Welcome to these online modules about research methods. Here you will learn about choosing a good research question, establishing the aims for your PhD, the different types of research methodologies, and detailed modules on statistical packages, study design and epidemiology.

Some of you may have forgotten the mathematics you were taught at school, so we even have a module linking you to basic arithmetic. At a minimum, you should understand the square root function, the natural logarithm and its inverse, the exponential function. Apart from that, we have tried to keep the maths to a minimum!

We hope you enjoy these modules and that they are helpful to you over your PhD journey.     

The beginning of the PhD journey, the end, and all that is in between depends to a large extent on your research question. Get it wrong and you are in for three or more years of pain.  Get it right, and the journey becomes immeasurably easier.  So how do you choose a suitable research question?

Firstly, the research question has to lead to a doable PhD. For example, take the following research question: “Does taking once daily probiotics protect against Parkinson’s Disease?” There is nothing wrong per se with the question itself, assuming it is biologically plausible.  It is a perfectly valid and laudable research question.  It is just that Parkinson’s might take 40 years to develop after exposure -  clearly not a feasible time frame for a PhD.  Here is another example: “Do women with only one breast after mastectomy have greater trouble breastfeeding than women with both breasts intact?” Since breast cancer is more often than not a disease of older age, you would have a great deal of trouble recruiting enough younger women with a mastectomy still breastfeeding.

Secondly, the research question cannot be too broad. For example, “Why are men reluctant to express their emotions?” Another perfectly valid research question, but you are likely to have grey hair and your supervisors retired before you complete your PhD!

Thirdly, your research question should not be too tight or closed. For example, “What is the prevalence of refugees from Yemen in South Australia?” – not nearly enough there for a PhD.

Finally, your research question has to generate something new - you cannot simply replicate what someone else has done; unless you can demonstrate that their work was flawed.

Where do you find a good research question? If you come from a clinical background, then you are in a good position to know what the major clinical problems are, and which aspects of your clinical practice lack evidence (usually most of it!). If you are from a non-clinical discipline, then have a look at recent editions of publications in your field. Most published studies end up with a conclusion something like “Clearly, more research is required in this area.” If you are joining a research centre with grant funding, then your research question might be given to you as part of the grant. Finally, a good supervisor should be able to sit down with you and help you develop your question.

Let us make up a research question. How about “Should AFL players wear protective helmets to prevent concussion during games?”  The next step is to set up a series of aims. Do not choose more than five or six aims, and each aim should be directly related to the research question.  You can use a table similar to the one below (Table 1). The first thing to note is that each aim begins with the word “To”. Secondly, each of the five aims directly addresses an aspect of the research question.

Table 1: The research question and aims

For each aim, we now add a suitable method to address the aim in a right hand column (Table 2).

Table 2: The research question, aims and methods

Is there enough depth in the above plan to make a suitable PhD? Think of an acceptable PhD thesis as consisting of four or five published papers bookended with a front Introductory chapter, and a back Discussion chapter. Then it looks like there is ample scope in the above table to end up with a good thesis. The above study is clearly doable.

Note that the methods and subsequent data collection are driven by the research question and aims, NOT the other way round.

The first step in any PhD is to find, read and critique the literature, in other words, a literature review.  The body of work reviewed might include published scientific papers, books, government or institution reports (called grey literature), or other theses. Consider using bibliographic software such as Endnote for storing information about each reference. There are in fact several type of literature review methods available.

Cochrane review

According to the Cochrane Organization, “Cochrane Reviews are systematic reviews of primary research in human health care and health policy, and are internationally recognised as the highest standard in evidence-based health care. They investigate the effects of interventions for prevention, treatment and rehabilitation.  Every study included in the review is carefully assessed for bias, and given a quality rating score. Cochrane reviews follow a very specific protocol, and the one you develop has to be accepted by the Cochrane Organization before you can start the review:

http://www.cochranelibrary.com/cochrane-database-of-systematic-reviews/

Cochrane reviews usually involve the review of randomised controlled trials, but may also include other types of controlled trials. Because of the rigour required, they can take 12 months or more to undertake, often require several researchers to be involved, and are therefore often not suited to the PhD timeframe.

Systematic review  

Standard systematic reviews are still rigorous, but not bound by the rules of the Cochrane Organization. For example, the researcher can choose or even create their own quality rating score for studies. They are again best suited to reviews of controlled trials. The key features of a general systematic review are that explicit and transparent methods are used, and a standard set of stages are undertaken. Students intending to undertake a systematic review should read the PRISMA statement, which provides guidelines for how to report them: http://prisma-statement.org/

A well-conducted systematic review can also take up to 12 months to conduct, so the student should give careful thought as to whether a systematic review is essential.

Scoping review

Scoping reviews are undertaken when feasibility of the research is considered to be a challenge, either because the relevant literature is thought to be vast and diverse and/or it is thought that little literature exists. In the scoping review, the same systematic, rigorous methods used by the systematic review are used to find studies and extract data. Analyses and syntheses are part of the scoping review, but the depth and type of analysis are different. Scoping reviews are ideal for a PhD in that they are still rigorous, but take a lot less time to undertake than systematic reviews. Here is a good reference:

http://ktdrr.org/products/update/v4n1/dijkers_ktupdate_v4n1_12-15.pdf

Integrative review

Integrative reviews are very similar to scoping reviews, and are the broadest type of research review method. They allow for the inclusion of experimental and non-experimental research, qualitative studies, and can also include theoretical papers. They are most commonly used in nursing research. Here is a reference to undertaking an integrative review:

http://www.aornjournal.org/article/S0001-2092(06)62653-7/fulltext

Narrative review

These are the least rigorous of the various types of literature review. The researcher still needs to describe the databases searched and the search terms used, but there is no formal attempt to rate papers on their quality. Instead, the researcher addresses the different studies in a narrative fashion, comparing and contrasting them as required. Because they lack rigour compared to the other types of literature review, it is more difficult to get them published.  None-the-less, they are still very commonly found in theses. Here is a reference to narrative reviews.

http://www.ease.org.uk/sites/default/files/writing-reviews.pdf

What type of literature review should you do?

The choice depends very much on the topic and what literature is out there. Is it mainly controlled trials, or a mixture of quantitative and qualitative research?  Whichever type of literature review you undertake, there are two golden rules.

Firstly, start broad and work towards the specific topic of interest. Think of the review as an upside down triangle. In the example of protective helmets for AFL players, you might start by looking at sports injuries in general, concussion as a result of playing a sport, the use of protective helmets in sports, concussion in AFL players, and finally, protective helmets for AFL players.

Secondly, there is only one main reason for undertaking a literature review as part of a thesis, and that is to justify why the research needs to be done. Make sure the final paragraph of your literature review reflects this.

In Table 2, the method for Aim 1 was a literature review, which we have already discussed. The methods for Aims 2 and 4 can be described as qualitative research, whereas the methods for Aims 3 and 5 are described quantitative research. Because the proposed study contains both qualitative and quantitative components, it could be best described as a mixed methods PhD.  

Qualitative research

In the qualitative approach the type of data to be acquired focuses on experiences, thought, feelings and behaviours. The data themselves are usually words or observations, unlike quantitative research, where the data are numbers. Qualitative Research is primarily exploratory research, used to gain an understanding of underlying reasons, opinions, and motivations. Qualitative research can be either a study in its own right, usually undertaken when little is known about a topic, or as an aid to undertaking quantitative research. For example, qualitative research is often used to help develop a questionnaire.  Qualitative research can also be used as an adjunct to quantitative research. For example, a clinical trial might show that an intervention is not effective, and additional qualitative research might help explain why it did not work. Some common methods of collecting qualitative data include focus groups (group discussions), individual interviews, and participation/observations. The sample size is typically small, and respondents deliberately selected because of their characteristics.

There are many different ways of analysing data obtained from qualitative research. Quite often, a simple thematic analysis is suitable, as long as it is undertaken in a systematic and reproducible fashion. Phenomenology or analysing the lived experience is another method of analysis. There are at least three schools of thought on how to undertake a phenomenological analysis. A second major method of analysing qualitative data is by using ethnography.  This is an approach developed by anthropologists to learn about cultures, and is now used to study behaviour and social interactions. Finally, grounded theory focuses on generating a theory from research data to describe what is happening in a social setting.  The above explanation only touches on many of the different theories used in qualitative research.

Quantitative research 

As its name implies, quantitative research is all about numbers. It is broadly covered by the disciplines of biostatistics and epidemiology.

Biostatistics

Biostatistics (also called medical statistics) is the discipline of statistical theory and methods applied in the health context. Whilst statisticians are usually trained in Schools of Mathematics and Statistics, biostatistics is taught in Schools of Public Health and Epidemiology, however, there is much overlap. Biostatistics focusses on areas such as design and analysis of clinical trials, the analysis of rates of disease, measures of the association between exposures and outcome, survival analysis, diagnostic testing, disease clustering and surveillance, disease screening, reliability and method comparison studies and biological assays. The analysis of genetic studies has formed its own discipline called bioinformatics, which is somewhat different. 

The wide availability of statistical software has made it much easier for students to undertake their own statistical analysis, but this has come with a downside. All statistical tests and procedures are based on assumptions, and it is very easy to either use the wrong statistical test, or break the assumptions required for a statistical test, and end up with nonsense results.

Epidemiology

Epidemiology is the study of the distribution, causation, diagnosis, prevention and treatment of diseases. It is very much interwoven with biostatistics since almost all epidemiological analyses require methods from biostatistics.

Descriptive epidemiology examines how much of a particular disease is in a community, whether it is endemic (always in the community), epidemic (rapidly increasing in the community), or pandemic (going globally). It also includes details of who gets the disease, when it first started and who is immune.  Disease causation (also called aetiology) is most often researched using case-control studies.  For example, case-control studies were used to demonstrate the link between smoking and lung cancer. Diagnostic testing including screening programs, requires an understanding of diagnostic accuracy, and covers indicators such as sensitivity and specificity. Epidemiology also covers the whole area of study design including clinical trials.

Which is better, qualitative or quantitative research?

That is a bit like in carpentry, asking what is better, a hammer or a saw. Each has a place in the toolbox of research. Quantitative research is better for providing the evidence for clinical practice, but addresses the question “Does it work?” rather than “Why does it work?” Qualitative research is ideal when you are investigating a very new area of research where very little is known. You might hate or be afraid of statistics, but qualitative research is equally if not more difficult to undertake – it is absolutely not an easy option.

In many ways, mixed methods is the ideal as it produces researchers trained in both qualitative and quantitative methods – a well-rounded researcher!

Having said that, the choice of a qualitative, quantitative or mixed methods PhD should not be based on personal preference, or the preference of your supervisors. It should be based on Table 2, a careful statement of your research question, aims, and the most appropriate method to address each aim. You must use the right tool for the job.

Now that you have read this very brief introduction to research methods, you will find other modules covering qualitative research, statistical software, biostatistical methods, study design, sample size calculation, and epidemiology.  

We hope that these will help you as you progress your way through your PhD.