Coping with decreasing response rates in Statistics Norway. Recommended practice for reducing the effect of nonresponse

Abstract

Over the last two decades, cooperation rates in most sample surveys appear to have declined in many countries, including Norway. As a consequence, the cost of conducting surveys has increased because repeated attempts must be made in order to seek information from reluctant sample members. In this report, we provide a review of experiences in Statistics Norway in the field of survey nonresponse. The report presents a recommended practice manual, RPM, based on these experiences. As nonresponse is part of quality, we treat the response process as part of the production process in a systematic quality approach, and recommend methods and techniques applied to reduce unit nonresponse and its effects on the final official statistics. The report is divided into five chapters: Chapter 1 provides an introduction of some basic concepts within the systematic quality approach and some recommendations. Chapter 2 presents some response rates and nonresponse biases and their development over time and some international comparisons. Chapter 3 is concerned with household surveys. Key factors that have important impact on nonresponse are identified. For many of these factors, key process variables are measured and applied to find a good balance between cost and quality. Chapter 4 gives a similar presentation concerning business surveys. Chapter 5 presents techniques used to investigate the nature of the effects on nonresponse on the population estimates together with the most commonly used weighting methods.