Welcome to SMIRK Unit
8 - Research design


SMILE main image
GCU logoSMILE logo
Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License.

What is research
design?

checklistWe ask questions all the time. Research is a formal way of going about asking questions. It uses methodologies (formal frameworks for collecting and analysing data). After you have decided what your research questions or your hypothesis are, you start designing your research. This means you plan your research. Much about this plan is determined by your research questions or your hypothesis.

What philosophical stance do you take to your research?
For example, do you think the best way to answer your research questions or prove your hypothesis is to set up experiments? That would be a positivist stance. Or do you want to explore how human behaviour impacts on public health? That would be an interpretivist stance. Have a look at the information on educational research the Open Lean web site from the Open University for more information.

How do we choose a research design? 
Methodology is the actual approach on how you go about answering your research questions (proving your hypothesis). A methodology can contain various research methods (tools with which you gather your data: questionnaires, experiments, interviews). Caution! You will find the literature sometimes treats the terms methods and methodology as the same, which is not correct. However, sometimes authors disagree if something is a method or a methodology, in this case there is no right or wrong. Your task is to justify your decision.

Some authors treat a case study as research method (the actual tool to gather data) while others tread it as a methodology (the approach on how to go about researching an issue and use), for example, interviews and focus groups as methods within the methodology ‘Case Study’.

It depends on whether we want to:

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License

Research methods

The most significant difference is made between quantitative and qualitative research methods. Quantitative methods are usually related to numbers, surveys, experiments, statistics and positivist research philosophy. Qualitative methods, interviews and observations, - are usually related to interpretivist research philosophy.

However, your methodology could benefit from including qualitative and quantitative methods. This depends on the aims and objectives for your research project. What do you want to find our or prove? What is the best way to answer your questions or prove your hypothesis? This section of SMIRK gives general information about research methods, to get specific advice relevent to your subject, contact your Learning Development Centre.

Here are the basic research methods and designs:

These are not mutually exclusive – you can use more than one. For example, a cross-sectional and case study, such as semi-structured interviews with households but within a particular neighbourhood. You could also use a case study and longitudinal study, such as in-depth research with a community over months or years. In this section we will also look at measurement and some general issues to consider when planning your project.

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License.

Quantitative research
and surveys

What is survey research?
Many research projects and dissertations need the collection of primary data from individuals. Questionnaires are often a good way to gather such information and views. However, a badly designed questionnaire may get only unusable responses or none at all. It is important to be able to identify and avoid any pitfalls of questionnaires to ensure a successful result.

What do you want to know?
Before you even write the first question, it is important that you have a very clear idea about what you want your questionnaire to achieve. Our section on defining a research question should help you with this. Write down your research goals, and think about what information you need to get from respondents to meet them. Think also about how you are going to analyse each question to get the results you need. Remember there is a difference between things you need to know, and those it would be nice to know. Eliminate unnecessary lines of questioning at the planning stage.

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License.

 

Qualitative research,
interviews, focus
groups and case studies

Face-to-face interviews

Focus groups

Case studies

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License.

Participatory and
experimental research

Participatory research

Experimental research

Classic experimental design

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License.

Experimental design
examples

Two groups

Experiment group Pre-test observation Experiment Post-test observation
Control group Pre-test observation No experiment Post-test observation

Analysis:  Compare Pre-test and post-test results. 

Temperature arrow Change in the state of a liquid
     
Independent variable (x) arrow Dependent variable

Quasi-experimental research
These are similar to experiments but without random assignment to the groups. There is a lack of internal validity – how can we be sure that x caused y if there was some engineering of the groups?

There are two main types:

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License.

Longitudinal

longitudinal data previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License.

Longitudinal and
cross-sectional research

Panel studies: data  from samples at 2 or more points in time, for example,

  • British Household Panel Survey (BHPS) – objective is to further understanding of social and economic change at the individual sample of more than 5,000 households in Britain with the same individuals re-interviewed in successive waves.

Cohort studies: returning to a sample of participants who are part of the same cohort, for example,

  • Millennium Cohort Study (MCS) – follows the lives of a sample of 19,000 babies born between 1 September 2000 and 31 August 2001 in England and Wales, and between 22 November 2000 and 11 January 2002 in Scotland and Northern Ireland.

Cross-sectional research

  • This is the most recognisable research design in social research.
  • Information is collected across more than one case at one point in time.
  • The classic ‘social survey’ design.
  • Do not confuse cross-sectional with a ‘survey’ used to describe a method of collecting data. Cross-sectional can use many other methods to collect data, such as structured observation or official statistics.
  • The aim is to collect information on variables to find out about the relationships between those variables.
  • previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License

Recap

  • There are four main research designs :
    • Experimental
    • Cross-sectional
    • Longitudinal
    • Case study
  • These are not mutually exclusive – you can use more than one.
    • cross-sectional and case study, such as semi-structured interviews with households but within a particular neighbourhood.
    • case study and longitudinal, such as in-depth research with a community over months or years.
  • Now let's go on and look at types of measurement.

If you would like to look at this topic in more detail, look at Strathclyde University's excellent flowchart on research problems.

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License

Measurement

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License

Planning your research

Key questions

Step 1: What?

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License

Key questions

Step 2: How? Where? Who?

Step 3: When?

Our section on writing a research proposal will help you with this and other issues associated with research.

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License

Getting the answer

Step 4: Why?

Good luck!

previous page next page

GCU logoSMILE logo

Creative Commons Licence
SMIRK - Research design by Learning from WOeRk, University of Plymouth modified by Marion Kelt, Glasgow Caledonian University is licensed under a Creative Commons Attribution 4.0 International License