This paper uses workforce statistics in the archive of the Engineering Employers Federation to re-examine data from official reports and contemporary accounts on the skill composition of the engineering workforce. It suggests that skill composition, at least among adult males,
changed more slowly than might be thought from reports suggesting a fall in the skilled proportion of the workforce from 60% in 1914 to 40% in 1926. This provides some support for those sceptical about deskilling who draw attention to the ways in which differentiated product markets and associated
technologies combined to sustain a higher level of demand for skill than might otherwise have been the case. At the same time, it is argued that workplace realities underpinning categories like 'craft' and 'skill' shifted rather more significantly than is often acknowledged. Qualitative and
quantitative changes combined to create a new hierarchy of skill quite different from that which had gone before. The implication, more fully developed elsewhere, is the need to revise long-standing views of the ways in which continuity of craft organization inspired later post-Second World
War shop floor trade unionism and the disorderly bargaining of the 1960s.