How did MRP do in Australia?

Welcome to the 84th edition of The Week in Polls, which takes a trip to the other side of the world to see how MPR did in the last Australian federal election.

Then it’s a look at the latest voting intention polls followed by, for paid-for subscribers, 10 insights from the last week’s polling and analysis. (If you’re a free subscriber, sign up for a free trial here to see what you’re missing.)

As ever, if you have any feedback or questions prompted by what follows, or spotted some other recent polling you’d like to see covered, just hit reply. I personally read every response.

---

Been forwarded this email by someone else? Sign up to get your own copy here.

Want to know more about political polling? Get my book Polling UnPacked: the history, uses and abuses of political opinion polling.

Or for more news about the Lib Dems specifically, get the free monthly Liberal Democrat Newswire.

---

The Australian Parliament. Photo by Michael on Unsplash.

How did MRP do in Australia?

My deep dive into MRP polling - what it is, what it can tell us and why it may get the next general election wrong - has been one of the most popular editions of The Week in Polls so far.

That piece concentrated on evidence from the UK. But MRP has been used in other countries too. How has it fared there?

Professor Murray Goot has kindly shared with me his analysis of YouGov’s MRP from last year’s Australian federal (national) election. Australia uses what we would call the alternative vote for its lower house elections, and the 2022 election came against a backdrop of declining two-party dominance in their electoral system and a new challenge from a group of independents (the Teal wave). So it was a good contest in which to see whether MRP could do better than traditional national uniform swing projections from normal polls.

What did he find?

First, how did traditional polls plus uniform swing (aka ‘the pendulum’) perform?

If Labor was going to win 53 per cent of the two-party vote - as the final polls from Ipsos, Roy Morgan and Newspoll suggested - according to the pendulum, it was likely to finish with 79 seats… If Labor was going to win 51 per cent of the two-party vote - as the final polls from Essential and Resolve suggested - it was likely to finish with 72 seats…

Labor won 52.1 per cent of the two-party vote - roughly the midpoint of these two estimates. This should have given it an extra seven seats, taking its tally to 76 - one less than it achieved.

In other words, the national polls were very accurate on vote share, and using uniform national swing translated that into roughly the right number of seats too.

Those seat number projections though, 79 seats or 72 seats, were either side of the magic 76 number, the seat tally that Labor needed to rule on its own. So although the polls and pendulum combined to be close, different pollsters were pointing to different answers on whether Labor would get enough seats for its own majority, which it just did with 77.

So did MRP do better, and allow such a close election result to be more accurately foretold?

Not so much. Here’s what he found analysing the MRP figures:

The MRP poll’s estimate of 80 - the midpoint in its range - overestimated Labor’s majority by three seats.

That’s not a bad result compared with the reality of 77, and better than some traditional pollsters, but it’s worse than the average of the traditional pollsters.

Moreover the individual constituency projections - something of particular interest when there is a new electoral dynamic such as the Teal independents - were not great:

[The MRP’s] mean errors were large - larger than those registered by RedBridge[‘s individual constituency polls] and not very different from those recorded by single-seat robopolls in 2013, 2016 or 2019. The standard deviations were also substantial, with large overestimates figuring alongside large underestimates.

The constituency errors included failing to predict four of the six wins for the Teals, a significant failure to tell the electoral story more accurately than simple national polling could.

Although in terms of getting individual constituency results wrong, it’s worth noting that:

There were nine winners it failed to pick … [which was] two less than the 11 misses one would have expected had the MRP poll been no more accurate than the 2017 UK original.

Overall then, not a bad public debut at all for MRP. (There was an unpublished one at a previous Australian election.) But not one that immediately established it as a gold standard with superior insight to traditional polling and swings.

Something to bear in mind as the next UK election nears and the MRPs here ramp up.

YouGov also did an MRP for the most recent Spanish elections, which performed very creditably. Details here.

Watershed: The 2022 Australian Federal Election which contains Murray Goot’s full analysis is available from Waterstones and Amazon.

---

Know other people interested in political polling?

Refer friends to sign-up to The Week in Polls too and you can get up to 6 months of free subscription to the paid-for version of this newsletter.

---

National voting intention polls

Yes, once again it was a week without a poll putting the Conservatives on more than 30%, extending the run stretching back to late June (when a Savanta poll gave them 31%). The lowest of low benchmarks, the Duke of Wellington’s crushing defeat in 1832, continues to be better than the current Conservative Party’s standing.

Here are the latest figures from each currently active pollster:

For more details and updates through the week, see my daily updated table here and for all the historic figures, including Parliamentary by-election polls, see PollBase.

Last week’s edition

The Times shows how not to report an opinion poll

---

My privacy policy and related legal information is available here. Links to purchase books online are usually affiliate links which pay a commission for each sale. Please note that if you are subscribed to other email lists of mine, unsubscribing from this list will not automatically remove you from the other lists. If you wish to be removed from all lists, simply hit reply and let me know.

---

Over 1 in 10 think the Elgin Marbles are marbles, and other polling news

The following 10 findings from the most recent polls and analysis are for paying subscribers only, but you can sign up for a free trial to read them straight away.