Detailed sample characteristics
Article 6
1. Only Member States that used the selected respondent model before 19 January 2020 shall be authorised to continue using the method over the following years. 2. Controlled substitutions of sample households or persons may be allowed only in the first year of each panel, if the response rate falls below 60 % and one of the following situations arises: (a) the sample household or person is not contacted because it/she/he cannot be located or is inaccessible; (b) the sample household or person is contacted, but the interview is not completed because the household or person refuses to cooperate, the whole household is temporarily away, the household or person is unable to respond or there are circumstances endangering the interviewer’s safety. 3. The set of sample households or persons for substitution shall be defined prior to data collection. There shall be no substitution with households or persons not belonging to that set. 4. Procedures shall be followed to ensure that the process of substitution is controlled to the maximum extent possible. Such procedures shall include using a design which ensures that the selected substitutes closely match the households or persons they replace in terms of their significant characteristics. 5. Small parts of a national territory amounting to no more than 2 % of the national population may be excluded from EU-SILC, as may French overseas departments and territories. Both exclusions shall not take place simultaneously. 6. If subpopulations covered by the survey are significantly under-represented in the achieved sample owing to specific non-response or attrition, Member States shall take measures such as oversampling, calibration or applying any relevant method to correct for the resulting bias. The methods used, their impact and any drawbacks shall be explained and assessed in the quality report.