These notes address, at a preliminary level, broad planning principles that apply to many different areas of research. Anyone who has a research degree should be aware of them, whether or not they arise in their own research. They give, also, pointers that may help in getting a clear view of where the researcher’s project is headed. I will have been successful in my endeavour if I kindle in at least some readers interest both in the research process itself and in the examples. There are several reasons why researchers should take an interest in broad-ranging issues in research planning: 1. The immediate research project may take twists and turns that are different from those for which earlier study has been a preparation. This is especially likely for highly applied projects, which typically demand a range of diverse skills. 2. Those who acquire a wide range of research skills are thereby better placed, after graduation, to turn their hand to tasks different from those for which their immediate research training has equipped them. 3. Broad-based research skills will best equip researchers to respond to changing demands, as they move from task to task and from job to job in the course of their careers. 4. Many of the skills are highly relevant to the planning of any substantial understanding. Designing the instrument panel on a large aeroplane may appear like an engineering problem. It has, also, a large human engineering component. A layout that has the potential to confuse pilots may, in an emergency, be fatal1. This is not a text on statistical methodology, even though there is extensive discussion of statistical issues. It discusses, with numerous examples, issues that should influence the design of data collection, the eventual analysis of the resulting data, and the reporting. The emphasis is on the way that statistical issues impact on the quality of the science. There is a strong focus on the critical and questioning role of scientific ways of thinking. It does not much matter where you start practicing scientific thinking. What is important is that you start. As Sagan (1997) notes2: Because its explanatory power is so great, once you get the hang of scientific reasoning you are bound to start applying it everywhere.
Criticism and questioning are in tension with the openness to imaginative insight that is equally important to the research process. Data may be in tension with the theoretical insights that generated their collection. The issue of evidence is central. There must be an assessment of the evidence in the literature that is the starting point for the research project. There must be a research strategy that will bring together data that address the research question. Statistical analysis will extract from the data evidence that relates to the research question. Finally, the new research evidence must be integrated into the body of earlier knowledge, creating a coherent account that will appear as a report or paper or thesis. My examples range widely, from social science through to pure and applied biology and physical science, with medical and health examples strongly represented. Most people are interested in their own health. I am hopeful that such examples will be of wide interest to non-medical as well as medical researchers. I have tried to find examples that are not unduly technical. I have found it helpful, at various points, to draw on ideas from the approach to clinical medicine that has the name “Evidence-based Medicine (EBM)”. For those who want to understand the practicalities of Evidence-Based Medicine, I recommend the book Smart Health Choices, subtitled How to make informed health decisions, by Judy Irwig and collaborators. These ideas may assist researchers both with their health needs and with their research planning! The first drafts of this monograph were written for a course that introduced a series of short courses on statistical design and analysis. Any statistical analysis must have a context. Data collection and data analysis serve the wider aims of the research project. This requires a clear view of the project’s aims. There are principles that should guide the design of data collection whenever this lies in the researcher’s control. Where the researcher does not have this control, it is important to examine the processes that generated the data. Focusing attention back onto the contexts from which data have come is important both for use of the data that the researcher may already have, and for thinking about any future data collection. Data do not just happen!
