Nice Newspoll, Shame about the Rates.
Posted by Possum Comitatus on November 6, 2007
This was me in Crikey today HERE. It’s free so you can have a squiz over there or keep reading.
When your luck is up, you just can’t take a trick.
Today we have the best Newspoll for the government since November, even if it’s only due to a bit of minor party preference noise, but as far as the image goes – a good poll is a good poll is a good poll.
Ordinarily, the popular media outlets would have a couple of stories splashed around hailing a comeback, the not so popular media outlets would probably be going hysterical and the nightly news would lead with the story; Laurie Oakes telling Nine viewers in tones of gravitas that the election is now a competition.
But this is no ordinary week.
Today’s papers are all about the ponies, tonight’s news will be about them as well, Wednesday’s papers will be all about the ponies and interest rates, Wednesday night’s news will probably be all about an interest rate rise, Thursdays papers will certainly be – and then the media attention and narrative turns to the first polls after the rate rise (assuming there is one).
Where does the best poll for the government in 12 months fit into that cycle? Well, it doesn’t. Not as far as normal people are concerned. But the particularly nasty piece of bad luck in this sequence of unfortunate events for the government is what is likely to come next.
The headline two-party preferred result of Newspoll has been bouncing around an awful lot lately – it’s almost become the great oscillator.
This week’s Newspoll figures have the problem of slightly undervaluing the preference flows the ALP receives from the minor parties, meaning that it’s more likely than not that the next Newspoll will probably fix that up. These slight rounding problems and sampling volatility of the minor parties all come out in the wash over a few polls.
When you combine that with the ALP primary looking rock solid at 47/48, it’s almost expected that in the next poll or two, the two party preferred headline figure will show the ALP increasing its lead – simply as a result of the high ALP primary vote combining with this minor party sampling error and rounding issues.
But should that happen, the headlines will undoubtedly scream “Interest Rate Backlash!” as some new 55/45 poll shows the ALP gaining a two point lead from the previous poll, the best poll the government had enjoyed for 12 months, but one which no-one paid attention to because the ponies were on.
Somehow, the gods look to have conspired against the government to the point where their best Newspoll result in a year will likely be completely ignored in terms of any beneficial media coverage that matters, but instead will create the platform for a very dominant media narrative that launches against them if rates rise tomorrow and the next Newspoll moves toward the ALP simply as a result of polling noise. And with it goes the second last week of the campaign.
Howard must be pulling out with absolute frustration those few remaining hairs he has left on his noggin.
He just cannot take a trick.
Possums Pollytics 
Doug said
Possum
Following comment on your discussion of the Newspoll by simon Jackman – has he got your argument right/
sampling error is random
Tuesday November 6, 2007
Filed under: Australian Politics, statistics — jackman @ 4:50 pm
The latest Newspoll has Labor on 47% of first preferences (down 1), the Greens on 5% (up 1), the Coalition on 42% and Others on 6% (both unchanged). Two-party preferred that resulted in 53-47, a one point change from the last Newspoll 54-46.
The devil there is buried in the rounding error. If the Greens go up 1 and Labor down 1, and 80% of preferences flow to Labor, then the net impact on Labor is about a .2 of a percentage point drop in ALP 2PP support. Toss in rounding to percentage points, and you get the 1 point drop in 2PP we saw in the poll.
Possum Comitatus said as much in a Crlkey post today. But Possum then said (my emphasis)
This week’s Newspoll figures have the problem of slightly undervaluing the preference flows the ALP receives from the minor parties, meaning that it’s more likely than not that the next Newspoll will probably fix that up. These slight rounding problems and sampling volatility of the minor parties all come out in the wash over a few polls. When you combine that with the ALP primary looking rock solid at 47/48, it’s almost expected that in the next poll or two, the two party preferred headline figure will show the ALP increasing its lead – simply as a result of the high ALP primary vote combining with this minor party sampling error and rounding issues.
Huh? How does that work? Lets make the big assumption that support for all parties is constant over time. Suppose that the vagaries of random sampling mean that I overestimate Green support this time. That doesn’t mean that in my next poll I am therefore more likely to underestimate Green support, the random fluctuations automatically canceling each other out over two or three polls.
Over the long run, yes, sampling errors around a stable parameter cancel out. But thats not what we’ve got here. Under random sampling, sampling errors are independent (both within and across surveys), meaning knowledge of today’s sampling error gives me no information about future (or past) sampling errors. Its like a roulette wheel in a casino; seeing the series “red, red, red” gives me no reason to think that “black” is any more likely on the next spin of the wheel (or at least it shouldn’t if the wheel is fair).
So I don’t quite see how Possum gets to argue that a bump down in ALP primary support will be followed by a bump up next time. Indeed, with the possibility of trends in the population, who knows…?
Enemy Combatant said
“He just cannot take a trick.”
Poor El Rodente tried so hard to be bold and true. He led with a $34 Billion trump, only to be finessed by fate, poor message control, and the Curse Of Crosby-Textor.
He’s never had the edge since Pixie came and Arthur went.
Bruce said
Hi Doug,
“Under random sampling, sampling errors are independent…, meaning knowledge of today’s sampling error gives me no information about future…sampling errors. Its like a roulette wheel in a casino; seeing the series “red, red, red” gives me no reason to think that “black” is any more likely on the next spin of the wheel (or at least it shouldn’t if the wheel is fair).”
For the sake of argument let’s assume that the population support has not shifted – say the 2PP is 55%. Then a sample that gives an estimate of 53% is near the lower end of the what is expected from a random sample. True it’s possible it could be 52%, and an outside chance it could be even lower.
Now consider the next sample. If we KNEW the population 2PP to be 55% our expectation would be the most probably sample result will be 55%, but it could also fall either side of 55% in an approximately Normal Distribution. However it is will most probably be more than 53%. This phenomenon is known as regression to the mean. More often (but not always) samples in the extremes of a distribution will be followed by samples closer to mean or on the other side of it.
To take your roulette example. Say you got a 3. While it’s possible the ball will fall on a 0, 1, 2 or 3 next time , it is much more likely it will fall somewhere between 4 and 36. There has been no loss of independence, it’s just when you are down the bottom of the distribution, the only way is up. When you are almost at the bottom you will probably go up.
This does assume we are looking at an underlying population 2PP that is stable, buty Possum has presented a strong argument on this based on the Labor’s primary preference.
Bruce
Rod said
I’ll make the same point here that I’ve made elsewhere about this poll (so apologies if you’ve heard it before). Of course all it probably really shows is variation due to minor sampling differences and the results fall so far within the margin for error that it is hard to take them too seriously one way or another, but if we pretend the changes really matter:
1) The Liberal vote is down 1%
2) The Nationals vote is up 1%
3) The Labor vote is down 1%
4) The Green vote is up 1%
What does this mean?
The Labor vote has almost certainly simply shifted across to the Greens , and will come back in preferences. It is accordingly of little or no consequence.
The DECLINE in the Liberal vote is a matter of some consequence if it is real. 30 of the 32 most marginal coalition seats are held by the Libs. Libs losing primary vote is bad news for them at this time. One percent decline could mean a swag of additional seats lost.
The INCREASE in the Nats vote is pretty meaningless. It probably simply involves people shifting from Family First to the Nats in rural areas after the Family First porno problems. In some cases it may mean the loss of Liberal or Labour votes to the Nats in safe Nat seats.
The overall impact, in terms of seats , is probably neutral for Labor, and negative for the Libs, given their declining primary vote.
Looked at broadly, if anything it may mean a reduction in the national 2PP vote needed for Labor to win government (as gains by the Nats in safe Nat seats inflates the coalition vote, but doesn’t deliver them any real electoral results.)
My own view is that this is all sampling error/ variation stuff, but if it is not, it is actually BAD news for the coalition, not good news. It looks like they are going backwards where it matters and scoring their “wins” in places that are already theirs.
Cheers
Rod
Possum Comitatus said
Just a quick reply doug as I have to run, I’ll repeat what I said to Simon.
It’s about the nature of the 2004 preference flow that Newspoll uses.
This poll, the ALP received the lowest preference flow level they have in any Newspoll since the 2004 election, at about 54.5%, and this was brought about by what must have been quite an unusual confluence of numbers to be rounded. So unusual, that if the ALP stays on 47 next poll, and the Coalition actually increases their primary by 1 point – the TPP remains the same according to the Newspoll formula.
The last time the “others” had a score of 11, the prefs were divided up 63.6% to the ALP, which is closer to reality according to both the published polls that ask for second preferences and party polling.So this distribution is either the result of:
(a) a grand Newspoll conspiracy going on that I dont know about
or
(b) a very very rare confluence of decimal points that provided this result, the likes of which we havent seen in 75 polls and probably wont again.
If the true value of public political sentiment remained the same until the next poll, the most likely outcome in the poll wouldnt be the 54.5% pref distribution, it would be the 63.6% distribution if they ended up with the same broad numbers.
Andos the Great said
Hey Possum,
Did you see Alan Kohler on ABC News tonight? He had a graph of the ratio of RBA cash interest rate to the OECD average over time since 1987. The average ratio worked out at 1.2 for both Labor and Liberal.
Looks like a pretty good analysis to me. What do you reckon?
Greeensborough Growler said
Can’t take a trick………
Kevin Rudd’s original Melbourne Cup pick is scratched.
John Howard says Rudd should Me Too on his pick Mahler.
Kevin Rudd’s new pick Efficient wins at around 25-1.
Moral of the story is when you are in a hole Johnny and you want to get out, stop digging.
Down and Out of Sài Gòn said
Quick question. Why don’t the bloody pollers go out and ask for preference intentions (as well as primary intentions)?
kiwipundit said
Rod said @4
“My own view is that this is all sampling error/ variation stuff, but if it is not, it is actually BAD news for the coalition, not good news. It looks like they are going backwards where it matters and scoring their “wins” in places that are already theirs.”
The last point you make could perhaps explain why the odds for the Coalition winning the election haven’t improved much at all since the details of the Galaxy were released last night – Centrebet for instance has Labor on $1.35 and the Coalition still out on $3.20. Of course, Labor’s 2PP lead of 53-47 in the latest Newspoll is around what they got in the Hawke landslide of 1983.
kiwipundit said
Oops,
I meant “the Newspoll” not “the Galaxy”
El Nino said
Rod, I think that the betting markets develop a bit of inertia as more money is punted. Unless something dramatic happens, I would be surprised if there is much more movement there before the election. The interesting odds now are on individual seats where maybe not too much money has been punted so far.
otiose said
GG at #7 – biggest irony is that ljh’s pick (mahler) is pjk’s fav composer 😉 asnd efficient best describes rudd’s campaign so far
Alan H said
My favourite composer too!
cheers,
Alan H
HarryH said
This suspect 53-47 poll off of 47-42 primaries with a drop of 1% from Labor to Green and a drop of 1% off Lib to Nats has been produced for a specific reason.
it is all about protecting workchoices’ reputation. this dubious result was produced to appear as though the workchoices party was on the way back and competitive until interest the rate rise went Kabam and knocked them out of the water.
the Murdoch narrative of the upcoming loss will be all about interest rates and protecting workchoices. Big business got out of Rudd what they wanted with workchoices lite.
Now the people can have their pyrrhic victory….but they will relentlessly be told that workchoices had little to do with it.
In reality , we can thank the Your Rights at Work campaign for the end of the Howard nightmare.
Stig said
Down & out @ #8: The quick answer is that they do, at least the times that I’ve been polled (Galaxy rather than Newspoll).
So, why don’t they distribute them that way? Maybe they understand their polls better than us, because they own them 🙂
Suspected Illegal Entry Vessel XI said
JWH might have tipped Mahler, but his soundtrack at the moment is distinctly Wagnerian. Sunshine Rudd, on the other hand, can’t do a lot wrong. Did he give his winnings to charity?
Norman said
Wasn’t Mahler Keating’s favourite? And we know where he ended up. Poor old Rodent.
Stephen Tardrew said
What do you guys make of the whole thing about pols showing that the Rodents mob are not responsible for interest rates. Is there a strong case for this argument if not what are the confounding factors in the the polling.
steve_e said
After the LIBS initially talking down a need for Interest Rates to increase, there is now an ambrace of the inevitable and a bold push to engage a fear campaign about the future (ecomoc uncertainty – fear of change, etc).
Where is the uncertainty? The change in Government is not uncertain. Increases in Interest Rates are not uncertain. The US is deeply in debt is not uncertain (debts are either repaid or repudiated).
What is clear is that there is no DIRECT connect between who leads the Government and Interest Rates. There is an indirect impact on inflation (wages, economic consumption stimulation – tax cuts, boosts to welfare payments, family tax benefits, etc). There is a longer term impact on inflation as a result of any shortage of skilled labour resulting in higher wages and a need to import replacement labour. There is a need for a comprehensive infrastructure and logistics program to move the population as well as goods and servcies around the country (this is part of how we export what is grown and dug up).
Lovie Scavuzzo said
Thanks For Sharing
baccarat online terpercaya said
situs game
Nice Newspoll, Shame about the Rates.