9.1: Understand the difference between the three main types of quantitative research designs (randomized experimental, quasi-experimental, and observational) and evaluate their internal and external validity.
9.2: Understand the distinctions between posttest-only, repeated-measurement, and multigroup experimental designs.
9.3: Discuss the use of field experiments and natural experiments in political science research and understand how they differ from laboratory experiments.
9.4: Describe the quasi-experimental design and its limitations.
9.5: Describe the two types of observational research designs--cross-sectional designs and longitudinal designs--and explain how they contribute to our understanding of political phenomena.
- There are various randomized experimental designs that are discussed in this chapter:
- The posttest design involves two groups and two variables, one independent and one dependent. Subjects are randomly assigned to one or the other of two groups. One group, the experimental group, is exposed to a treatment or stimulus, and the other, the control group, is not or is given a placebo.
- The repeated-measurement design is a plan that calls for making more than one measure or observation on a dependent variable at different times over the course of the study.
- The multiple-group design is an experimental design with more than one control and experimental group. Multiple-group designs may involve a posttest only or both a pretest and a posttest. They may also include repeated measurements.
- A field experiment adopts the logic of randomization and variable manipulation but applies these techniques in naturally occurring situations not in laboratories.
- In natural experiments, forces outside of the investigators control assign individuals or units to “treatment” and “control” groups.
- Assignment to groups may involve true random assignment or, in some cases, the assignment is “as-if” random.
- Researchers are able to observe but do not themselves manipulate the operation of the “experimental” factor.
- The chapter also discussed nonrandomized designs, such as:
- A quasi-experimental design contains treatment and control groups, but the experimenter does not randomly assign individual units to these groups.
- As one moves from randomized designs to quasi-experiments, inferences about causal effects demand more and more of the researcher’s substantive knowledge and analytic skills.
- Observational studies are used to describe designs in which the researcher neither manipulates experimental variables nor randomly assigns subjects to treatments but instead merely observes causal sequences and covariations.
- Cross-sectional designs: characterized by measurements of the independent and dependent variables at approximately the same time including surveys and aggregate data analysis.
- Longitudinal or time-series designs are characterized by the availability of measures of variables at different points in time.
- Benefits of longitudinal studies include the fact that they can in principle estimate three kinds of effects: age effects, period effects, and cohorts.
- In an intervention analysis or “interrupted time-series analysis,” measurements of a dependent variable are taken.
- A trend analysis is a research design that measures a dependent variable at different times and attempts to determine whether the level of the variable is changing and, if it is, why.
- No matter which design, however, the emphasis is on discovering the causes of political phenomena.