Common standards for editing, imputation, weighting and estimation
Article 9
1. Administrative or register data, results from previous interviews, and results from interviews of another person shall not be used to replace or impute information on the variables WKSTAT, ABSREAS, JATTACH, SEEKWORK, ACTMETNE and AVAILBLE referred to in Annex I of this Regulation. Simplified rules may be applied for certain groups of persons, as specified in Annex II of this Regulation. 2. Where information on other variables is missing, invalid or inconsistent, methods of statistical imputation may be applied where appropriate, except for the variable INCGROSS referred to in Annex I of this Regulation, to which statistical imputation shall always be applied when non-response exceeds 5 %. 3. The reference population for weighting shall be the (real or estimated) population usually residing in private households. 4. Weighting factors for estimates at individual level based on quarterly, annual, and biennial variables shall fulfil the following requirements: (a) the weighting factors shall be calculated taking into account the probability of selection and external data on the distribution of the population being surveyed, by sex, age groups and region (NUTS 2 level). Five-year age groups shall be the standard. However, given the sample size and the quality and availability of the external data, aggregations that include more than one five-year age group are allowed only to the necessary extent; (b) consistency between annual totals of sub-samples for annual and biennial structural variables and full-sample annual averages shall be ensured for employment, unemployment and outside the labour force by sex and for the following age groups: 25-34, 35-44, and 45-54. Consistency for the groups of persons aged 15-24, 55-64, and 65 and over shall be achieved to the fullest extent possible. 5. Weighting factors for estimates at household level, using the average weight of the household members, and at individual level by specific household characteristics shall fulfil the following requirements: (a) the weighting factors shall be calculated taking into account the probability of selection and data on the distribution of the private households being surveyed, namely the (real or estimated) number of households and the (real or estimated) household size (household level), and the distribution of the population being surveyed, by sex and at least age groups 0-14 and 15+ (individual level). Estimates of the number of households and the household size shall be based on the best sources and concepts available; (b) consistency between annual full-sample or sub-sample totals using the household weighting factors and full-sample annual averages using the individual weighting factors defined in paragraph 4 shall be ensured for employment, unemployment and outside the labour force by sex and for the following age groups: 25-34, 35-44, and 45-54. Consistency for the groups of persons 15-24, 55-64, and 65 and over shall be achieved to the fullest extent possible. 6. Weighting factors for estimates at individual level based on eight-yearly/ad hoc subject variables shall fulfil the following requirements: (a) the weighting factors shall be calculated taking into account the probability of selection and data relating to the distribution of the population being surveyed, by sex and eight-yearly target population age groups as defined in paragraph 3 of Article 4 of this Regulation. Five-year age groups shall be the standard. However, given the sample size and the quality and availability of the data, aggregations which include more than one five-year age group are allowed only to the necessary extent; (b) consistency between annual sub-sample totals using the eight-yearly weighting factors and annual sub-sample totals or, if not applicable, full-sample annual averages using individual weighting factors referred to in paragraph 4 shall be ensured for the eight-yearly target population defined in paragraph 3 of Article 4 of this Regulation and for employment, unemployment and outside the labour force by sex; (c) the requirements shall apply accordingly to ad hoc subjects and their target populations. 7. Information on the variables WKSTAT, ABSREAS, JATTACH, SEEKWORK, ACTMETNE and AVAILBLE referred to in Annex I of this Regulation shall not be used in the weighting process.