It aggregates and averages all federal opinion polls asking primary voting intention to generate a two-party preferred result.
This averager includes all published polls that include a voting intention question. If I’ve missed a poll, please contact me.
Latest Polling Average
Updated 9 September 2014
|TPP 8 Sep||Swing||Seat +/-||TPP 25 Aug||TPP 11 Aug||TPP 28 Jul||TPP 8 Jul||TPP 30 Jun||TPP 2 Jun||TPP 20 May|
The average of the polls shows a two-party-preferred -5.73% swing against the Coalition, which equates to a -27 seat change from the 2013 election result. This is a change in the average TPP since the average of the polls in June which was -7.61% (and -36 seats) away from the Coalition. The addition of the most recent Morgan poll and Newspoll halts the slow improvement for the Coalition’s TPP. Ultimately the average of the polls reflects a continued and substantial collapse in the Coalition’s TPP vote since the 2013 election.
A majority of 76 seats is needed to form government. On the current polling, it is projected that Labor would have 82 seats (currently 55), and the Coalition 63 (currently 90).
Share this on social media
The most recent polls added to the Polling Average is the Morgan poll and Newspoll.
Share on social media
Get the latest updates on Twitter
Sign up to be notified when this averager is updated
Use the form below to sign up to receive an email for when this averager is updated. It will generally be updated within a day or two of a new poll being published.
(No nasty spam, I promise.)
About this poll averager
State polling: This poll averager does not consider state-level polls that ask federal voting intention, or seat-level polls. It only averages national-level opinion polls.
Newspoll: Unfortunately, Newspoll does not separately ask about the Palmer United Party in its primary vote questions. As a result, this impacts the two-party preferred result due to preference flows. E.g. PUP preferences went 53.6% to the ALP, compared to “Other” which was 50.2%. This means Newspoll slightly favours the Coalition in its results.
Morgan Poll: Morgan poll has an odd effect that slightly skews results in favour of the LNP. This is for two reasons. Firstly, like Newspoll, Morgan does not seem to ask about voting intention for the Palmer United Party; PUP is included in “Other” and thus this advantages the LNP. Secondly, Morgan has a large sample size (over 3000) which appears to be due to adding two weeks of polls (e.g. 19/20 Jul and 26/27 Jul). Because there’s no way to disaggregate these result, I have taken the large sample size at face value.
Greens, Independents & Palmer: For the purposes of seat level swings and allocation, this polling averager assumes that the Greens party retains Melbourne, Katter retains Kennedy, Wilkie retains Denison, McGowan retains Indi, and Palmer retains Fairfax.
Preference distribution: This averager uses preference flows from the 2013 election.
Primary vote sample sizes: The primary vote table shows the (known) sample size for recent polls. Where no sample size is known, I have used the sample size for the immediate preceding poll for that publisher.
About poll averaging
The poll results that are averaged are weighted based on sample size and recency. This means that a poll of 3000 people is weighted more than a poll of 800 people; it also means that a poll released yesterday is weighted more than a poll released three weeks ago. A poll that is over a month old has a 0% weighting.
Averaging polls is generally regarded as more accurate than individual polls. This is because it smooths out the individual variation that can occur with rogue polls. Additionally, because averaging many polls aggregates samples, the final two-party preferred average has a lower margin of error than a single poll.
Because pollsters ask voting intention questions differently (amongst other reasons), this poll averager is not scientific. For everyday use however, it should provide a much better picture of public opinion and its trend than a single poll could.