The Use and Misuse of Surveys in International Development: An Experiment From Nepal

Developing countries rely heavily on Western survey research methods to gather data for national planning. Although it is known that particular problems with "non-sampling errors" may arise when surveys are conducted cross-culturally, little attention has been given to determining the magnitude of the problem. An experiment in Nepal was conducted to measure the accuracy of a popular KAP (knowledge, attitude, practice) survey. This experiment demonstrated that non-sampling error for the kind of knowledge and attitudinal variables measured in the survey was far greater than what is normally computed as a sampling error. One reason for inaccuracy concerned the cultural reinterpretation of survey questions by respondents. In more general terms it is argued that surveys are limited by a "contextual bias" (reliance on only one context for gathering data) that may render survey data inaccurate or misleading. However, it is by no means recommended that Third World surveys be discontinued or be replaced by qualita...