Construct a Sampling Plan (Step 5)

In disaster studies it is necessary to obtain information from several groups of those involved and affected by the disaster and from various types of medical and administrative documents. We have suggested in Chapter 8 that information be obtained from, at a minimum, (1) survivors/victims/families; (2) professional responders and coordinators (both public safety and EMS/medical); (3) officials of governmental and non-governmental organizations; (4) medical records; and (5) administrative documents. This typically involves accessing a large number of individuals and reports. It is therefore almost always necessary to select a sample from each group and source. When possible some form of random (probability) sampling should be used; a different type of sampling called ‘purposive’ may be employed for key informants. ‘Convenience samples’ are not generally used in scientific evaluation studies due to their great potential for introducing bias into the data. Preparing and implementing a scientific sample design will prove to be one of the most challenging aspects of disaster evaluation studies. It will usually be necessary to consult with an individual who has statistical expertise when preparing the sampling plan; therefore, in this chapter we present some basic concepts in sampling and then conclude with four descriptions of sample designs used in past evaluation studies.