The Power of Survey Design

Are you in favor of financial incentives for poor countries?" If this question were asked in a survey many would be inclined to agree. Yet the result of this poll would be different if the question asked was "Are you in favor of subsidies for poor countries?" This is a simple example of how one single word, incentives or subsidies, can change the answers to virtually the same question.

How we ask the question can often lead the respondent in one direction or the other. And this effect can be significant, in the order of up to 30%. Hence a skillful questionnaire designer can "demonstrate" popular support by wording the question in a way congenial to his or her desired objective.

Similarly the international comparison of survey results is today a common occurrence, yet rarely survey design effects are taken into account when results are presented. So for instance underreporting will certainly be present if a question on corruption is asked by a government official. Henceforth if we wish to obtain a meaningful comparison of this phenomenon across countries we must control for this effect. If we don't China's corruption level will appear lower than Honduras', contrary to what Transparency International reports.

This handbook aims at showing the multitude of survey design factors that play a subtle but crucial role in the accuracy of survey data and can taint its interpretation. A practical how-to guide on all the steps involved with survey implementation, this volume covers survey management, questionnaire design, sampling, respondent's psychology and survey participation, and data management. A comprehensive and practical reference for those who both use and produce survey data.