Social mobility research has consistently demonstrated that a strong link exists between an individual's social class origins and their social class destination. Yet, in absolute terms, the relationship between class origins and destinations does appear to be weaker now than it was in the past. There has been a increase in the numbers of those of all class origins entering higher-level service class occupations, and a decrease in those entering manual employment. However, while this would appear to be evidence of a high degree of fluidity in the class structure, these patterns can largely be explained by changes in the occupational structure: the expansion of non-manual occupations while manual ones have contracted (Goldthorpe 1987:59). When these structural effects are controlled for by looking at relative mobility rates, it is clear that there has been virtual stability in such rates over time. A relationship between origins and destinations is still in evidence (Goldthorpe 1987; Erikson and Goldthorpe 1992). The relationship between class origins and destinations has therefore become widely established in social mobility research, and as a result, the emphasis is now moving towards explaining how these trends are created (Saunders 1995; Breen and Goldthorpe 1999).