This question requires us to understand the differences between the following research methods: observational studies, surveys, and experiments. We will do this by identifying the marked characteristics of each research method and understand when each should be used in research.

First, we will discuss observational studies. Observational studies can be described as anthropologic or ethnographic studies. In these studies, researchers observe a particular event and develop conclusions on these groups or phenomena based upon direct observation. For example, researchers have explored tribal groups in many countries as well as social phenomena such as the baby boomer generation as well as deviant groups (e.g. gangs and paramilitary organizations)

Second, surveys are typically composed of questionnaires or interviews. They are used to ask a sample of participants questions in order to gather information that is to be generalized to the greater population. Survey research often employs the use of random sampling techniques in order to get a representative sample of the population. In this way, they have a high level of external validity. In other words, the sample statistics derived from the study may be used to make valid inferences of the population parameters of society at large. An example of survey research could include restaurant questionnaires related to satisfaction as well as the United States Census (it is important to note that a census has high external validity because it attempts to sample every member of a population). Another survey with high external validity could be a survey that takes a simple random sample of high school, college, and secondary school graduates and their occupational wages post graduation. This survey could find a correlation between level of education attained and potential income. It is important to note that correlation does not equal causation. A researcher could find that shark attacks and ice creams sales at the beach are positively correlated; however, this does not mean that shark attacks cause ice cream sales or that ice cream consumption leads to shark attacks. Instead, both of these particular phenomena are caused by a hidden or lurking variable: warm weather and greater numbers of beach goers.

Last, we will discuss experiments. Experiments employ methods of random sampling; furthermore, they use procedures to assign subjects and variables to control or treatment groups. The treatment group receives the treatment in the trial while the control group is given a substitute that would not alter standard conditions. For example, scientists studying the effect of a blood pressure pill on a group of individuals with high blood pressure would give the treatment group the pill and substitute the treatment for a placebo or sugar pill in the control group. Experiments use double blind methodologies to assign subjects to control and experimental groups. These methods randomly assign subjects to either group; furthermore, they hide the assignments from both subjects and researchers. This prevents experimenter biases as well as limits the placebo effect—in which subjects tend to display the symptoms of the treatment group while being part of the control group. These studies collect representative samples and control for extraneous variables; therefore, they have strong external validity. It is important to note that experiments directly manipulate variables and can be used to identify causal relationships; therefore, they can be said to be strong predictors of causality.

Now, let's use this information to answer the question. The characteristics of these research methods have been tabulated in the following table:

Observational studies use purposeful samples not random samples. In other words, they pick a particular group or event and study it directly in order to learn about it with little desire to extend this information to the greater population; therefore, they also lack external validity.

In the problem, researchers used a questionnaire in order to find the preferences and satisfaction of customers at a specialty cheese company. If we look at the characteristics given we can see that they are not assigning a treatment and they are not trying to indicate causality (instead, satisfaction). On the other hand, they are trying to generalize these results to the greater population of cheese consumers (i.e. from the feedback of recent consumers) and they are using a random sample of recent consumers. From this information, we can see that they have no treatment assignment or investigation of causality; furthermore, they want a high external validity and are using methods of random sampling. They are using a survey research design.