Not Willing, Not Able: Causes of Measurement Error in Business Surveys

National statistical institutes must collect accurate data from businesses in a timely and cost-effective way and without causing too much response burden. An adequate design of the information request is critical in achieving this goal. This paper describes the lessons we have learned about the design of business survey questionnaires from a thorough evaluation of the questionnaires of a typical business survey for official statistics, the Structural Business Survey. The paper presents a framework for understanding factors that contribute to missing and inaccurate data and draws a number of conclusions regarding how the design of business surveys can be improved to take these factors into account.