it is true that the income and consumption survey determines expenditure and not need; but in a society in which people do generally support themselves with products they produce themselves it is sufficiently plausible to conclude to need from consumption. As the income and consumption survey collects data on household items and consumer behaviour of private households by way of a voluntary random survey, this data is just as open to error as any other empirical collection of data, yet it depicts consumer behaviour of the population in a way that is statistically sufficiently reliable. The Federal Statistical Office (Statistisches Bundesamt) attempts to verify the voluntary entries in the household diaries of the respondent reference groups in a technically appropriate manner, and monitors the results by way of plausibility checks. There is also no objection to the fact that individual items from the income and consumption survey included in the calculation of standard needs are not backed by published figures. Certain data are not shown by the Federal Statistical Office with precise numbers for reasons of data privacy, which must be respected here as well. […] bb) There is no constitutional objection to the selection of reference households by the legislature in order to determine standard needs. 96 It is generally permissible for the legislature to decide whether to orient the level of benefits for standard needs to the consumer behaviour of the lower income groups as determined in the income and consumption survey. The legislature may assume that expenditure is increasingly made over and above what is necessary for a minimum existence in higher income groups (cf. BVerfGE 125, 175 <234>). 97 The Constitution does not oblige the legislature to orient its determination of the level of standard benefits to the lower 20 % of single person households stratified by net income, like it did with the 2003 income and consumption survey. The decision to use only the lower 15 % of households as the reference group with regard to the 2008 income and consumption survey […] does not violate the Constitution. It is not for the Federal Constitutional Court to examine whether the choice of a different reference group would have been more appropriate. The decisive factor is that the choice of reference group is objectively justifiable. This is the case. […] 98 The reference group is also broad enough to collect statistically reliable data (cf. BVerfGE 125, 175 <236>). Generally[…], the reference group must be large enough to provide a sufficiently large random sample for statistical purposes. Here, it was determined subject to the proportion of households removed from the survey (Bundestag printed paper 17/3404, p. 89; Bundestag printed paper 17/3982, p. 2). With 1,678 one-person households (Bundestag printed paper 17/3404, p. 139; Bundestag printed paper 17/3982, p. 1), it was sufficiently large. 99 cc) Similarly, there is no constitutional objection to the legislature using the standard needs assessed for one-person households, that is, standard needs level 1, as the basis for setting the value of the level of benefits for standard needs of such adults who live in a joint household with other adults who are also eligible for benefits, that 100 17/29