Reliability and validity

In the absence of direct prevalence measures of alcohol and drug use in the community, indirect measurement is common. This is achieved by using data pertaining to alcohol and drug-related harms. These data tend to be derived from administrative data sets collected by agencies involved in the provision of some kind of service such as hospitalisations or alcohol and drug treatment or registers such as deaths. The key issue here is that these data were collected for administrative or operational purposes of the organisation and not primarily designed for research. Consequently when using these data for secondary analysis there are unavoidable deficits such as incomplete data or inadequate coding. For example, the Victorian police data captures information on reported assaults in the community; however this data is inadequately coded for our purposes regarding alcohol, warranting the application of a surrogate measure. That is a proxy measure of alcoholrelated assault is required. Time of day is used as an estimate of alcohol involvement.

 << Previous Page       Return to Index     Next Page >>