Statistics Norway has carried out a Consumer Expenditure Survey household budget survey every year since 1974. During the last years considerable resources have been allocated to keep the response rate at a “tolerable” level of about 50 percent. The Consumer expenditure survey is a survey with a high response burden for the respondent. First, there is face to face interview. Second, the household fill in a diary for two weeks. Third, there is a closing face to face interview when the interviewer comes and collects the diary. During the data collection for the 2008 survey, SSB conducted an experiment in order to test whether a reduced burden could have a positive impact on the response rate. An experiment group was asked to fill in the diary for one week instead of two weeks, and the response rates were compared with a control group. Our finding supports the hypothesis that reduced burden has a positive impact. But, we also need to find out whether the data on household expenditure can be of acceptable quality even if the respondents fill in for one week instead of two. Our working hypothesis is that the measurement error in the data can be said to be equal.

Keywords: Non sampling errors; measurements errors; Consumer Expenditure Survey

Biography: Øyvin Kleven is Head of Division at the Division for Survey Planning and User Testing at Statistics Norway. He has a master in political science at University of Oslo in 2000. He joined Statistics Norway in 1998. From 1998-2000 he was a project leader for several surveys and from 2000-2004 he was group leader at the division for sample surveys. From 2004 he's main task was research and development in sample surveys. His current research interests include studies of political participation, sample survey design and analysis, non-response errors and measurement errors.