URS 6806: Designs for Description and Explanation
[Chapters covered in this unit: 2,3]
Research components
- Research Statement/ Question
- Research hypothesis
- Research Design/ Analysis:
- Systematic definition of variables
- Operational definition of variable
- Measurement of variable
- Systematic data collection pertaining to variables
- Inference: Conclusion about research hypothesis (Section 2 of
class deals with this)
a. Research Statement/ Question
- An interesting question about an event or observed phenomenon
- Key emphasis on interesting: Larger social or policy interest [a
convincing answer to the big SO WHAT? question]
Types of Research Question (Purposes of Research)
- Exploration: Better understanding of an observed phenomenon or process
[usual questions: How? What?]
- Description: Profile of a place, people, or events [usual question: What?
Where? When? How?]
- Explanation: Explanation of a phenomenon or process [usual question: Why?]
b. Research Hypothesis
- A specific testable expectation about an observed phenomenon or
process
- A model explaining the characteristics of a variable or
relationships between two or more variables
- Usually based on a general proposition
- Based on one or more observed concepts/characteristics related to the
observed phenomenon or process [i.e. variables]
- Concept: A notion or a social observation [e.g. politeness, crime,
intelligence, unemployment]
- Dimension: A specific aspect of a concept [e.g. politeness could include
several dimensions: a reciprocal smile; opening the door for another person;
waving in recognition; indulging in small talk]
c. Research Design
- Empirical method for testing the premise
- Provide evidence to substantiate or negate the hypothesis
- Three BIG questions:
- How to provide the explanation?
- What type of evidence needs to be provided?
- How convincing is the explanation/ evidence?
- Pretesting and post testing
How to provide the explanation?
- Experimental research designs
- Laboratory based experiments [e.g Double blind experiments]
- Natural experiments
- Quasi experimental research designs
- Non experimental research designs
What type of evidence needs to be provided?
- Cross sectional research design
- Longitudinal research design
- Time series (Long term trends; cyclical trends; seasonal trends;
irregular trends; forecasting)
- Panel Design
- Qualitative research design (case study; focus groups; textual analysis)
Notation for Research Designs
R = Random
O = Dependent variable
X = Independent variableExperimental [with pre-test
(O1)and post-test(O2)]:
R O1 X
O2
(experimental group)
R O1
O2
(control group) Quasi-Experimental [with pre-test (O1)and
post-test(O2)]:
O1 X O2
(experimental group)
O1
O2
(comparison group) Time Series [ pre-test:
O1
through
Oj; post-test: Oj+1
through
Oj+j]
O1
O2 O3
O4 O5 ...
Oj X Oj+1
Oj+2 Oj+3
Oj+4 Oj+5 ...
Oj+j |
How convincing is the explanation/ evidence?
- Internal Validity Threats:
- History, Maturation, Testing, Instrumentation, Statistical Regression,
Selection bias, Experimental Mortality, Design Contamination
- External Validity Threats
- Uniqueness, Selection, Setting, History, Testing, Reaction