Crimes against Psephology: Christopher Pearson –you’re nicked.

The really dismal thing about elections, apart from being inundated with images of some of the most truly unphotogenic people in the country, would have to be the absolute fucktardery over polling that masquerades around as fact in opinion columns.

The latest cab off the rank is that savant psephologist, Christopher Pearson. On September 22, he wrote (if by wrote you mean ‘scrawled a random stream of consciousness from a parallel universe’) an article in The Australian with the humble title “To win the unwinnable poll”.

“Interesting”, I thought….. “Maybe a polemic on what the government needs to do”, I mused.

But alas, as one read through the piece it quickly became apparent that the article had much to be humble about.

The first line was a corker and really set the scene: “The Government is getting near the level of support it needs in the seats where it matters most.”

….which was just the first in a rather long line of WTF? moments that ensued.

We only have to look at the swings in the marginals identified in the quarterly Newspoll breakdown to clearly demonstrate this to be nothing more than an exercise in make believe.

The seats where the government needs support is in the marginals. The ALP have, according to the latest quarterly Newspoll breakdown, 50% of the primary vote in those marginals. That is a 9.2% primary vote swing to the ALP and an 8.3% two party preferred swing in those seats that are supposedly the ones that “matter most”, seats held by less than a 6% margin.

The ALP has an average of 2.3% more vote than it needs to take every single one of the least marginal of those marginal seats, let alone those seats with a margin well under 6%. For the government to be getting “near the level of support it needs in the seats where it matters most” , that swing to the ALP in the marginals needs to be halved – let alone the swing against the safe government seats likewise halving as well, simply to stop the government held seats on 6% – 7% buffers from falling.

After a bit of self indulgent navel gazing and chastising of other commentators for their apparent inability to understand reality, this electoral Man of Letters declared:

“It’s often forgotten that his victory in 1998 was achieved with a primary vote in the House of Representatives election of only 39.5 per cent and that Labor won in 1990 with a primary vote of 39.4 per cent. If the Coalition were to wage a dogged campaign concentrating on holding its marginal seats, it could win by maintaining its present primary vote if it also managed to cut Labor’s two-party preferred margin to about two points, as in 1998 when Labor led with 51 points to the Coalition’s 49 and still lost.”

What Christopher Pearson seems to have forgotten is that little thing called One Nation – just how one could forget One Nation is beyond me, but memory loss and delusion do seem to walk hand in hand in the political psychopathology stakes.

In 2007, a primary vote in the low forties will deliver the Coalition exactly nothing but the opposition benches, simply as a consequence of the minor party make up and vote share. The 1990 election quoted was memorable for the high 17.1% minor party primary vote – mostly the Democrats and the Greens, which forced 91 seats to be decided on preferences. Of those 91 seats, the Coalition gained 33, the ALP 57 and Others 1. That was simply a result of the ALP benefiting from a high preference flow from those minor parties – parties generally from the centre left, in an election where the environment was a dominant issue.

In the 1998 election, One Nation was the dominant minor party and being a party from the conservative side of politics, sent a majority of preferences back to the Coalition helping to push Howard over the line in a large number of seats. We can see the One Nation effect on the primary vote of the Coalition by simply graphing the primary vote swing of the monthly Newspoll aggregates (which is simply the difference between the Newspoll primary vote estimate and the primary vote obtained at the previous election).

The One Nation effect is marked and represents the period from the month when the One Nation party formed through to the 1998 election. The blue numbers at the top are the primary vote swing achieved at each election. If we do the same for the ALP primary vote swing we get:

There was little to no One Nation effect on the ALP primary vote. However, the ALP vote seemed to grow at the end of 1997, possibly as a reaction to the Coalitions handling of the One Nation saga. From the primary vote swings we can clearly see that the Coalition primary vote dropped substantially (-7.75%) compared to the small ALP rise (+1.34%). In 1998, the Coalition could win government with a small primary vote simply because the preference flows from One Nation were benefiting the Coalition compared to the ALP – leading to the Coalition winning 62 seats on preferences vs. the 35 seats the ALP won on preferences.We’ve modeled the One Nation effect many times before, and to readers here it is nothing new.

Unfortunately for Pearsons fantasies, preferences are currently flowing between 65-75% to the ALP if we look at the ACNielson and Morgan preference allocation distributions.

So no Christopher – the Coalition cannot win with a primary vote in the low forties in 2007 simply because of the lack of minor party support from the right. The ALP can win with a low primary because of the high minor party preference flow to them that is a function of the political composition of the minor party vote, but the Coalition simply cannot – so let us have no more of that horseshit eh?

The next piece of ignorance to emanate from the pages was this gem:

“In Western Australia, the two-party vote has moved from a 50-50 split to 51-49 in the Coalition’s favour, which would help deliver the Government Labor’s two marginal seats in Perth. In Queensland the two-party split moved from 54-46 in Labor’s favour to 52-48.”

If we use WA as an example to show why this is nothing but buffoonery of the most inane kind, in WA during the period from Quarter 2 to Quarter 3 2007, the governments primary vote swing has gone from -5.8% to -4.8%, the ALP primary vote swing has stayed the same at +5.3% and the ALP TPP swing has reduced from +5.4% to +4.4%.This would deliver the ALP the two Coalition marginals of Stirling and Hasluck on buffers of 2.4% and 2.6% respectively according to Antony Greens spiffy election calculator. For the government to gain the two ALP marginals, they need a swing TO them, not a swing AWAY from them. On current standing, the Coalition needs a 4.5% swing to them in WA between now and the election to pick up Swan, and a 5.2% swing to them between now and the election to pick up Cowan.

You see Christopher, it’s not about the 50/50 split on TPP that makes the difference, it’s the SWING that matters. In Qld, to give another example to beat the stupidity out of you with, a 52-48 split would represent a 9.1% TPP swing to the ALP. If this was uniform, the ALP would pick up 10 seats in Qld.

Yes, 10. From smallest to largest winning margin they would be Bowman, Dickson, Hinkler, Flynn, Petrie, Longman, Herbert, Blair, Moreton and Bonner.

So please, let us put this type of gross misunderstanding of polling results to bed as well.

Pearson continues:

“On these figures, a recapitulation of the Coalition’s victory in 1998 is quite on the cards.”

Only if you’re smoking crack Christopher.

Now remember folks, this article was printed in The Australian, the very newspaper that thundered in a self-indulgent hissyfit that “we understand Newspoll because we own it“. They declared that “THE measure of good journalism is objectivity and a fearless regard for truth“, before stating that “Not properly understanding how polls work gives our critics licence to project their own bias onto analysis of our reporting.”

Allowing columnists such as Christopher Pearson to produce articles like “To win the unwinnable poll”, makes a complete mockery of the standards of “objectivity and a fearless regard for truth” that the paper declares is so important to the measure of good journalism. Publishing such articles that, by any objective measure, fall victim to “Not properly understanding how polls work” undermines The Australians self declared superiority of polling analysis to the point of it verging sharply toward hypocrisy.

One doesn’t need to own Newspoll to understand the Yooniverse, but one certainly needs to have a modicum of understanding for the very basics of statistics and electoral history, to produce polling analysis that can be said to have even a mediocre relationship with observable reality. On this point, Christopher Pearson and consequently The Australian, fail.

If The Australian wishes for its political analysis to be seen as national best practice, then it needs to uphold a higher quality of journalistic standard and vigorously enforce stronger quality control over its published content. For as long as articles like Christopher Pearson’s “To win the unwinnable poll” are published by The Australian, the self-declaration on the superior quality of The Australians polling analysis will ring hollow.