following the publication of Making a Difference (Department of Health 1999) with the aim of widening access to the study of nursing and widening the ethnic and socio-economic background of the nursing workforce. However, an evaluation of this policy showed mixed results and it is unclear what the future of cadet nursing programmes will be (Watson et al 2005).

The purpose of research is to generate knowledge and insights into our world. However, the nature of the inquiry will vary depending upon the stance that we take to view the world and our understanding of ‘truth’ as being either universal or context-specific. For example, if we see the world as being ‘external and concrete’, then to generate and test theory based on this belief structure, quantitative approaches will be used to measure representations of universal truth. On the other hand, if we view the world as ‘fluid and plural’, then qualitative approaches will be used to represent/generate theory within a version of the ‘truth’ that is context-specific and located within a meaning that the event has for individual human beings.

At its most fundamental level, randomised controlled trials (RCTs) are used as a vehicle to establish the existence of a fixed (universal) truth and this is operationalised through the generation and testing of hypothesis. Here, objectivity and distance are the watchwords of theory production. The hypothesis is worked up by the research team to logical deduction (e.g. using the findings of previous research studies), samples of similar characteristics are then recruited into the study and divided (blindly) into experimental and control groups, the measures agreed, bias neutralised and the interventions conducted in a randomised, rigorous and blind manner with the impact then measured (objectively) and compared across and between groups. In this approach, should the interventions in the experimental group be found to have a demonstrably positive impact (in contrast to results from the control group) derived via scores from the agreed measures, then the resulting evidence is seen to symbolise a universal ‘truth’ that is replicable in all similar situations.

Theory generated and tested in this manner is seen as a ‘gold standard’ and actively promoted as the evidence base necessary to advance health care professions and underpin clinical decision-making. Accordingly, the search for high quality RCTs pervades the discourse on reliability in evidence-based health care and is promoted as such through organisations such as the Cochrane Collaboration (www.cochrane.org).

Distilling the production (and, with certain data analysis styles, the testing) of qualitative research down to a few lines is challenging to say the least as comprehensive and weighty handbooks, spread over several volumes, have been dedicated to the subject (Denzin 1994, Denzin 2000 and Denzin 2005). However, it is probably safe to assume that qualitative research approaches, such as phenomenology, grounded theory or ethnography, whilst differing in their epistemology, are held together by a representation of human experience that is grounded in the individual and their subjective interpretation of events and the meaning that it holds. Therefore, producing theory from individual experience that holds as a truism, or a reality, for all representations of human experience makes little conceptual sense. In following a qualitative research approach, fieldwork is (generally) conducted in naturalistic settings and, as Field and Morse (1985) stated in an early and influential text on nursing research through qualitative approaches, with the aim of seeing ‘research as a process that builds theory inductively, over a period of time, step by step’ (p. 11).

Of the existing qualitative approaches, it is grounded theory (Glaser & Strauss 1967) that makes the claim of developing testable qualitative research. Whilst Cutcliffe in this book (Chapter 21) develops this argument and position further, it is important to acknowledge that grounded theory stands alone in developing mid-range (qualitative) theories that seek to explain and predict phenomena under study. Indeed, the central ideas and philosophy delineated in Glaser and Strauss’ (1967) seminal text The Discovery of Grounded Theory: Strategies for Qualitative Research were only ever meant to be a beginning. This was especially evident in the development of ‘theory as process’ (p. 9) and in the strategic method of comparative analysis, which Glaser and Strauss (1967) saw as a general approach ‘just as statistics exist for the experimental methods’ (p. 21), with the addendum that both approaches use the logic of comparison. Grounded theories exist to ‘take hard study of much data’ and Glaser and Strauss (1967 p. 3 slightly abridged) believed that the interrelated role of theory within sociology was to:

As such, Glaser and Strauss (1967) saw the role of theory in sociology as a strategy for handling data in research which provided modes of conceptualisation for describing and explaining. Thus, in conducting grounded theory the researcher attempts to give the data a more general sociological meaning, as well as to account for, and interpret, what has been found.

Glaser and Strauss (1967) believed that, by making theory generation a legitimate enterprise, they would be able to free research from the ‘rigorous rules’ of objective verification, assimilating verification instead into the ongoing process of generating theory. Accordingly, the canons of the deductive approach exist in grounded theory not only as tests of the generalisability of the study, but also as a method for theory modification. Whilst the approach to generating grounded theory may have developed over the years, it is important not to lose sight of this important separation of grounded theory from the mainstream of other qualitative approaches which are described elsewhere in this book.