Conducting Research Projects Guide: Research design

The purpose of action research is to investigate an issue that will then result in some kind of action in order to improve the quality of a programme, service or to address the need for social change.

The three parts of action research are:

The methodology used for action research can be quantitative, qualitative or mixed-methods.

The result of action research is that some kind of change is enacted at the end of the project that changes the status quo.

(Kumar, 2014).

Cohort design is based on choosing a particular sub-group of the population who have something in common, e.g. same year of birth, marital status, or other factor.  

(Kumar, 2014).

Causal design is used to determine the impact that a distinct change may have on an existing situation or assumption. By finding causal links between variables, researchers can determine why certain outcomes occur, studies can be replicated and internal validity is higher. 

There are three conditions that are essential in determining causation:

1.  empirical association - an association found between an independent and dependent variable leads to a valid conclusion.

2.  appropriate time order - participants of the study must be subjected to a variation in the independent variable before the dependent variable.

3.  non-spuriousness - the connection between two variables is not caused by a variation in a third variable.

Causal design is not always useful because causality can be difficult to identify and measure.  There may be a level of co-incidence, so researchers must be careful to weed out those variables that muddy the waters.  Causality cannot always be proven.

(University of Southern California, 2025).

Descriptive design is used when the researcher wishes to report on the what or how of what they wish to study.  As the title suggests, it is a straight description.  It does not look for connections or causes.

(Kumar, 2014).

Exploratory design can be used in two circumstances:

•  when the researcher seeks to explore an area or issue that has had little research conducted on it.  
• when the researcher tests if a study or an area or issue is practical and reasonable.

(Kumar, 2014).

Longitudinal studies are so named because they continue over a long period of time.  The researcher will visit or have contact with the participants several times over the period of the study.  The same information will be collected each time and the results compared to note trends over time.

An example of a longitudinal study is the Dunedin Study where researchers from Otago University have followed new born babies born in 1972-1973 across the years to track their progress through life. To see more about this study: https://dunedinstudy.otago.ac.nz/

(Kumar, 2014).

Meta-analysis uses quantitative methodology to summarise and analyse a variety of research reports that explore the entire population rather than a subset of the population.

(Vogt et al., 2012).

Observational design is used to observe a situation, interaction or communication as it is occurring. This gives data in real time. Observational design is purposeful as participants are chosen based on who will produce/provide the best data for the study.  It is also systematic in the way the observation is conducted (see the sampling variables in the research tools tab). 

(Kumar, 2014)

A case study is usually conducted using a particular individual, group or organisation who are studied in great detail.   

Cases are chosen because they can provide a window into the particular issue being researched.

Case studies provide the researcher the opportunity to gain a deeper understanding of the topic due to the small sample group.  However, this means that results cannot be generalised across the wider population.

It is essential for the researcher to build a positive rapport with the subject(s) of the study, particularly if this is a community group. Empathy, open-mindedness and being non-judgmental are essential qualities of the researcher undertaking case studies.

(Kumar, 2014)

Cross-sectional studies aim to obtain a snapshot of the incidence and frequency of what is being studied by sampling a cross-section of the population involved.  It is most useful for social science research but only gives a picture of a specific moment in time.

The design process:

One limitation of cross-sectional design is that it cannot measure change because there is only one contact event with the participants.

(Kumar, 2014).

Experimental design is used when the researcher knows the cause of some change and observes the outcomes of this change.

A sample of types of experimental design:

• the after-only experimental design - an intervention has been implemented and the researcher seeks to ascertain the results of the intervention.
• the before-and-after experimental design - The intervention is set up and then the results of the intervention are evaluated.
• the control group design - a control group and an experimental group are set up to study the effects of the intervention on the experimental group compared to the control group who do not participate in the intervention.
• the double-control design - this design has two control groups to overcome any effects that may be due to the research instrument.  
• the comparative design - this can be experimental or non-experimental. The study population is separated out into groups and the results of the intervention compared. 
• the matched control experimental design - used mostly in the testing of new drugs.  Individuals are matched according to a set of criteria.  One is put in the control group and the other into the experimental group.  The outcome is then compared.
• the placebo design - the participants are divided into two or three groups. One group receives the drug, one group the placebo and one group has no intervention.  The results of the three groups are then compared.

(Kumar, 2014).

Field research studies individuals or groups in their own environment.  It is generally based on observations and interviews. The researcher takes notes of what they observe while in the environment.  These observations are then interpreted and conclusions developed from themes that become apparent during the observation period.

This design is helpful in:

1. filling in gaps in existing research based on the local situation.

2. putting the research problem in a specific context thereby assisting researchers to more fully understand the problem and ultimately develop a solution.

3. corroborating or refuting previous research into the problem.

4. asking questions that are specific to the context of the research setting.

5,  understanding the unique perspective of the local situation that may differ from that of the general understanding in the literature.

Field research design:

1. takes a lot of time and resources because it includes several distinct steps.

2.  commitment to the people and situation so that the fullest picture can be obtained.

3.  requires flexibility due to a lack of control of the situation being studied.

4. can be difficult to interpret without other documentation to verify what was observed.

5. may not be able to be generalised to other situations due to the specific nature of the environment being studied.

6. requires researchers to acknowledge how their presence may have affected the behaviour of participants.

(University of Southern California, 2025). 

Mixed-methods design uses both qualitative and quantitative methods in order to gain a rich variety of data.  This design is used when parts of the study are more suited to one method than the other.  By using both methods, the quality of data collected can be improved.

(Kumar, 2014).

A systematic review is an analysis and evaluation of the existing research findings on the topic being studied.  It seeks to gain a body of evidence to form the basis of further research, policy making or process formulation.

(Holly et al, 2011).