For this
reason the design phase should be very detailed, and followed by a pilot testing
phase aimed at assuring that the survey instrumentation and procedures are
adequate.
However, in planning the activities the decisions made during the early steps
affect the choices remaining at later steps (see Figure 2). It is not possible to
proceed step by step: constraints and limitations in the later steps should be
considered in the earlier steps. For these reasons, major decisions about data
collection (telephone, interview and mail) and time horizon (cross-sectional or
longitudinal) must always be made prior to designing and selecting a sample
and constructing the questionnaire and the other material. It is important to
match the capabilities and the limitations of the data-processing methods with
the sampling and instrumentation. For more details on project planning see
Alreck and Settle (1985).