Identifying and formalizing the research question are where many problems with research occur. Without a well-defined and specific research question or hypothesis, even if the research is carried out to the end, it is unlikely to provide much information. A tightly focused research question or hypothesis dictates what data are collected, how they are collected and analysed as well as providing a context for interpreting the results.
The aim of this post is to discuss the process of developing a research question from the initial idea through to the final research question, which is defined in working terms, is unambiguous, measurable and worthy of the research effort.
The research question – Stages in its development
Most research originates from a problem or question identified from the literature or arising out of the experiences of the researcher relating to some general problem. For example, it might have been observed that non-attendance rates at an outpatient clinic seem to be very high for a particular group of patients and the purpose of the research would be to explore the reasons why. Alternatively, the researcher may want to consider whether the implementation of a programme, e.g. to facilitate a patient-centred approach to the management of a specific chronic illness, has beneficial outcomes in terms of patient’s well-being compared to no programme, or which rehabilitation programme has better outcomes for patients than another following a myocardial infarction (MI).
Reviewing the literature
The first essential stage in developing the research question should be reviewing the literature. By reviewing the literature at an early stage, existing research associated with the problem can be identified and will help in drafting the raw research question.
Having decided on the raw, or rough, question to be addressed by the research, two questions will need to be answered to refine the research question(s). First, what kind of information needs to be collected to answer the question? Second, how is this information collected? In other words, the design of the study or research needs to be decided and the right methodology to fit the type of question needs to be selected. A comprehensive review of the literature helps to answer these questions by identifying earlier research and the methodologies used.
An effective way of identifying relevant literature is through the use of electronic sources. Relevant literature can be searched over specified years, using keywords, by study area, and by author.
|British Medical Journal||www.bmj.com|
Medline and EMBASE provide coverage of the literature in many health-care areas, but do not record all publications from all medical journals. While Medline has a wide coverage of English language journals, EMBASE provides an increased coverage of articles in other European languages.
Scanning the reference lists of retrieved articles, and the bibliographies of systematic and non-systematic review articles, can find additional literature not recovered by the search. Whenever possible, key journals in the field should be searched by hand to find articles which may have been missed or overlooked in the various database searches. Hand searching is also useful in identifying recent articles which have not yet been cited or indexed electronically (NHS Centre for Reviews and Dissemination, 1996).
Published literature can provide a rich source of research ideas through either extending or refining earlier research. Other sources of ideas include:
- The Research Findings Register provides a summary of research projects funded by the NHS, including information such as the research question, methods and findings. The research projects listed are fairly current, which is sometimes not the case with journal articles.
- The National Research Register is a database of ongoing and recently completed research projects funded by, or of interest to, the UK’s health service.
- The NHS Centre For Reviews and Dissemination (www.york.ac.uk/inst/crd/ holds abstracts of quality-assessed systematic reviews, economic evaluations of health-care interventions and publications and projects by a variety of health-care technology assessment agencies
Conference proceedings can be a useful source of information of research in progress and completed. They are generally an unreliable source of data, so where possible, reports from the author(s) should be obtained before any reference is made to the stud.
Research question – The different types
Why the question is being asked, and the type of information to be collected to answer it, determines the type of research design to be used. For example, differences in the experience of mental illness across different cultures or ethnic groups could be investigated using standardized survey techniques and quantitative analysis (i.e. fixed design) to describe differences in rates of illness such as incidence and prevalence. On the other hand, by taking a qualitative approach (flexible design) the researcher would gain insights into, and an understanding of, the factors and experiences from the analysis of the narrative, obtained from in-depth interviews, which underlie the identified differences (Nazroo and O’Connor, 2002).
Research question or hypothesis?
Choosing whether the study will ask a research question or test a hypothesis is, of course, dependent on the purpose of the study. If a prediction is to be tested, which is related to some underlying theory, a hypothesis for the study needs to be developed. However, if a description of what exists in the study population is needed, or an examination of relationships between variables and/or factors, or examination of the causes and effects of a new treatment, the relevant research question needs to be asked. How this question is asked is dependent on whether the researcher is undertaking a quantitative or qualitative research study.
Operationalizing the research question involves first, identifying the concepts referred to in the research question and second, converting these concepts into operational definitions and expressing them as measurable indicators. In other words, having identified the ‘concepts’ in the research question, the researcher must decide how these concepts will be measured. Determining what data need to be collected and the best approach to collecting the information, will then follow on. A concept has been described as:
…a label we put on a phenomenon that enables us to link separate observations and to make generalizations (e.g. depression, quality of life, disability). A convenience, a name we give to observations and events.
Example 1: Which treatment is better for treating psychiatric disorders?
In its present format there is no way to test the question – which treatments and psychiatric disorders are of interest? However, following a review of the literature it may be possible to move on to a more structured question such as:
Is psychotherapy more beneficial than psychiatric medication for people with a psychiatric disorder?
This question is testable, but in its present form is too general. There are three key concepts: psychotherapy, psychiatric disorder and psychiatric medication . To operationally define these, it needs to be clear about what is meant by psychiatric disorder (i.e. acute or chronic; mild or severe), what type of psychotherapy, what type and dosage of medication will be given, over what period of time, and how would the outcome be measured? A more specific question could be:
Will there be a statistically significant improvement in mildly depressed women as assessed on the Beck Depression Inventory (BDI) after 6 months of cognitive therapy compared to fluoxetine (Prozac)?
This question operationally defines both the type and degree of illness, against an objective measure or measurable indicator – the BDI (Beck et al, 1961), as well as the time period and treatment. In addition, by stating a statistically significant improvement, it can be determined whether any difference found will not be by chance, and not just random improvement.
Example 2: Does rehabilitation improve quality of life of patients post-MI?
Again, the question is testable but needs to be more precise. First, a definition is needed of what kind of rehabilitation is provided. Also, there is no mention of a control group. There may be an improvement, but it may also be found that quality of life improves in patients not attending a rehabilitation programme, so a control group is needed to compare and measure the effectiveness of any changes. What does the rehabilitation programme include? How long after the completion of the rehabilitation programme is it of interest to find improvements in quality of life? What does quality of life mean?
Quality of life is an ill-defined concept and can cover a number of non-health as well as health-related dimensions including psychological well-being, pain, social activity, role and physical functioning. Many so-called quality of life measures are, in fact, measures of health status. As a result, the researcher needs to be specific in predicting which areas to expect changes to occur, e.g. social and role activity, and psychological well-being.
The population of interest also needs to be clarified – are we thinking about all patients who have suffered MI? Anxiety levels arising from participating in a rehabilitation programme might, such as, be higher for patients with a second MI compared to first MI, therefore possibly reducing any beneficial effect on psychological well-being What if patients with diabetes were to be included? Will they be more concerned over increased risks of acute complications such as increased hypoglycaemic episodes? Age may also be a key factor of outcome. The length of the programme needs to be defined. Too short, and any potential improvement may be missed. But, what is meant by improvement? What levels of improvement are being looked for and compared, to justify the programme’s effectiveness?
Finally, when developing the research question or hypothesis, the researcher will need to consider whether the study is possible or not, e.g. he/she may have to make a trade-off between rigour and practicality. Will the researcher have the resources to complete the research? Can the required cooperation be obtained from other staff and patients to carry out the project in the specified time? Is the length of the project reasonable? Can the project be completed in the time available? What is the possible
impact on day-to-day work and colleagues? Is the research relevant?
If you would like to know more about developing the research question email us to get your free copy of the full length version of this paper.
Anastas JW, MacDonald ML (1994) Research Design for Social Work and the Human Sciences. Lexington, New York
Argyrous G (2000) Statistics for Social and Health Research. SAGE Publications, London
Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961) Inventory for measuring depression. Arch Gen Psych. 4: 561-571
Cooper G (2001) Conceptualising social life. In: Gilbert G, ed. Researching Social Life. SAGE, London
Fink A (1995) How to Ask Survey Questions. SAGE Publications Inc, Thousand Oaks, CA; London
Guildord J P, Frucher B 1(1978) Fundamental Statistics in Psychology and Education. 6th edn. McGraw-HillInc, London
Meadows KA (2003) So you want to do research? An overview of the research process. Br J Comm Nurs 8(8): 369–75
Miles M B, Huberman (1994) Qualitative Data Analysis: An Expanded Source Book. 2nd edn. SAGEPublications Inc, Thousand Oaks, CA; London
Nazroo J, O’Connor W (2002) Introduction. In: O’Connor, Nazroo J eds. Ethnic Differences in the Context and Experience of Psychiatric Illness: A Qualitative Study. The Stationary Office, London: 9-16
NHS Centre for Reviews and Dissemination (2001) Undertaking Systematic Reviews of Research on Effectiveness. CRD Guidelines for the carrying out or commissioning reviews. CRD Report No 4. 2nd edn. University of York, York
Pope C, Mays N (1995) Qualitative research: reaching the parts other methods cannot reach: an introduction to qualitative methods in health and health services research. BMJ 311(6996): 42–5
Robson C (2002) Real World Research: A Resource for Social Scientists and Practitioners. 2nd edn.Blackwell, Oxford
Siegel S, CastellanNJ (1988) Nonparametric Statistics for the Behavioural Sciences. McGraw-Hill Inc, London
If you would like to read the full and free article on ‘Developing the research question@ why not contact us at: firstname.lastname@example.org