With the exception of 2001, these data have been created for Vision of Britain by re-districting statistics originally reported for other units. We have also had to deal with variations in the categories and classifications used in statistical reporting over the years.

• 1841: The data come from the Occupational Abstract of the 1841 census, which the project has computerised in full. The calculation of numbers in each of the six sectors used here was based on the occupations listed in the census report data having already been assigned to the occupation units used by the 1921 census occupational tables, using the 'Alphabetical List' (pp. 91-191) in Census 1921. Classification of Occupations (London: HMSO, 1924). We then assigned these occupation orders to social classes by following table A, 'Occupational Mortality, Legitimate Fertility and Infant Mortality', pp. ciii-cxiv in the Registrar General's Decennial Supplement for 1921, Part II. Occupational Mortality, Fertility and Infant Mortality (London: HMSO, 1927).
• 1881: The data are derived from the full transcription of the 1881 Census enumerators' books organised by the Genealogical Society of Utah. The raw data therefore contained all the individual occupational descriptions, but we worked with a dataset supplied by Kevin Schürer and Matthew Woollard of Essex University, which had already been grouped into the 414 occupational categories used in the county-level tables of the 1881 census reports. These were then assigned to social classes using the 1921 Classification of Occupations and , as above.
• 1931: The 1931 census did not tabulate social class, only occupations, but national social class statistics were published by the Registrar General and we have been able to almost exactly reproduce those calculations at a local level. The original sources are tables 16 'Occupations of Males and Females aged 14 years and over', for county totals and individual large towns, and table 17 'Occupations (Condensed List) of Males and Females aged 14 years', for small towns and Rural Districts, in Census, 1931: Classification of Occupations (London: HMSO, 1934). These tables use an occupational classification whose most detailed level is 674 'occupation units', and table 1 in the Registrar General's Decennial Supplement, Part IIA, Occupational Mortality puts each of these into one of the five social classes. The census table for large towns listed numbers in all the individual occupation units, so we were able to precisely compute numbers in each social class. The census table for other districts, table 17, lists numbers in the 32 occupation orders plus selected occupation units. Probably not by coincidence, these selected units are particularly useful ones for the calculation of social class. They include '658-9 Messengers and porters', '920-930 General and undefined labourers' and '940 Unskilled workers in factories, works, etc.', which together comprised 62% nationally of all men in social class 5. Similarly, the data provide precise numbers for both farmers (occupation unit 11, social class 2) and agricultural labourers (occupation units 20-23, social class 4), so the class divide in rural areas is recorded very directly. Our method for estimating numbers in each social class in small towns and rural districts was therefore as follows: (i) We re-cast the data for districts in table 17 into the most detailed mutually exclusive groups possible, i.e. re-computing numbers in certain orders to exclude those more detailed categories which were separately tabulated; (ii) Then, using the more detailed data for counties, we computed precise numbers in each combination of a class and one of these mutually exclusive groups, but excluding from these county totals all persons in the separately tabulated large towns; (iii) We used the county-level proportions of each class in each group to estimate the numbers in each class within the group data for the particular district; (iv) Finally, we summed the number in each class in each group across all groups to obtain class totals for each district. The resulting class totals for each 1931 district were then re-districted to the modern local authorities.
• 1951: The original source is table 27 in the County Reports of the 1951 census, 'Social Class distribution of Occupied and Retired Males aged 15 and over'. The limitation of this table to males, and the fact that it does not sub-divide social class 3 into manual and non-manual, explain why we limit all our other social class data in the same way.
• 1971: The data are computed from table 23 of the 10% sample from the population census, via the Linking Censuses through Time system. The figures cover all economically active males. The source table provides data on 17 'Socio-Economic Groups'. Of these, SEGs 16 and 17 cover the armed forces and 'inadequately described occupations', which are excluded from the calculation of social classes. Eight of the remaining SEGs, covering 59% of the remaining men, fall wholly within a single class. Numbers in the other seven SEGs were assigned to classes on the assumption that each 199 LAU had the same 1971 class distribution within each SEG as the totals for Great Britain as a whole, which appear in table 37, 'Males by area of usual residence and socio-economic class', in the Economic Activity Report of the 1971 census. NB in all seven SEGs, over 50% nationally of each SEG fell in a single class, and in all but two SEGs over 79% fell in a single class. An exact calculation using the original 1971 report data will be possible later.
• 1981: The data are computed from table 52, 'Residents in Private Households, 10% sample: Social Class' in the 1981 Small Area Statistics, via the Linking Censuses through Time system. The data are for males only, aged 16 to 64. The figures for social class 3 combine data for manual and non-manual workers.
• 1991: The data are computed from table S91, 'Social class and economic position (10% sample): Residents aged 16 and over in households', in the 10% sample of households in the 1991 Small Area Statistics, via the Linking Censuses through Time system. The data used here cover all males aged 16 and over resident in households, regardless of 'economic position'. The data have been adjusted to allow for under-enumeration in 1991.
• 2001: The data are taken from table KS14 'Socio-economic Classification: Census 2001, Key Statistics for local Authorities', using data from KS14b for all males aged 16-74 in employment. We hope to use more detailed 2001 data once they are available, but for now we must work with the fairly simple classification used in the initial 2001 results. Class 1 combines 'large employers and higher managerial occupations' with 'higher professional occupations'. Class 2 combines 'lower managerial and professional occupations' and 'intermediate occupations'. Class 3 combines 'small employers and own account workers' and 'lower supervisory and technical occupations'. Class 4 consists of 'semi-routine occupations'. Class 5 consists of 'routine occupations'.