ABSTRACT
Survey methodology finds many applications in environmental psychology. The existence of large-scale longitudinal data sets has inspired both methodological and substantive research and has drawn attention to the need for developing new methodological tools for their analyses. A survey is used to obtain information, not from every member of the population, but only from a sample selected using probability methods. Probability sampling methods have a long history of theoretical development, but recently methodological attention has turned to those factors that can destory the value of information from surveys even when probability samples are employed—problems of nonresponse and mistaken responses. The Panel reviewed and compared procedures used for incomplete data, summarized theory and methods for field procedures, data processing, and estimation, and made suggestions for reporting surveys so that results of nonresponse can be taken into account.
