Please enable JavaScript.
Coggle requires JavaScript to display documents.
Chapter Five: Research Design, Requires, ensures - Coggle Diagram
Chapter Five: Research Design
Research Design:
comprehensive plan for data collection in an empirical research project
Must specify:
-data collection process
-instrument development process
-sampling process
Data Collection
can be grouped into 2 categories:
Positivist Methods:
lab experiments and survey research; aimed at theory testing; relies heavily on quantitative data
Mixed-Mode Designs:
combine qualitative and quantitative data use for situations where it may help generate unique insight that either type alone could not
Popular Research Designs:
Positivist
Experimental Studies:
intended to test cause-effect in a controlled setting; can be in a lab or in the field
true experimental design:
subjects must be randomly assigned
quasi experimental:
random assignment is not followed
Field Surveys:
non-experimental designs that don't control for or manipulate independent variables or treatments; measure variables and test their effects using statistical methods
cross-sectional field surveys:
independent and dependent variables are measured at the same point in time
longitudinal field surveys:
dependent varialbes are measured at a later point in time
Secondary Data Analysis:
an analysis of data that has previously been collected and tabulated by others
limitations: data may not have been collected well or for the same purpose you are studying
Case Research:
in-depth investigation of a problem in one or more real-life settings over an extended period of time
data collected via interviews, observations, and internal or external documents
good for discovering social, cultural, and political factors; interpretation may depend on observational and integrative ability of researcher
multiple case design:
improves generalizability by replicating and comparing the analysis in other case sites
Interpretive Methods:
inductive approach; starts with data and tries to drive a theory; relies heavily on qualitative data
Popular Research Designs:
Interpretive
Focus Group Research:
bringing in a small group of subjects at one location and having them discuss a phenonmenon of interest
best suited for exploratory research; can be positivist sometimes
Action Research:
complex social phenomena are best understood by introducing interventions or "actions" and observing effects
Ethnography:
emphasizes research phenomenon must be studied within the context of its culture; inspired by anthropology
researcher is deeply immersed in a certain culture over an extended period of time
Quality of Research Designs:
can be defined in terms of 4 key design attributes
Internal Validity:
causality
examines whether the observed change in dependent variable is caujsed by a change in an independent variable
covariation of cause and effect:
if cuase happens, then effect happens too
temporal precedence:
cause must precede effect
spurious correlation:
no plausible alternative explanation
lab experiments are strong in internal validity; field surveys are not
Improving Internal and
External Validity:
best research designs can ensure high levels of internal and external validity
Manipulation:
researcher manipulates independent variables and compares the effects of treatments against a control group.
ex: medication testing
Elimination:
eliminate extraneous variables by holding them constant across treatments;
ex: restricting a study to a single gender
Inclusion:
role of extraneous variables is considered by including them in the research design and separately estimating their effects;
ex: factorial designs where one factor is gender
Statistical Control:
extraneous variables are measured and used as covariates during the statistical testing process
Radomization:
aimed at cancelling out the effects of extraneous variables through random sampling
random selection:
sample is selected randomly from a population
random assignment:
subjects selected in a non-random manner and randomly assigned to treatment groups
External Validity:
generalizability
observed associations can be generalized from the sample
population validity:
can be generalized to the population
ecological validity:
can be generalized to other people, organizations, contexts, or time
survey research can be more easily generalized than lab research
Construct Validity:
how well a given measurement scale is measuring the theoretical construct that it is supposed to
assessed in positivist research based on correlational or factor analysis of pilot test data
Statistical Conclusion Validity:
extent to which conclusions derived using a statistical procedure are valid
not applicable for interpretive research designs
Selecting Research Designs
many researchers will choose what they are comfortable with, but the research design should be based on the nature of what is being studied
Positivist for theory testing;
Interpretive for theory building
strive to collect quantitative and qualitative data using a combination of techniques such as questionnaires, interviews, observations, documents, or secondary data
goal of the researcher should be to collect as much and as diverse data as possible that can help generate the best possible insights about the phenomenon of interest
Requires
ensures