Anomalies of collection : a study of the validity and value of ethnic data : a thesis presented in partial fulfilment of the requirements for the degree of Masters of Arts in Social Policy at Massey University

Abstract

From the outset the task was to investigate the validity of ethnic mortality data collection and suggest ways it could he improved. This began with a concern about an anomaly between the hospitalisation rates for young Maori involved in road accidents and the reported official road mortality rates for these same young Maori. The hospitalisation figures suggested Maori were over-represented and the mortality rates suggested Maori were under-represented. Using a sample of those who died on the road in 1990, the ethnic data collected by police was compared with ethnic data collected by funeral directors, the responses of next-of-kin and the ascribed ethnicity taken from a search of death notices. Major differences of ethnic classification were revealed, particularly between next-of-kin and police and funeral directors, raising serious questions about the validity of the two official sets of data. An assessment of the validity of the two sets of data was made by linking an historical analysis of the concepts and definitions of race and ethnicity with a qualitative study of the ethnic data collection practices of police and funeral directors. The final ethnic data sets were revealed to be a construction of the 'commonsense' conceptual understandings of race and ethnicity held by police and funeral directors being applied to the practices of collection and classification. These concepts, understandings and practices contained a variable mixture of historical paradigms and discourses. On the basis of validity as it is understood by statisticians, both official collections could reasonably be judged invalid. However, the deeper issue underlying this finding concerns the value of ethnic statistics. Their value lies in their use for justifying resource allocation and distribution between different ethnic groups. Ethnic data and its collection cannot be viewed outside the politics of resource allocation and the project of improving ethnic data is firmly located in this context.