A research design refers to how you go about conducting your research. The choice of a research strategy is largely determined by the type of research question, the available means, and the units of analysis. Most introductory text books (including the one written by Earl Babbie, which is used at this faculty) do not distinguish clearly between research designs and data collection methods.
Research designs that can be used to explain things can be broadly categorised as either (quasi-)experimental, and as non-experimental. Within these two broad categories there are several sub-categories, depending on factors such as the number of groups analysed, the number of times each case is observed, and the number of cases.
(Quasi-)Experimental designs
Experimental designs refer to the broad category of research designs that involves at least two groups, administering a treatment and observing the consequences. Within this category, various sub-categories exist, depending on:
- How many groups are examined. A two-group design involves one control group and one treatment group. A factorial design involves several treatment groups, relating to different independent variables and different levels of these variables.
- How subjects are allocated to different groups. In true experiments, subjects are allocated to different groups on the basis of random assignment. In quasi-experiments, subjects are not allocated to different groups on the basis of random assignment.
- How many times the groups are observed. In post-test only designs, observations are only made after the treatment is administered. In pre-test post-test designs, observations are made both before and after the treatment is administered. When a series of observations are made before and/or after the treatment the design is even more complex.
Experiments (preferably using random assignment) are generally the best way to make valid causal inferences; however, many research questions are not amenable to experiments and the external validity of experiments is disputed.
Non-experimental designs
Non-experimental designs involve making observations with a single group. Within this category, various sub-categories exist, depending on:
- How many cases are studied. Quantitative non-experimental designs involve a large number of cases, often chosen by random selection, are also called ‘correlational designs’; qualitative designs involve a small number of cases or a single case.
- The number of time points at which observations are made. In cross-sectional designs observations are made at one point in time. In longitudinal designs repeated observations are made over time. Longitudinal designs can be further distinguished depending on whether the same units of observation are observed over time (panel studies), whether the same cohort is followed over time (cohort analysis) or whether only a series over random samples is used with different units of observation (trend studies).
Quantitative, non-experimental designs can be used to make accurate descriptive inferences about a population. This means they are strong at external validity. As ‘correlational designs’ they are also used to tackle explanatory questions. Qualitative research is the best approach to gaining insights into a topic with a view to developing hypotheses (exploratory research). Qualitative designs are also used to provide tentative tests of hypotheses on research questions that cannot be answered using experiments or quantitative non-experimental designs. The external validity of qualitative designs, however, is disputed. An extensive overview of qualitative designs can be found here.
Readings
Basic readings
- Babbie, Earl (2020). The Practice of Social Research (15th edition). Cengage. Chapter 4.
- Shadish, William R., Thomas D. Cook and Donald T. Cambell (2002). Experimental and Quasi-Experimental Designs for Generalized Causal Inference. Boston: Houghton Mifflin. Chapters 1, 4-6.
- De Vaus, David (2001). Research Design in Social Research. London: Sage.
Additional readings
- Spector, Paul E. (1981). Research Designs. Beverly Hills: Sage.
- Przeworski, Adam and Henry Teune (1970). The Logic of Comparative Social Inquiry. Malabar, Florida: Kreidger Publishing Company.
- King, Gary, Robert O. Keohane and Sidney Verba (1994). Designing Social Inquiry: scientific inference in qualitative research. Princeton: Princeton University Press.
- Lieberson, Stanley (1985). Making it Count: the improvement of social research and theory. Berkeley: University of California Press.
- Yin, Robert K (1994). Case Study Research: Design and Methods (2nd edition). Thousand Oaks, California: Sage.