Research Methodology and Quantitative Research Designs
Whats is Research Methodology
Research Design and Sampling Methods
The methods are derived from the research design and generally include sample, interventions (if applicable), instruments, data collection procedures, and plans for data analysis.
A research design, according to
Kerlinger (1986), “expresses both the structure of the research problem and the
plan of investigation used to obtain empirical evidence on the relations of the
problem” (p. 279).
Selection of Best Research Design
There is no “best” design. The appropriate design is the one that
fits the theoretical formulation underlying the research questions or
hypotheses. Theory generation often requires qualitative approaches, whereas
relation, association, and theory testing often require quantitative data.
Types of Quantitative Research Designs
Quantitative designs are often divided into experimental, quasi-experimental, or non experimental. We often think of experiments as having been around for a long time, but actually it was only in the 1930s that the first experiments were conducted.
Sir Ronald Fisher's (1935) book The Design of Experiments provided the first details of experimental techniques. The purpose of experimental design was to gain greater control and thus improve validity.
The aim is to associate a treatment with its outcome by minimizing the effect of other variables on the outcome and reducing error introduced by extraneous or confounding variables. Random assignment to groups, manipulation of the independent variable, and control of extraneous variation are the key elements in experimental design.
Originally, experiments were conducted in laboratories; then the social sciences adopted the techniques, and other designs emerged, such as quasi-experimental designs. In quasi-experimental designs there is an experimental intervention, but one or more of the other elements of experimental design are missing.
There may be no random assignment to groups. In such cases, the investigator should address the issue of group equivalence by comparing the groups on relevant variables. There may be no control group, as when a group of subjects is measured over time or under different conditions.
This is usually referred to as a within subject's design. In non experimental designs there is no investigator-controlled intervention. Because the investigator does not control the independent variable, it is more difficult to test the direct effects of one variable on another.
What is usually tested is
the relationship between and among variables. This includes the testing of
models through techniques such as path analysis and structural equation
modelling.
Quantitative Experimental Research Design in Healthcare
One type of experimental design that is of special interest to health care professionals is the randomized clinical trial. In such an experimental design the intervention is tested in practice rather than in a controlled laboratory experiment.
In the United States the first randomized clinical trial was reported in 1951 by Yale researcher Cadman (1994). She has studied the effectiveness of penicillin in treating pneumococcal pneumonia.
Another Yale researcher, Dumas (Dumas, RG, & Leonard, 1963), published the
first report of the use of experimental design in nursing research. In 1963 she
reported on nursing interventions to reduce postoperative vomiting.
Sampling Methods i Quantitative Designs and Types
In all designs, sample selection is crucial. Whether the sample
consists of an N of 1, or of thousands, the sample must represent the
population of interest. Additionally, the size of the sample must be adequate
for subsequent analyses.
Sample designs are often divided into probability and
non-probability designs. Some form of random sampling is used in probability
sampling. This enables the researcher to make use of probability theory to
determine the accuracy of results through the computation of standard errors.
The notion is that all potential subjects have an equal possibility of being
included in the sample.
Non probability sampling includes several techniques, including
selecting subjects based on some criteria (purposive or judgmental), taking
those subjects that are available when the study is conducted (convenience or
accidental), accruing a set number of subjects in various categories (quota),
and advertising for volunteers.
The procedures for implementation of the study and for data
collection are designed to maintain the integrity of the study. In experimental
designs, methods for assignment to groups and implementation of the
experimental conditions must be determined. Careful attention should be paid to
ensuring that there is no contamination of experimental intervention across
study groups.
Internal and External Validity of Data
DT Campbell and Stanley (1963) coined the phrases “internal” and
“external” validity. Internal validity refers to the integrity of the study
through which we can infer the relationships among the variables under study.
External validity refers to how generalizable the results of the study are to
other samples, settings, and so forth.
All variables included in the design must be defined and measured.
Selection of psychometrically sound instruments and establishment of controlled
methods for data collection are necessary for the integrity of the data. Data
analysis is based on the questions being answered, the characteristics of the
data collected, the size of the sample, and the assumptions underlying the
statistical techniques.
Quantitative research methods from design through data collection
and analysis must be carefully explained prior to embarking on a study. Careful
attention to all aspects of the methodology is necessary to produce valid
results.
Give your opinion if have any.