With Newspoll having released its quarterly polling breakdowns, we can now combine them with the Nielsen data for the last three months and  run our election simulation – giving us the most likely result were an election held over the Jan-Mar period and the results matched the polling.


While the headline results showed a sharp move towards the Coalition  – clawing back some 12 seats since the final quarter of 2009 – the swing is not only still towards the government, but our phone poll trend has moved further away from the Coalition over the last few weeks.

On the distribution results of the simulation, the simulated ALP seats in Parliament probabilities and distributions came in like this (click to expand):

simfreqmar10 simcummulativemar10

The first chart gives is the spread of the simulation results, the second chart the cumulative probabilities. To use the second chart, simply choose the number of seats the ALP would win from the bottom axis, trace it vertically until it intersects the red line, then trace horizontally across to the left axis – this tells you the probability of the Labor Party winning at least this many seats were an election held during the last three months.

For instance, there was an approximately 20% probability of the ALP winning at least 93 seats were an election held over the last 3 months, and where the results came within the margin or error of the polls.