© John Wiley & Sons, Inc.

FIGURE 13-1: These designations for cell counts and totals are used throughout this chapter.

Choosing the Correct Sampling Strategy

In this section, we assume you are designing a cross-sectional study (see Chapter 7 for a review of

study design terminology). Using such a design, though you could not assess cause-and-effect, you

could evaluate the association between an exposure (hypothesized cause) and outcome. For example,

you may hypothesize that being obese (exposure) causes a patient to develop hypertension

(abbreviated HTN, outcome) over time. However, in a cross-sectional study, exposure and outcome

are measured at the same time, so you can only look for associations. If your exposure and outcome are

binary (such as obese: yes/no and HTN: yes/no), you can use a fourfold table for this evaluation. But

you would have to develop a sampling strategy that would support your analytic plan.

The example given here is for the use of a fourfold table to interpret a cross-sectional study. If

you have heard of a fourfold table being analyzed as part of a cohort study or longitudinal study,

that is referring to a series of cross-sectional studies done over time to the same group or cohort.

Each round of data collection is called a wave, and fourfold tables can be developed cross-

sectionally (using data from one wave), or longitudinally (using data from two waves).

As described in Chapter 6, you could try simple random sampling (SRS), but this may not provide you

with a balanced number of participants who are positive for the exposure compared to negative to the

exposure. If you are worried about this, you could try stratified sampling on the exposure (such as

requiring half the sample to be obese, and half the sample to not be obese). Although other sampling

strategies described in Chapter 6 could be used, SRS and stratified sampling are the most common to

use in cross-sectional study. Why is your sampling strategy so important? As you see in the rest of this

chapter, some indices are meaningful only if the sampling is done accordingly so as to support a

particular study design.

Producing Fourfold Tables in a Variety of