This study provides a way of calculating confidence intervals for a spatial microsimulation model using estimates of statistical validity calculated from Census data. The calculation of confidence intervals has so far been difficult using spatial microsimulation models, but this paper provides a technique that estimates confidence intervals based on the z statistic published previously. This z statistic uses small area Census data and results from the spatial microsimulation model. An application of the method to estimates of housing stress shows that about 95.7% of SLAs show statistically accurate small area estimates with reasonably narrow confidence intervals.
|Title of host publication||IMA|
|Subtitle of host publication||5th Proceedings|
|Place of Publication||Luxembourg|
|Publisher||Luxembourg Institute of Socio-Economic Research (LISER)|
|Number of pages||17|
|Publication status||Published - 2015|
|Event||World Congress of the International Microsimulation Association (IMA) - Luxembourg|
Duration: 02 Sep 2015 → 04 Sep 2015
|Conference||World Congress of the International Microsimulation Association (IMA)|
|Period||02/09/15 → 04/09/15|
Rahman, A. (2015). Calculating Confidence Intervals for a Spatial Microsimulation Model Using a Z-statistic Error. In IMA: 5th Proceedings (pp. 1-17). Luxembourg Institute of Socio-Economic Research (LISER).