Tuesday, January 26, 2016

Is Morgan cooling?

The three most recent Morgan poll results suggest a cooling in voting intention for the Coalition. Taking preferences distributed as they were in the 2013 election; Morgan has the Coalition's two-party preferred (TPP) voting intention as follows

  • 56 per cent for the period December 5/6 and 12/13, 2015
  • 55.5 per cent for the period January 2/3 and 9/10, 2016; and
  • 54 per cent for the period January 16/17 and 23/24, 2016

But I am a little cautious of the final result. It includes the weekend near Australia day. Because many households use the Australia day weekend as a long weekend, this can affect the comparability of polling results. And to highlight my caution, as we will see below, some of the primary vote results suggest significant movements in voting intention. It will be interesting to see how Morgan behaves in its next fortnightly print.

Notwithstanding this caution, adding these numbers to the aggregation, we get an estimated two-party preferred result for the Coalition of 55 per cent

Turning to the primary vote aggregations, we can see the more surprising statistics in the most recent Morgan print. Morgan had the Coalition's primary vote well down on recent polls (from 47 to 43.5). On the up-side we have the Greens (from 13 to 15) and the other parties (from 11 to 13.5). The Morgan symbol on these charts is the upwards pointing triangle.

Note: I have upgraded the analytical engine to JAGS 4.1.0. A description of the aggregation models can be found here.


  1. Mark, thank you:

    If you compare the Morgan for November with the latest Morgan, you find that more labor ( + 0. 7 ) and greens ( + 2.7 ) voters prefer Shorten as PM in this poll than in the November poll. There is also a rise of 1.8% in coalition voters supporting Turnbull.

    That being the case, how does Morgan come up with a PPM for Shorten dropping from 29 to 19. Surely, it should be at least 25.

    Can you think of an explanation for that?

    Thank you,


  2. Methinks FRED means ReachTEL. I have raised this issue in detail after first spotting it late last week here: http://kevinbonham.blogspot.com.au/2016/01/poll-roundup-shortens-latest-shocker-or.html

  3. FRED - I think Kevin is right, it's most likely either a data entry error (a typo) or a calculation error.