As you could probably imagine, this place gets all sorts of requests to look at various bits of polling data, or to compare some voting intention against some economic indicator – right through to some snarky requests involving the extraction of various body parts from various orifices.
Instead of doing these things on an ad hoc basis, it might be a good time to formalise the whole affair. Every Monday I’ll throw open a Requestathon thread where you can make requests for what sort of bits and pieces you’d like to see analysed. When enough of them turn up that are doable, we can throw the results together for a Sunday article that will act as a substitute for the Nerdy Sunday posts (because, though try as I might, I can’t actually think of 52 separate nerdy things to talk about a year!).
I recognise that some of you folks are shy (or just plain old unsociable – I can’t quite figure out which 
), so you can either drop the requests into the comments or send them to me by email like you usually do. I’ll repost them in the comments without your name attached.
No request is too peculiar (why do I suddenly feel nervous), and by opening it up to a crowd that knows more about things than I do, some interesting little nuggets might come out of the whole shebang.
24 Comments
Can you post a picture of you without that newspaper hiding all the good stuff?
Perv!
Any chance of some ACT polling analysis? I appreciate that the ACT government is a glorified local council and no one outside of it cares at all, but it might make some of us who live here (NOT by choice!) feel included.
Hi Possum
What I would most like right now is a calculation of the number of Coalition voters dying and not being replaced per day, as per your excellent (and personally heartening) ‘Coalition Train Wreck’ post. I would like to be able to say breezily, on any appropriate occasion, with authority, that ‘x Coalition voters died today and were not replaced’. I think this stat needs to get around and yeah, verily even unto the highest councils of the Liberal Party. Let them sweat on what to do about it!
Cheers
I’d love to see stats for gen Ys skipping the vote at a federal level and/or state level against historical figures for the same age group. Anecdotally almost noone I know (am gen y meself, laugh if you will) is actually voting on a regular basis. I have a theory that we’re a bit of a cynical mob who as a group are avoiding democracy a bit more than in the past. It’s the kind of thing that needs actual testing for veracity though, so that my lazy friends have something to blame at the pub
Anonymous Person 1 asks:
“I’ve been wondering, from a numbers point of view, which previous presidential contender 2 weeks out from the election does Obama most look like? Who does McCain most look like? I mean in terms of who has been a ahead/behind to a similar degree at this stage in their campaign. Doesn’t matter if the parties are switched – it would just be interesting to draw a historical analysis between this campaign and previous ones where we know the outcome.
My explanation is a bit torturous, but essentially I’m wondering if the data support a statement like “Obama’s win probability looks like Reagan’s” or “McCain’s looks like Dukakis’” etc.”
Anonymous Person 2 asks.
“I’ve got one for you Poss,
Why did Nielsen get it so wrong at the last election considering that they have such a great reputation and you psephs hold that poll in such high regard? Have they since fixed whatever it was that caused that fuck up?”
Anonymous Person 3 asks:
“Re your reqestathon-a-majiggy.
Find out what Mumbles has been smoking with his US Election thought bubbles.Model it if you must, make a spiffy graph, even bring Dennis Shanahan into it if you have to, but please get to the bottom of it!?!”
heh! I’ll put it to Peter and see what he says!
Or I’ll put it to Peter and duck
Hey Possum, awesome idea!
I would love to see something like the ALP’s primary vote amongst the youngest age bracket over time (If such data is available… I know that many pollsters don’t break down their voting intentions into demographics) and then have it compared to the Greens primary vote amongst that same group. I think it would be interesting to see if the Greens are “stealing” an increasing, or decreasing, proportion of this vote. I believe that if they can improve their vote in this group in the long-run they are going to keep more throughout their life and grow as a serious political force.
Just a suggestion, I dunno.
CountArach – I don’t want to steal Possum’s thunder but I’ve been reading this report:
http://www.aph.gov.au/Library/Pubs/RP/2008-09/09rp08.pdf
Which shows the Green vote amongst those < 25 skyrocketing over the last few elections. 8% in 2001 to almost 20% in 2004. I presume it would be even higher in 2007.
How about a generalised discussion on the difference between correlation and causation.
Psephys love to confuse the two.
A good title might be “How any random statistic re-inforces your prejudices”
Oz @ 11 – Thanks heaps for that mate! That’s exactly the sort of thing I was looking for.
Poss, Any stats on where people think each party falls on the (admittedly largely irrelevant) political spectrum. With Humbull trotting out the old communist red book stuff today, it got me wondering just how many people actually think Labor is socialist.
Given the recent shellacking the NSW government got, I’d be interested to see a graph of primaries, TPP highlighting the terms of Rees, Iemma and Carr?.
i’d be interested to see if Rees can turn the polling around, the people of NSW seem to have said at the last election “this is your last chance”, and its been shown in the recent by-elections.
In the nicest possible way, I request that this blog returns to it’s former glory. It used to be witty, insightful, informative and downright clever. (and regular dammit!)
Lately, not so much
Nerdy Sunday has been moved to another time zone.
For those waiting in the queue of anticipation.
http://www.youtube.com/watch?v=tIdIqbv7SPo
Spam and Greeny – I’ve been moving house this week, have the truck today.
So things will resume tomorrow, and you’ve got a very good point Spam.
This place will be returning more to what it was.
Anonymous person 4 asks:
I’ve been wondering about an analysis for some time, and your Requestathon has prompted me to email you. I’d be keen for some kind of analysis of the average sentence and word length used by Rudd through time, and to see whether this corresponds with his approval/satisfaction rating, PPM of even primary/TPP vote for Labor. The hypothesis would be that his approval/PPM/Labor vote is higher when he speaks clearly and directly than when we get a barrage of bureaucratese.
My observation is that he has had periods (e.g. AWB as Shadow Foreign Affairs Minister; Opposition Leader; the financial crisis) when he speaks much more clearly and more directly, than other times such as climate change details when we get huge sentences and complex words or jargon. I suspect that as he resorts to bureaucratese people switch off and his ratings decline. While this might seem a bit trivial and probably a time consuming analysis to do, my gut feeling is that it might affect the government’s re-election chances. Turnbull seems to be targeting it a bit, which might suggest they’re trying to reinforce it in voters minds.
OK, here’s my request.
I’d like to see all the Intrade two-cornered political contest data you have, as of the day before the relevant election, sorted into 10% (or 5%, if there’s enough) bins like in the recent Intrade Mondayish post. For each bin, calculate the percentage of those elections that the candidate actually won and plot it, along with a line of best fit. (You’d only need to plot probabilities of 50% and higher, since for each election you can just pick the candidate that had a higher-than-50% probability).
If Intrade probability matches real probability, we’d expect a linear relationship. How the relationship deviates from that should give us a way of exactly accounting for long-shot bias (and similar effects) in future.
Here’s one to ponder on Possum, is there an Australian conventional wisdom v facts?
http://www.bloomberg.com/apps/news?pid=newsarchive&sid=axt7Qr7cOyVo#
The XHTML guide to be put back up
Would Peter Andren have gotten a seat in “The other place” (Senate) ?