CHAPTER 6 Taking All Kinds of Samples 85

know what students think about a new policy on campus, they can just ask who-

ever is in their classes, as those students are a convenient sample of the student

population.

The problem is that the answer they get may be very biased. Most of the students

in their classes may come from the sciences, and those studying art or literature

may feel very differently about the same policy. Although our convenience

sample would be a valid sample of the background population of students, it would

be such a biased sample that the results would probably be rejected by the rest of

the faculty — especially those from the art and literature departments!

Given that the results from convenience samples are usually biased, you may

think that convenience sampling is not a good strategy. In actuality, convenience

sampling comes in handy if you have a relatively low-stakes research question.

Customer satisfaction surveys are usually done with convenience samples, such as

those placing an order on a restaurant’s app. It is simple to program such a survey

into an app, and if the food quality is terrific and the service terrible, it will be

immediately evident even from a small convenience sample of app users complet-

ing the survey.

While low-stakes situations are fine for convenience sampling, high-stakes

situations — like studying whether a new drug is safe and/or effective — require

study designs and sampling approaches completely focused on minimizing bias.

As with SRS, convenience sampling is prone to omitting important subgroups

from the sample. Minimizing bias through sampling and other strategies is cov-

ered in detail in Chapter 5, which examines clinical research and describes how

researchers must present a well-defined protocol that includes selection criteria,

a sampling plan, and an analytic plan that undergoes regulatory approval prior to

the commencement of research activities. Other strategies for minimizing bias are

presented in Chapters 7 and 20, which cover study designs and causal inference.

Sampling in multiple stages

When conducting large, epidemiologic surveillance studies, it is necessary to do

an especially good job of sampling, because governments use results from these

studies on which to base public policy. As an example, because being obese puts

community members at risk for serious health conditions, government public

health agencies have a vested interest in making accurate estimates of the rates of

obesity in their communities.

For this reason, to strive to obtain a representative sample, researchers designing

large epidemiologic surveillance studies use multi-stage sampling. Multi-stage

sampling is a general term for using multiple sampling approaches at different