Designing social research: a guide for the bewildered
2011, Ian Greener
1. Introduction to social research design – or what are you talking about? (p. 1 – 14)
There are several terms that you need to understand with regards to social research.
Primary and secondary data
If you aim to collect your own data, you are conducting a primary research collecting primary data. If
you aim to use existing data or library resources, you are conducting a secondary research using
secondary data. The distinction between primary and secondary research and data is good as a
starting point and as a way of labelling research in an outline way, but there are other (more
Quantitative, qualitative and mixed-methods research and data
Quantitative research is primarily concerned with techniques that analyse numbers (calculating
averages, probabilities or exploring numerical relationships), while qualitative data is non-numeric
data (words or images). Quantitative research is powerful when the consumers of research want a
specific answer, while qualitative research is more powerful when the consumers of research want a
more elaborate answer. When these types of research are used simultaneously, it is a mixed-
methods research which can be a way of overcoming the problems associated with each of the
approaches. However, different methods often lead to the production of different kinds of
knowledge and having more data is not always better.
Induction and deduction
Inductive research is that which tries to work from data (usually primary data) to build theory
data-driven (usually qualitative research). Deductive research is that which tests theory through the
use of (usually) quantitative data driven by theory (usually quantitative research). However, it is
possible to do deductive qualitative research and inductive quantitative research. Inductive research
tries to start without clearly defined hypotheses or propositions.
Epistemology, methodology (and method) and ontology
Epistemology is the theory of knowledge: ‘what is knowledge?’ ‘what counts as good knowledge?’.
Epistemology examines what counts as truth, what we can say we know and how we know what we
know. Different methods and different research strategies favour different kinds of knowledge.
Experiments might well work in some circumstances, but not in others: we need to be aware of their
limits and problems before we go ahead with them. Methods are the tools and techniques that are used in social research practice. Methodology studies
methods; when we are examining methods, comparing them or thinking about the kinds of
knowledge they produce, we are doing methodology. The choice of method has to be related to the
particular problem, the research question and the method implicates the kind of knowledge that the
research is able to produce later on. Ontology is the theory of being. When you are considering questions of ontology you are thinking
about issues such as whether the world exists independently of your perceptions of it. Particular
methods and the way they are used in particular research programmes carry with them a view of the
world that has consequences for what we can claim from the research.
Having a good research question increases the chance that you will know where to start with the
Questions or hypotheses in quantitative methods
In quantitative designs, researchers use questions and hypotheses to provide sharp focus to the
purpose of the study. Quantitative study is (usually) testing a theory or hypothesis, so the research
question they are attempting to answer should follow directly from that aim. Quantitative research
questions tend to be concerned with the variables that the research wants to know about. Quantitative hypotheses are predictions the research makes about the expected relationships among
the variables that are included in the study. Research hypotheses are typically either ‘null’ (no
relationship exists) or ‘alternative’ (there is a relationship). An alternative hypothesis can be
directional (‘one way street’) or non-directional where it is not sure which direction the relationship
is going in.
Qualitative research questions
Withing qualitative work, research questions are likely to be expressed in a more open and less
precise way. A good starting point for putting together questions in qualitative research is to ask the
biggest question you can (a central question) and then start narrowing it down/making the general
terms more specific (asking sub-questions). Beginning a question with ‘what’ can be answered using qualitative work but can also fit with
quantitative designs. Beginning a question with ‘how’ are generally more amenable to qualitative
research, except when you add ‘many’. ‘Why’ questions can be useful in both quantitative as
qualitative work. Experiments are a method that involves putting forward a hypothesis and null hypothesis of what the
experiment is designed to test. Then, a test of that hypothesis is designed to control for the variables
within it. Once this test has been repeated sufficient times, the hypothesis is tested and a conclusion
can be drawn from the experiment that either confirms the hypothesis or the null hypothesis. Experiments have three stages.
1. Working out what you are going to test + forming a hypothesis and null hypothesis.
2. Designing the specific test of the hypotheses. (deductive)
3. Carrying out the experiment.
Finally, the results needs to be interpreted. (inductive) Experiments have two specific features that are important.
1. The idea that the external reality that we are testing can be controlled so as to allow
individual elements within it to be investigated (control the variables).
2. Their treatment of causation (constant conjunction).
3. Surveys and questionnaires – or how can I conduct research with people at a
There is a difference between surveys and questionnaires. A survey is a research design that takes a
It is based on a sample, to have as large a sample as is necessary to capture all of the
variation in the population.
It occurs at a single point in time.
It is predominantly quantitative.
The aim is to seek patterns within that quantitative data.
A questionnaire is a type of survey asking subjects to respond to a range of questions, often in a self-
They are relatively simple top put together.
They are cheap: researchers do not gather data personally (less time spend).
They are versatile: they can be used in a variety of ways and they can produce data to test
hypotheses as well as gather qualitative responses.
They are easy to get wrong: different subjects respond to/interpret words differently and
you do not get to get a clarification if they understand incorrectly.
Different perspectives on social research might make differences in a questionnaire design.
Hypothetico-deductive perspective: theory generation through hypotheses testing –
deductive methods – quantitative approach – closed-response questions.
Realist approach: find out about the external world; find underlying, generative causes of
behaviour – mixed-methods – open and closed-response questions.
Idealist approach: to understand our subjectivities and interpretations of the world –
qualitative approach – open-response questions.
The purpose of a questionnaire in your research design will considerably affect its design and use.
E.g. learn about something you do not know much about or testing hypotheses, generated
from existing research or your own ideas.
The aim of a questionnaire is likely to shape a range of design issues withing it, the extent to which: it
utilises open and closed responses, sampling is a significant issues and the type of questions we ask
within it. Open-ended questions ask respondents to give a response in their own terms. Reason to use this
kind of questions is if it is impractical to give all of the possible options and to avoid imposing the
researcher’s ideas and concepts upon the respondent. But: risk that respondent does not answer
everything + takes time to analyse. Closes-ended questions ask respondents to give their answers according to possibilities that the
researcher has predefined. Therefore the researcher needs to have a good idea of the likely
responses to the questions. Reason to use this is to reduce ambiguity, they are easier to interpret,
they can be answered quickly and the data is easy to analyse. But: fail if the response options are
inadequate + misunderstandings are difficult/impossible to pick up.
Designing social research: a guide for the bewildered