Chapter 7
Having Designs on Study Design
IN THIS CHAPTER
Grasping the hierarchy of study designs in epidemiology
Appreciating relative strength of different study designs for causal inference
Getting the details about different study designs
Biostatistics can be seen as the application of a set of tools to answer questions posed through human
research. When studying samples, these tools are used in conjunction with epidemiologic study designs
in such a way as to facilitate causal inference, or the ability to determine cause and effect. Some study
designs are better than others at facilitating causal inference. Nevertheless, regardless of the study
design selected, an appropriate sampling strategy and statistical analysis that complements the study
design must be used in conjunction with it.
In this chapter, we provide an overview of epidemiologic study designs and present them in a
hierarchy so that you can relate them to the biostatistical approaches described in the different chapters
of this book. We start by looking at broad study design categories such as observational, experimental,
descriptive, and analytic, and move into descriptive study designs including expert opinion, case
studies and case series, ecological (correlational) studies, and cross-sectional studies. We present
analytic study designs next — case-control studies and longitudinal cohort studies — which are
superior to descriptive designs in terms of developing evidence for causal inference, and then move
into the highest-level designs: systematic reviews and meta-analyses.
For a deeper dive into epidemiologic study designs, we encourage you to read Epidemiology
For Dummies by Amal K. Mitra (Wiley) and pay special attention to Chapters 16 and 17, which
are about causal inference and study design.
Presenting the Study Design Hierarchy
Figure 7-1 illustrates the epidemiologic study designs in terms of their relationship with each other in
a hierarchy as they apply to human health research (not animal research or other domains of human
research like psychology). As shown in Figure 7-1, human health research may be split into two types:
observational and experimental.