Item 2.2 Scientific Use Files for the Time Use Survey Wilhelmus.Kloek@ec.europa.eu
Scientific Use Files (SUFs) Confidential microdata made available for research purposes Inside the research entity Partial anonymisation The research entity is responsible for output checking Regulation 557/2013
Time use survey 1 HETUS: Harmonised European Time Use Survey Collected once in 10 years: current proposal concerns 2010 data Microdata available at Eurostat for 18 countries (voluntary) In total 17 countries agreed in principle to produce SUFs
Time use survey 2 The value added is in the specific time use variables These variables are neither identifying nor sensitive The design differs over the countries, notably the number of diary days per respondent and the sample size The sample size is relatively small compared to the Labour Force Survey and Statistics in Income and Living Conditions
Approach to the partial protection Complete removal of some variables: exact date of diary, current education Aggregation: age in 5-years age group, marital status, country of birth, country of citizenship The survey does not have regional breakdowns
Analysis of the disclosure risk Indirect identifiers: age, sex, marital status, country of birth, country of citizenship, life cycle Analyse disclosure risk on the basis of four dimensional tables of the indirect identifiers Analysis of four example countries, two with a relatively large sample (AT, DE) and two with the smallest sample (LU, NL)
Number of unique records after the proposed treatment Austria Germany Luxembourg The Netherlands Number % Table 1 25 0.3 28 19 0.9 Table 2 31 0.4 32 35 1.7 8 Table 3 138 107 5.1 Table 4 54 0.7 74 0.8 3 0.1 Table 5 55 41 0.5 5 0.2 Table 6 165 2.0 1.2
Conclusions from the analysis The number of unique records remains below 1% in most cases Some unique records contain missing values on an indirect identifier Perhaps some need of country specific solutions: e.g. Luxembourg for the education variable
Opinion of the Expert Group on Statistical Disclosure Control 9 out of 12 Members replied All positive Comments on the coherent aggregation of age Request for additional analysis on economic activity and occupation Identification of some mistakes and unclarity in the document Some country specific situations
Conclusion The WG is requested to agree to the general approach for the protection of SUFs for HETUS 2010 Under the condition that the document is updated on the basis of the comments of the Expert Group (notably a coherent aggregation of age)
Next steps Eurostat will update the document and send it to both the Expert Group and the WG If no further comments the method is approved Eurostat will produce the SUFs Add information on the identification risk including the requested tables for economic activity and occupation Send to the specific MS with request to approve This might lead to some country specific protection requests