The measurement of urban poverty: a missing dimension

Abstract

destitution? Departmental archives reveal that the Board of Trade was acutely aware of the demand from social scientists and reformers for national and comparative regional data on the extent of poverty and 'urban degeneration'.2 Moreover, many of its leading statisticians were familiar with contemporary developments in the theory of sampling and with the application of probability theory to social investigation. Representatives of the department regularly attended the International Statistical Institute, the British Association, and the Royal Statistical Society, where the value of such data to the measurement of social and economic phenomena was discussed. Several of the Board's investigators were introduced to the mathematical principles and practical application of sampling at the statistical classes run by Bowley at the L.S.E., and Bowley himself was retained as a technical adviser to the Board of Trade during many of its most significant enquiries after I900.3 Nonetheless, prior to I914, no social investigation undertaken by the Board fully incorporated random sampling techniques. Several explanations can be advanced for this shortfall in the technical sophistication of civil intelligence. Firstly, a range of practical constraints operated upon Whitehall's statistical output. Weaknesses in the data produced by other departments of social administration and by the census inhibited the ability of the Board of Trade in obtaining information of sufficient quality and comparability for the application of more sophisticated quantitative techniques.4 Similarly, Treasury control militated against new methods. They were resisted by the Treasury as being cost-ineffective and as creating an additional field of expertise outwith the normal recruitment procedures of open competition.' In the case of social statistics, the normal bureaucratic constraints of Treasury control were reinforced by more personal, ideological