Selektivitätsprozesse bei der Verknüpfung von Befragungs- mit Prozessdaten : Record Linkage mit Daten des Panels "Arbeitsmarkt und soziale Sicherung" und administrativen Daten der Bundesagentur für Arbeit

"Linking survey data with register data can combine the advantages of both types of data and increase the possibilities for analysis compared to the use of the single data sets. The exact linkage of records from the same individual is only legal with the content of the respondent. Moreover, not all consenting persons can actually be merged with the process data. It is possible that certain characteristics of the respondents have an impact on the consent behavior. Also the likelihood of actually realizable linkage in a technical sense may depend on person-specific characteristics. Therefore it must be examined, whether this leads to a selective sample, and eventually to biased results when using linked data. In this report selectivity bias due to the linkage of the panel 'Labour Market and Social Policy' with administrative data from the Federal Employment Agency (BA) is investigated. The results show a low selectivity of the sample due to the response to the question of the content to link BA register data. The selectivity due to the realized merges is limited primarily to groups of people without any information stored in the BA register data. To verify whether the research results will be affected if solely individuals are included in the analysis that agree to the linkage, and subsequently can actually be linked, a regression with the logarithm of gross hourly wage as the dependent variable is performed. In this case, no significant differences could be found between the regression coefficients of a full and a reduced model, which would indicate a bias of the data." (Author's abstract, IAB-Doku) ((en))