14
votes
Prime Minister Scott Morrison has defied years of opinion polls and public expectation to lead the Coalition to a shock election victory
Link information
This data is scraped automatically and may be incorrect.
- Title
- 'I have always believed in miracles': Scott Morrison declares election victory
- Published
- May 18 2019
- Word count
- 846 words
Wow.
I didn't expect this outcome, that's for sure. This was Labour's election to lose, and somehow they did.
Anyone have good theories on what happened before everyone and their uncle's finished their post-mortems of the campaign?
I don't know. I just heard that 56 Newspolls in a row (more than 2 years of polling) had Labor in the lead. This was supposed to be a shoo-in. But here we are.
A lot of people will come up with guesses, and many will pretend that their guess isn't really a guess, but I think noone actually knows what happened. Even Senator Arthur Sinodinos, that very experienced Liberal strategist, seemed surprised by the outcome.
But, if we're going to guess...
My guess is that, when push comes to shove, Australians are conservative. Not capital-C Conservative, but small-c conservative. They just don't like change. "If it ain't broke, don't fix it." is a very Aussie thing to say (even if it didn't come from Australia). And, despite the chaos in the Coalition, too many Aussies got scared by the agenda for change that Labor presented them with.
There's also the hip-pocket factor. Some of Labor's proposed policies were going to hit certain groups quite hard in their hip pockets, and those people would have voted against Labor to protect their wallets/purses.
there also seems to have been a late break against labor in the 2-party-preferred polling which played a part and, as anybody whose job it is to interpret polls will tell you: margin of error is a thing for a reason. 51-49 is as good as 'too close to call' in all but the largest polls and could go either way. so it was still labor's election to lose, but it looks like it narrowed to within what amounts to a coin flip by election day, and they just happened to be on the wrong end of said coin flip, probably for some of the reasons you listed.
For most of the past 2 years, the difference in polling between the Coalition and Labor was larger than the margin of error. In fact, Newspoll (which is considered to be extremely reliable, and has been cited by previous Prime Ministers as a reason to depose other previous Prime Ministers) actually showed an increase in Labor support just the day before the election.
A Labor victory at this election was widely believed to be a sure thing by everyone concerned - even the Liberal and National parties (which is why Scott Morrison delayed the election as long as he could, until almost the last possible day he was allowed to). The Liberal party Senator on the ABC's discussion panel tonight was just as surprised by this outcome as everyone else.
We've had more than 2 years of polling tell us that Labor would win this election.
And then they didn't.
keep in mind that margin of error is in both directions, though. 51.5% to 48.% might seem to have been a decisive advantage for labor, but all that means with respect to 2PP is that on a 3% margin of error (which i'm guessing is what newspoll's MoE is based on their sample sizes and it being pretty standard in polling), the final outcome should be between 48.5% to 54.5% for labor and 45.5% to 51.5% for the coalition if the pollster got it right. by all accounts, the result is going to be to one end of that range but still comfortably within it (and it would actually still be that way even if you gave labor an additional half-point to make it 52-48 as at least one other pollster had it).
so really, this looks to me like another case of the polling picking up on a very late swing that the punditry did not, leading to everybody being blindsided by an outcome which actually had a pretty reasonable chance of occurring at the end of the day.
But, surely, if the voting public's intention was to vote for the Coalition more than for Labor, this should have shown up in at least one poll in the past few years. Every poll has consistently put Labor in front, margin of error or not. That's why it seemed such a sure thing: because no poll for the past 2 years ever showed anything different. It was only the degree of the expected win which seemed to vary (whether Labor was going to romp it in, or whether it was going to be a close call).
That means that multiple polls by multiple polling agencies and political parties, using different polling methods, all had exactly the same error in them. I'm not buying it.
nope! or else florida would have governor andrew gillum right now, since he lead in literally every significant and good poll in that race before losing by 0.4% on election day. voters are not necessarily rational actors, and polling is not and cannot be down to the dot perfect at its projections, which is why margins of error exist. a winning 51 or 52% on a 3% margin of error either way is not distinguishable from a losing 48 or 49% in the 2PP for a reason, and this is why. there's also a reason psephologists like nate silver and sean trende are responding like this:
I don't know a ton about Australian polls but from a cursory glance, most polls correctly showed the conservatives as the largest individual party and the 2-party vote as being within a couple points. Doesn't seem like that big an upset?
IDK, if I were looking at this chart I'd see a hard break toward conservatives in the closing weeks. I'd say Labor was the favorite based on this trend, but certainly not the prohibitive one.
and not flipping out about what an unexpected, crazy, improbable result it is. it was a possibility--an unlikely one--that happened to happen this time around.
they had the same error because they all had this as a reasonable possibility, which then happened because it was a reasonable if unlikely possibility due to margins of error. like i mentioned above, a winning 51 or 52% on a 3% margin of error either way is not distinguishable from a losing 48 or 49% here.
even if you grant that this is an error, though (which it's really not, this fell in the acceptable range of outcomes projected by every final poll), the final polling error across the board is going to be something like, 2.5%? which is actually very good for polling. in the US for comparison, the average polling error in presidential elections is 4%, in senate races it's 5.4%, in house races it's 6.2%, and the average of polling errors at all levels where polling is conducted is about 5.9%.
Okay. Now I know what it feels like for someone else when I go Full Pedant on them. Thanks for that.
I need to start by pointing out that the two psephologists you've cited are not Australian. They're not here on the ground, living here every day, reading our news and listening to our broadcasters. One of them even says "I don't know a ton about Australian polls". They're just looking at a graph from the other side of the world, and with the benefit of hindsight. Where are their tweets from last week, predicting a resounding Coalition victory?
I am here, and I have been following this election quite closely (because I'm a politics nerd from way back), and everyone here in Australia was saying that Labor was headed for a win, and the only uncertainty was whether it would be a close victory or a total thumping. EDIT: Just to demonstrate how widespread this belief in a Labor win was, a high-profile betting agency paid out $1.3 million dollars in winnings to everyone who bet on Labor... last Thursday.
In that context, it's only natural that a lot of people were taken by surprise when the election went the other way.