The Australian Bureau of Statistics (ABS) is embarking on a transformation program, which includes, amongst other things, re-engineering, using different collection modes for survey data, and using different, but more efficient, sampling frames and estimation methods for official statistics. Whilst this transformation is expected to bring about positive changes to official statistics, there is a risk that such changes could induce statistical impacts in some ABS time series, which could be misinterpreted as real world changes. The resulting challenge is to develop methodologies to monitor, measure and, where needed, adjust for impacts for any affected ABS time series.
In this research, a methodology to measure such statistical impacts in time series is proposed. To estimate the change that occurs in the target survey variable, the method uses related data series, which measure a similar concept to the target survey variable, but which are not subject to measurement change. Using this method, the statistical impact can be assessed by intervention analysis, taking advantage of the cross-correlations and leading properties between the target survey variable and the related series. We illustrate the power of this method by estimating Australian Labour Force Survey supplementary survey effects as an example.