Landline telephone surveys are suffering from increasing levels of non-coverage, with an increasing number of households dropping their landline connections and becoming “mobile-only” households. This mobile-only population differs demographically from the population that is reachable by landline telephones, which leads to significant non-coverage bias for telephone surveys that rely only on landlines. A dual frame survey design which uses a landline sampling frame and a mobile sampling frame can be used to reduce this non-coverage bias. However, a dual frame design introduces a number of new complexities to the weighting of the survey. Firstly, there is no consensus on which of the dual frame weighting strategies proposed in the literature is optimal. Secondly, it is not known when poststratification should be performed during the weighting process to obtain the most representative estimates of the population. In real life the true population totals are not known; using a simulated population where the population totals are known will allow the different estimators and poststratification methods to be evaluated accurately. Data was simulated for telephone usage (landline only, dual user and mobile only), demographic variables used for poststratification (gender, age, geographic region and education), and various outcomes. This population was then sampled from using the general sampling method for dual frame telephone surveys in Australia, with generated patterns of non-response to mimic a real life sampling process. This allows the bias of the various estimators and poststratification methods to be compared. This paper evaluates preliminary results from this simulation study which examines the optimal weighting strategy and the optimal order of poststratification for dual frame surveys in Australia.