Hard News: Gower Speaks
206 Responses
First ←Older Page 1 … 3 4 5 6 7 … 9 Newer→ Last
-
And here it goes
https://github.com/thoughtfulbloke/OneVoteTwoVoteRedVoteBlueVote -
George Darroch, in reply to
When someone says high, I always want to know how high.
I think that's called poll jumping.
-
Peter Green, in reply to
Does that need a library(gtools) at the top?
-
David Hood, in reply to
I was actually adding some code for loading gtools while you were writing that :)
-
Interesting. In today’s Horizon Poll, two thirds of New Zealanders expect National to lead a potential coalition government post-election. A plurality of New Zealanders want Labour to lead a coalition government post-election.
"The survey starkly illustrates the difference between perceptions over which party might win and what electors would prefer, Horizon’s Manager, Grant McInman says."
There are a number of ways to explain this discrepancy.
-
Interestingly, support for a Labour-led government is extremely high among 2011 Green voters, with 89.4% supporting the prospect. This compares to 93.1% of 2011 Labour voters.
The other figures show that National supporters are very confident of their side winning, Labour and Greens supporters are much less so. Bolstering those perceptions (filtered through the dozen or so people given license to shape their opinions) is essential for the left. Reinforcing the inevitability of a National win is its job.
-
Rob Stowell, in reply to
In today’s Horizon Poll, two thirds of New Zealanders expect National to lead a potential coalition government post-election. A plurality of New Zealanders want Labour to lead a coalition government post-election.
Fascinating, very understandable (most people I talk to are dubious about Labour/Greens getting over the line), and infuriating all at once :)
From a left perspective, vital the narrative changes from 'National leading' to 'election on a knife-edge'. Since that's a more realistic reading of what we know of the polls, it shouldn't be hard. (Some in National will be wanting the same story, to energise and focus their own vote.)
But there's some powerful voodoo against it. -
danduran, in reply to
This might be a dumb question, but how does one run this at home? I guess I have to install a language or something?
-
George Darroch, in reply to
This might be a dumb question, but how does one run this at home? I guess I have to install a language or something?
Go to http://www.r-project.org/ and install R. This is pretty easy, and can be done in about ten to fifteen minutes without specialist knowledge. I'd also go to https://www.rstudio.com/ and install R-Studio, because it makes this work nicer and easier if you're going to do more of it, but it's not at all necessary.
Then it's simply a matter of cutting and pasting the code at https://github.com/thoughtfulbloke/OneVoteTwoVoteRedVoteBlueVote/blob/master/oneVoteTwoVoteRedVoteBlueVote.R into your R session and pressing enter. Because it is the first time you're using it, it will require the installation of packages, but this is fairly automatic. Just choose the Auckland server and they'll install directly.
-
David Hood, in reply to
how does one run this at home
You do indeed need to install something- the language R, from
http://cran.r-project.org
Where in theory you could open this as a source file and run it. However if dead keen to give it a go, your life would probably be much easier if after installing R you install RStudio as well
https://www.rstudio.com/ide/download/
Because then it is a little more obvious that if you use the File menu and Open the OneVoteTwoVoteRedVoteBlueVote.R file, you get a source button in the top right of the window with the script in it (the window with the script is the upper left one) and can run the script by clicking the script button.(On the other hand, George might have pretty much covered this one while I was writing my response)
-
Snap.
Quite happy to help if for any reason things are difficult (95% of the time it's as easy as pie, but sometimes it isn't).
-
Russell Brown, in reply to
Just to note that I'm very keen to host and promote anything that emerges from these exercises.
-
Bart Janssen, in reply to
There are a number of ways to explain this discrepancy.
Most New Zealanders want Labour to form a coalition government
But most New Zealanders expect Labour to do something so utterly compellingly stupid before the election that they will fail to win
-
David Hood, in reply to
Just to note that I’m very keen to host and promote anything that emerges from these exercises
Well, we can probably get it making pretty graphs of the result for inclusion in posts about election things. Beyond that I can't think of much as it is really just reinterpreting the figures collected by the polling agencies (though will keep that in mind).
-
Matthew Poole, in reply to
But most New Zealanders expect Labour to do something so utterly compellingly stupid before the election that they will fail to win
I suspect it's actually more that most New Zealanders have been conditioned by the polls and reportage thereof to believe that National's victory on the day is assured.
-
But most New Zealanders expect Labour to do something so utterly compellingly stupid before the election that they will fail to win
This is a way to explain it.
I suspect it’s actually more that most New Zealanders have been conditioned by the polls and reportage thereof to believe that National’s victory on the day is assured.
As is this.
It's also possible that A is the cause of B, or B is the cause of A.
-
I decided to run the simulation with the figures produced by Peter Green’s poll aggregator (which uses a generalised additive model(GitHub)), and made assumptions about voter enrolment (3,150,000; slighly below average increase) and turnout (77%; slightly below the average of recent turnout). Just for kicks I set the loop to 100,000 simulations, and ran it once.
national_led
0.39955nzf_decides
0.60045This may be an abuse of both models, and an extrapolation too far. However, it speaks to a significantly different narrative than the one being communicated, one which adheres to a reality in which:
“you’ve got to stick with the information that you’re given. And you’ve got to stick with it from month to month and pretend that there is nothing else out there, for the sanctity of that information.
I don’t judge him. A man must have a code, and Gower sticks to his.
-
Sacha, in reply to
must have a code
or must have code :)
-
More playing… taking away an Epsom ACT electorate win takes National’s chances of an outright victory down from 40% to around 33%. That’s maths.
-
Matthew Poole, in reply to
taking away an Epsom ACT electorate win takes National’s chances of an outright victory down from 40% to around 33%.
What happens if you plug in CCCP with an electorate and whatever their current share is of the party vote?
-
David Hood, in reply to
Good things for National, but you are starting to make what I would call a big assumption that they can gift the Conservatives a seat and not lose any significant support over doing so. I would imagine this is a topic of conversation in National strategic circles though.
-
Download the code and play with the numbers. It’s a lot of fun.
I would start making other assumptions in that case. Firstly, that National would take a hit of a few percent, made disproportionately from young women and ‘aspirational liberals’, and that the Conservative Party would gain votes, some from NZFirst and others from National. Labour and the Greens would also be beneficiaries.
However, none of this moves things sufficiently towards Labour, or away from National. NZ First has enough catchment in the middle at the moment to be the main beneficiaries. Organic growth of 4-5% is necessary for Labour to start looking like they’re in a position to win with the Greens outright.
Peter Green’s model responsibly assumes that the Green Party's ability to turn out polled ‘Green-voters’ is similar to past elections. I’m in the Green Party and we are working towards both higher support and higher turnout, and hopeful that we can achieve that – but it isn’t a given, and I wouldn’t expect anyone to assume it.
-
JonathanM, in reply to
It's probably actually a little tricky to combine the two models due to the lack of independence between the larger parties at least? You can probably get reasonable estimates by sampling party support from small to large using rnorm with appropriate mean + sd as per Peter's estimates + CIs (i.e. replace the rdirichlet), with the largest party's (Nats) proportion set to one minus the sum of the others.
That way you drop the largest problems with dependence (Nats+Lab) and also assure they add to 1. (You don't get rid of the likely dependence between Greens + Labour though, but it's probably not much of an issue).
-
bmk, in reply to
Just to note that I'm very keen to host and promote anything that emerges from these exercises.
Maybe closer to the election there could be a weekly post with updated graphs as David says:
Well, we can probably get it making pretty graphs of the result for inclusion in posts about election things.
I know I'd find this really interesting and it could be a regular thing up to the election. Ideally it would be a stand-alone post with the latest graphs and then you could add your thoughts or someone else could add some analysis. But even if you just had the graphs it would be a great service and be sure to start some interesting conversations.
-
bmk,
Thank you for producing and sharing the code David and Jonathan. I'm installing R now on my home laptop (already got it at work and am familiar with R although hardly fluent in it) so that I can play out some scenarios on it.
Post your response…
This topic is closed.