Abstract

For more than 100 years the Registrar General has reviewed mortality in depth in a series of supplements relating extra information provided by decennial censuses to deaths in a period before and after the census. The volume describing occupationl mortiality in 1970--72 was recently published (Registrar General, 1978). Here we consider in more detail one of the questions raised by occupational mortality studies: how much does mortality of an occupation group reflect work environment and how much way of life? We first describe the traditional method of distinguishing these direct and indirect influences (that is, the comparison of the mortality of men following an occupation with that of their wives) and then introduce an alternative which we call 'social class standardisation'.