Methodology Report Coverage Estimation for the Swiss Population Census 2000

Coverage of the Swiss population census is estimated for the first time for the census 2000. Both undercoverage and overcoverage are analyzed apart and then combined by using the dual system methodology. The estimates are based on two samples: the Enumeration sample (E-sample) and the Population sample (P-sample) in order to capture both the overcoverage and the undercoverage components. Similar to results in other countries, we determined that 1.6% of the resident population were overlooked in the census (undercount) and that 0.4% were counted erroneously (overcount). The resulting overall rate of net undercoverage is 1.4% with larger values for some subgroups of the population such as 20-31 years-old people (2.8%) or foreigners (2.9-3.5%). Other types of errors were analyzed such as error in the type of domicile, time delay between census day and effective data collection day for movers around the census day, or potential misclassification variables. The results and experience gained during the project can be used to improve the subsequent censuses.