Bitte verwenden Sie diesen Link, um diese Publikation zu zitieren, oder auf sie als Internetquelle zu verweisen: https://hdl.handle.net/10419/95716 
Autor:innen: 
Erscheinungsjahr: 
1993
Quellenangabe: 
[Publisher:] Deutsches Institut für Wirtschaftsforschung (DIW) [Place:] Berlin [Year:] 1993
Schriftenreihe/Nr.: 
DIW Discussion Papers No. 81
Verlag: 
Deutsches Institut für Wirtschaftsforschung (DIW), Berlin
Zusammenfassung: 
Panel attrition has not only the potential to bias population estimates but it may also inflate the variance of the estimates from panel surveys. Thus it is essential for an ongoing panel survey to monitor not only the size of the panel attrition and the potential biases that may occur but also the decrease in the precision of estimates. Losses in the precision of cross-sectional estimates are also due to a second source, which is specific to the design of household panels. This effect, stems from different inclusion probabilities of households depending on the number of occasions when persons move into the household. Panel surveys offer substantial efficiency gains in trend analysis, where differences in time with respect to population charcteristics are estimated. Such an analysis may also be performed on the basis of a sequence of independent cross-sectional surveys. It is investigated here, how the panel attrition effects the efficiency gains of the panel compared to a sequence of independent cross-sections. Panel attrition also offers the'possibility of'different population estimates for the same population characteristic. It is investigated which of the alternative estimates is more precise. In order to give an answer to the above topics two panel specific approaches of variance estimation are presented. Numerical examples are displayed from the German Socio-economic Panel (SOEP).
Schlagwörter: 
Sampling variance
household panel
panel attrition
Dokumentart: 
Working Paper
Dokumentversion: 
Digitized Version

Datei(en):
Datei
Größe
1.49 MB





Publikationen in EconStor sind urheberrechtlich geschützt.