Consider the types of information different researchers may need regarding the number and characteristics of cancer patients. Researchers may be looking for:
The number of people who began to suffer from different types of cancer in a particular year
How cancer incidence is related to certain risk factors—such as family history or exposure to tobacco or alcohol or
The outcome of alternative treatment protocols—the survival rates of patients who were treated with a particular procedure and their quality of life.
Each of these inquires implies a different standard of precision and accuracy.
It may only be possible to know the number of diagnosed cases that might be only a fraction of those who were affected by the early stages of the disease. By comparing the known cases of one type of cancer with others or with previously reported rates, the researcher may be able to answer the underlying question implied. Good comparison figures may be more important than the last degree of precision.
Relating cancer to identified risk factors involves inferences that depend on the availability of controls that can hold other factors constant.
The effectiveness of protocols is best established through controlled studies. Reliance on statistical data from a population alone—could yield only vague results.
Thus, the evaluation of statistical results reported depends as much on the questions asked as it does on the data available. The search therefore critically depends on understanding the connection between the implicit methodological questions that underlie each search.
We discuss cancer statistics further later in this course. See the section on SEER.