D annan a simple guide to resarch writing indd

31 
 

---

 
 
 
 
Figure 3.2: Research philosophy
 
 
RESEARCH PROCESS WITHIN A MIXED METHOD FRAMEWORK
 
The two main paradigms represent two extremes of a continuum; therefore, it is possible for 
the researcher to use a blend of philosophical assumptions and methodologies. Creswell 
(2014) noted following about the mixed method framework:
• It is a – two-phase design in which the researcher conducts entirely separate positivist 
and phenomenological examinations of the problem.
• The dominant/less dominant is where one paradigm dominates while the other is 
used for a small component of the study (e.g. the researcher may use observation 
and in-depth study to generate themes and variables and then use a quantitative, 
hypothesis-testing design for the study as a whole).
The use of a mixed-methodology design allows the researcher to combine aspects of both 
paradigms using the advantages of both when appropriate.

32 
 

---

 
 
 
RESEARCH STRATEGIES
 
A brief description of some of the important research strategies available to students/ 
researchers is provided next. Each strategy differs from others in two distinct areas: the 
nature of the question asked and the method used to answer it.
Positivist research strategies/experimental design
 
The basic experimental designs used in research have been adapted from the physical 
and biological sciences. Experimental studies seek to identify causal relationships.
• The aim is to manipulate the independent variable in order to observe the effect on 
the dependent variable.
Because of the greater difficulty to control experimental conditions in the human sciences, 
experimental methods can only be used under very clearly defined conditions. The typical 
experimental design in for instance in business research involves:
• the selection of a sample of subjects; random assignment of these subjects to 
experimental and control groups;
• the exposure of the experimental group to the independent variable which is withheld 
from the control group;
• finally, the evaluation of the two groups on the dependent variable. There are other 
variations in experimental design 
– repeated measures design, matched-pairs design 
and single-subject design.
In a quasi-experimental design, the researcher does not have as high a degree of control 
over the independent variable. People, for example, are already assigned to groups such as 
social class, type of injury, type of occupation and income group, for which the researcher 
is testing the effects. The experimental designs isolate and control all the factors that could 
be responsible for any effects except the one under examination. The most important use 
of quasi-experimental research designs is where researchers cannot, in good conscience, 
assign people to groups and test the effect of the group’s membership on some other 
outcome. For instance, researchers interested in the effect of retrenchment on staff morale 
could not encourage the process or actively identify those to retrench in any organisation. 
Rather they would seek out companies where retrenchments are occurring and then 
conduct the research.
Quasi-experimental research is also called post-hoc research or after-the-fact research 
means that the actual research takes place after the assignment of groups such as 
employed versus unemployed, male versus female, etc.