Activity time: 30 Minutes
Types of media: Webpage, Helpsheet
Andy Hagyard (University of Lincoln)
This guide provides an overview of research methods and methodologies, including practical advice for those new to research, and tips on how to avoid some of the most common pitfalls for new researchers. It covers a background to research; getting started; quantitative, qualitative and mixed methods.
(This resource can be freely repurposed and reused)
This information/resource was last updated in June 2021.
This post was originally added to LearnHigher on: January 14, 2012
About this resource
This guide is aimed at anyone who is keen to develop their understanding of research. This may be because you need to design a research project of your own, or simply to help you understand and critically evaluate research findings that are presented to you.
It aims to demystify the vocabulary surrounding methods and methodologies, and provide practical advice on some of the issues to be considered when designing a research project. In particular it highlights some of the most common mistakes made by new researchers.
Research is at the heart of academic life, one of the defining characteristics of Higher Education which differentiates it from other forms of adult training. Academic staff undertake research in order to create new knowledge and understanding in their subject area, hopefully involving and engaging students in the process and using outcomes to inform curriculum developments.
As a student you will almost certainly have to engage in research at some stage. Typically this will be in the form of an individual project or during the final year of undergraduate study, although universities are increasingly introducing elements of research-based or enquiry-based learning into all levels of study.
Part of the research process involves exposing your findings to peer review and ultimately to public scrutiny. Unfortunately there are large numbers of small-scale research projects whose findings are rejected because their methodology is not appropriate, their methods are flawed or lack rigour, or their conclusions are invalid. Consequently it is essential that anyone embarking on research is able to justify their methodology, identify appropriate methods and comment on the validity of their findings.
Whether consciously or not, any research question will be investigated from a particular standpoint. The way in which we view the world is called a paradigm.
When it comes to doing research, our paradigms are determined in particular by our views on ontology (how do we know what is real) and epistemology (the theory of knowledge – how do we know what we know).
The philosophical arguments around the range of ontological and epistemological perspectives can be extremely complex, and this guide does not aim to go into these in any detail. At the risk of over-simplification, it seeks instead to give a brief overview of the most relevant research paradigms and attempts to demonstrate the implications of these for conducting a small-scale research project.
A positivist standpoint adopts a scientific approach to research. In terms of ontology and epistemology it assumes that the world has an objective reality, that knowledge exists and that it can be observed and measured. A positivist approach generally translates into quantitative research methodology.
One of the cornerstones of the positivist approach is the process of deduction, whereby empirical evidence is collected in order to prove or disprove a hypothesis. Strictly speaking, the research aims to falsify the hypothesis, testing what is known as the null hypothesis. If the null hypothesis is rejected, then the theoretical hypothesis is supported.
For example: you may develop a hypothesis that plagiarism is more prevalent among lower achieving students. A quantitative, positivist research method would collect data to test the null hypothesis (that there is no significant relationship between academic achievement and plagiarism). If the null hypothesis is rejected, then this supports the original hypothesis.
Critics of the positivist approach would claim that it is too mechanistic and inflexible, limited in scope and pseudo-scientific.
Interpretivism supports the view that people and their institutions are fundamentally different from the natural sciences. The study of the social world therefore requires a different approach and seeks an understanding of human behaviour, an empathic understanding of human action.
There is a view that all research is interpretive, that research is guided by the researcher’s set of beliefs and feelings about the world and how it should be understood and studied. Interpretive research methods are prone to criticism because they embrace multiple, individually constructed realities. If reality is individually constructed it implies we are active and implicated in that process. This is in contrast to positivist approaches within which the researcher is independent of reality. In an interpretive paradigm, the researcher is always part of the reality they are attempting to understand.
An interpretivist approach will always adopt an inductive process, meaning that theory is developed from the evidence base. Because of its underpinning belief that reality is personally constructed, it inevitably uses qualitative methods to gain insight into each individual’s experience of a phenomenon.
Critics of the interpretivist approach would argue that it is unscientific and value-laden, open to conjecture and subjective interpretation.
Quantitative research involves the collection and analysis of data in numerical format. A prime aim of quantitative research is to analyse evidence from a sample in order to produce results which can be extended to the whole population. This also allows direct comparisons to be made Consequently, a quantitative study needs to pay particular attention to issues of reliability and validity, to ensure that these claims stand up to scrutiny.
These issues of generalisability and comparability make quantitative research methods particularly attractive to managers.
Findings such as:
Our revised induction programme resulted in a 15% improvement in student retention
Students at our university have an overall satisfaction rating of 3.9, compared to the national average of 3.8
are lent an air of credibility and authority by virtue of their mathematical and scientific nature. Unfortunately this appearance of accuracy can be quite illusory, particularly when investigating a social phenomenon such as learning, and researchers should be cautious when publishing statistical analysis, ensuring that they are able to defend the robustness of their methodology. The following are particular aspects of quantitative methods where care should be exercised.
As stated above, most quantitative research involves investigating a sample and then extending the findings to the whole population. For the results to be generalisable, the sample needs to be of a certain size, but also needs to be representative of the population.
For example: a final lecture on a module was attended by 40% of the registered students who were asked to complete a questionnaire. However, it is reasonable to assume that the characteristics of the 60% who did not attend are significantly different from those who were present. What does this mean for the validity of the questionnaire results?
When samples are genuinely random they can be surprisingly small: national opinion polls typically use a sample of around 1000, yet are able to extend their findings to a population of millions with considerable accuracy. Conversely, large samples may not necessarily tell us anything about the population if they are not representative. This is particularly true when the sample is self-selecting, for example, with internet or telephone polls. As it becomes easier and cheaper to conduct surveys electronically, this should be a major consideration in the interpretation of results. Ultimately the extent to which a sample is representative of the population may be a subjective judgement, or it may be possible to assess it by other research means, but it is a factor that any quantitative researcher needs to take seriously.
Inference and significance
Much student research will inevitably involve small sample sizes and it is therefore important to be careful about drawing conclusions which may not be statistically significant.
For example: you may know that average student marks for their second assignment are approximately the same as their first one. However, out of 20 students who attended an essay-writing workshop, 14 achieved a higher mark in their second assignment. Does this prove the effectiveness of the workshop?
Remember the principle of the null hypothesis – statistically you need to investigate the probability that there is NO correlation and that this observed difference is purely due to random factors.
Suppose that you toss a coin 20 times – how many times would you expect it to land on ‘heads’? On average, you would clearly expect 10 heads and 10 tails, but it is equally intuitive to realise that there will be a range of results. Getting 14 heads and 6 tails is within the normal range of outcomes, and does not lead you to believe that the coin is biased. Normally a result is considered to be statistically significant if there is less than a 5% probability of it occurring by chance. In the above example this requires a ratio of 15:5.
Therefore this case may suggest that attendance at the workshop leads to improved performance, but it is not proven to any acceptable level of statistical significance without further evidence.
A common mistake is to conclude that there is a causal link between variables when the quantitative data only demonstrates correlation. Taking the above example again, even if the improvement in performance is statistically significant, it cannot be claimed that this is because of the workshop. If the students have self-selected to attend the workshop then it is quite plausible that they are more motivated to improve, and that this motivation is the key factor.
Research has shown that overseas students who go to the cinema more than twice a month have significantly better English language skills than those who go less often. This may mean that cinema-going improves their English, but it may also be the case that these students go to the cinema more because their English is better. The relationship between cause and effect can only be proven if there is a clear time sequence. In this case we need to be sure that the improvement in language skills came after the visits to the cinema.
Another error is to assume a causal relationship when in fact there is a third factor which correlates with each of the other two. Research evidence can demonstrate that children who learn a musical instrument perform better at school than those who don’t. However it is quite possible that each of these variables correlates with a third one, that of socio-economic status. As a result it would be invalid to conclude that taking up a musical instrument will directly lead to higher academic achievement
If you are aware of possible external factors, then they can be accounted for by using control groups. These are groups that have the same characteristics as the research sample, but who are purposefully prevented from experiencing the same phenomenon. However, in an educational context there are ethical considerations around depriving groups of students access to a particular service or programme.
Alternatively, qualitative evidence may be gathered separately in order to gain some insight into the causes for the correlation. See the section on Mixed Methods.
The main aim of qualitative research is to discover how research subjects, or participants, feel about their lived experiences. The research usually begins with gut instinct, or with a hunch, rather than with a hypothesis to be tested. Research questions aim to reveal broad social perceptions and are therefore broad and exploratory in nature. Emphasis is given to how participants express themselves in their discourse, with particular attention being paid to the use of metaphor and imagery.
Because qualitative research seeks an in-depth understanding of a phenomenon from a participant’s point of view the sample will be significantly smaller than can be accommodated in quantitative research methods. However, participants are drawn from the group under study and their views are therefore contextualised. Because the data gathered in this way is subjective the results are not necessarily extendable to the whole population but can provide an insight into how members of the group under study might feel about the phenomenon in question.
Qualitative research demands a high level of trust between participant and researcher. It is also a reflexive process in which the researcher needs to reflect on their own role in the research and also on how their choice of research method has impacted on the results, and awareness of the role of the researcher in interpreting meaning. Reflexivity allows the researcher to justify and defend their stance and should provide the rigour for the work. Qualitative research recognises that there is no single truth that can provide an explanation, rather the interest lies in understanding and making sense of the sometimes partial and always subjective view of the participant about a particular phenomenon at a particular point in time.
Qualitative research methods range from interviews to participant observation and can include collection of written data including email, and both researcher diaries and participant diaries.
Interviews are probably the most widely used method for data gathering. Open ended interviews are a useful method of providing an in depth insight into a participant’s feelings or experiences.
Interviews provide the opportunity for conversation to develop and for the participant to express their point of view. But, interviews are also problematic for the researcher in terms of ensuring that the participant says what they mean and means what they say, and that the researcher reports the views of the participant without bias. The researcher needs to analyse the discourse of the interview to develop an understanding of the participant’s point of view through their use of language and through their use of non verbal communication.
It is through analysis and interpretation of the data collected in interviews that meaning is derived. Interviews can also be an iterative process, where the researcher and participant can potentially meet several times to develop themes that emerge through analysis. The in-depth nature of interviews provides their strength but also provides the researcher with significant ethical issues, including responsibility for the psychological and physiological safety of the participant and the guarantee of confidentiality. Interviews may be thought of as intrusive as participants need to make themselves available and will be expected to disclose their views and feelings.
Observation and participation
Observation usually requires the long term presence of the researcher and therefore requires immersion in the field. An example might be that to understand the experiences of a cohort of students on a particular course, the researcher would themself enrol and participate in the course. In this case, it is essential that the analysis is reflexive and that the process is transparently documented in order to avoid the criticism of ‘going native’ or simply providing a descriptive account. Observation and participation provide the opportunity for in-depth investigation of a particular setting or situation and can offer what is termed thick description of what is really going on.
Written data can be described as data that is found rather than derived through the researchers intervention and can therefore be considered as less intrusive. It may, for example, be that a participant feels uncomfortable in the presence of the researcher or for practical reasons, like distance, may physically not be able to meet with the researcher. In these cases a participant diary is a useful source of data. The researcher’s own personal research diary is also a useful reflexive instrument. Data can also be gathered in other creative ways, for example, through drawing and through model building.
Mixed methods is the term used when different data sources and research methods are combined to provide a more accurate account of the phenomenon under investigation. While some care should be taken to ensure that the methods combined do not stem from incompatible research paradigms, at a practical level there are obvious benefits to be derived from blending the strengths of alternative methods to produce a fuller picture.
Firstly, a combination of methods can greatly enhance the validity of research findings (provided of course that they reach similar conclusions!). Perhaps of greater interest, however, is the fact that quantitative and qualitative studies provide answers to different questions. The former give information about ‘what’ and ‘how much’, while the latter answer questions about ‘why’ and ‘how’.
For example: recent research has shown that children are now two to three years behind where they were 15 years ago in terms of their cognitive and conceptual development, yet it offers nothing more than speculation on the possible reasons for the decline. Here is a case where a quantitative study leaves questions unanswered, inviting further qualitative investigation to provide real insight into the factors behind children’s intellectual development.
Broadly speaking there are three ways in which methods can be combined:
- Although the term is often used synonymously with ‘mixed methods’, triangulation refers strictly speaking to the simultaneous collection of qualitative and quantitative data, so that the findings can be compared to see if they validate each other.
- An exploratory design collects qualitative data and uses the findings to inform a quantitative enquiry. This data is then used to validate or extend the qualitative findings.
- In an explanatory design, a quantitative enquiry may produce findings which are then investigated using qualitative methods. This data therefore provides insight into the causes of the relationships identified in the quantitative study. (Fraenkel and Wallen, 2003)
Research is a fundamental part of university life and students are increasingly provided with opportunities to engage in small-scale research projects. However, it is important that research is conducted in a manner which is methodologically and ethically sound, so that research outputs can stand the test of peer-review and public scrutiny. Researchers must be able to justify the rationale for the methodology they have used and not simply use the method which is most convenient.
If using quantitative methods, it is essential that the sample is both large enough and representative of the population and that conclusions drawn from the data are valid and statistically significant. Qualitative methods, on the other hand, can give powerful insight into human behaviour but do not claim to produce results that can be extended to the rest of the population. Consequently, a mixed methods approach can combine the strengths of different approaches and enhance the validity or research findings.
Bryman, A. (2004) Social research methods. Oxford: Oxford University Press.
Crace, J. (2006) ‘Children are less able than they used to be’ The Guardian, 24 January. ONLINE: https://education.guardian.co.uk/egweekly/story/0,,1692952,00.html
Denzin, N.K. and Lincoln, Y.S. (1994) Handbook of qualitative research. London: Sage.
Fraenkel, J.R and Wallen, N.E. (2003) How to design and evaluate research in education (5th ed.) New York: Mcgraw-Hill.
Gordon, W. (1999) Goodthinking: a guide to qualitative research. UK: Admap. Holliday, A. (2002) Doing and writing qualitative research. London Sage.
Knight, P. (2002) Small-scale research. London: Sage.
Opie, C. (2004) Doing educational research. London: Sage.
Tight, M. (2003) Researching higher education. Maidenhead: SRHE.