Abstract
The classification schemes frequently used to group farms (farm households) according to socio-economic characteristics rely on more or less intuitively based methods; and employ a few — sometimes arbitrarily defined — criteria only. These methodical and conceptual deficiencies of traditional classification schemes will be discussed in the article, referring to the socio-economic classification schemes currently used in the Federal Republic of Germany. Next, the methodical aspects of using cluster analyses techniques for constructing a revised socio-economic classification scheme are outlined and the variable selection procedure is explained. The classification obtained here consists of four types of farm households, divided according to whether the operator and/or other family members have non-farm employment or not.