Posts by steve black
Last ←Newer Page 1 2 3 4 5 Older→ First
-
Speaker: Losing cultural treasures under…, in reply to Sacha,
some separate legal instrument that comes into effect upon your death
is that a thing?
It is called a will, last I heard. I am currently acting as the executor of a friend’s Estate right now. My job is to distribute the property in his Estate as he requested.
-
Speaker: Losing cultural treasures under…, in reply to Jeremy Malcolm,
He asked about how he could STOP the copyright at his death not EXTEND it.
If the publisher owns the copyright it is in their hands.
If you own the copyright yourself I presume you can state in your will that you wish the executors of your estate to place any copyright you hold into the public domain. Property is property, last I heard.
-
Speaker: Inequality: Too big to ignore, in reply to Ian Dalziel,
Yes thank you Ian. I do like to get the link in there.
So here we have an economist from each end of the spectrum (if you believe that the left/right distinction is a sophisticated enough model) who agree: bogus methodology. I too always point out bogus methodology even for conclusions I agree with because for me if bogus methodology is allowed to stand unchallenged than we don't have a chance to lift our game. We deserve better.
-
Speaker: Inequality: Too big to ignore, in reply to Bart Janssen,
If after that you conclude their findings are scientifically flawed, go for it with your criticism. But at the moment it feels like you are taking pot shots at this work without any basis for it other than your reacons about cause and effect. It may turn out that you are right and their conclusions are flawed…
Ben is off the hook.
This was nicely done by an interview with Tim Hazeldine (economist of the left) and Eric Crampton (economist of the right) and they were agreement that the methodology was as dodgy as I suspected in my first response to Ben (far above). However, in their critique they came up with a number of new sorts of dodgy which I hadn’t guessed at beyond using ecological correlations (without reading the methodology myself).
I tried to link to the audio of this earlier but it wasn’t visible to me. Now I feel I should post this even though I haven’t found the link.
-
Speaker: Inequality: Too big to ignore, in reply to Sacha,
Bill English tells RNZ the report is wrong (and Guyon does not let him off the hook). Then Grant Robertson says it’s right while Susie Ferguson spends most of her time trying to blame previous Labour govts.
Alas, Guyon failed to pick Bill English up on his use of the bogus "50% of people pay no net tax" statistic. I remarked to my wife at the time that Keith Ng has had a bit to say about this sort* of ill defined bogus Nat party (based on Republican party) ideological spin dressed up as an informative and well defined statistic. I was going to have to do the research so I could point it out here in an informed manner, but fortunately my mate Thomas Lumley has already done it:
why "50% pay no net tax" is bogus
And Susie Ferguson is in good company when she blames growth in inequality on the previous Labour Governments. Our Prime Minister agrees and goes further blaming the Green Party.
Key blames Greens for growing inequality
It really must be silly season.
* note: I say this sort because Keith was responding to a slightly different (equally bogus) series of claims about the tax burden on higher versus lower income people
I'm sure Keith has specifically mentioned the bogus stat "net tax" issue before but I haven't found it. Maybe somebody else can help us out... Meanwhile, I must say that reading up on how earlier NAct budgets were pitched (versus how things turned out) in Keith's older posts makes very interesting reading.
-
Speaker: Inequality: Too big to ignore, in reply to BenWilson,
It’s not like it’s a crime to look at your observations and then predict a model that looks like the pattern you see there.
Well…changing the goal posts (hypotheses under test in an experiment) leads to some sticky ethical as well as scientific issues (incorrect p values) in my neck of the woods, thus is frowned upon:
-
Ben,
You are describing a single case crossover design, and Matthew Poole's dismissal of your suggestion leads me to think he might be a Rocket Scientist rather than a Statistician. ;-)
I'll try and keep this to a level which is appropriate here in a place of intelligent people who aren't statisticians, but be warned I don't always succeed.
As to how strong the inference is from these sorts of international comparison studies, it depends on whether the OECD and the IMF have considered each country separately (as a single case crossover design with substantial policy changes within each country representing changed treatments) and then looked at whether they find the same effects (magnitude and direction) replicated across counties. If instead they have done what is far too commonly done in macroeconomics and international comparisons and looked at ecological correlations based on using the countries as cases (albeit in some cross-correlation in a lagged time series style) then the correlations they are working with are unrelated to the real underlying correlations (let alone causes) and are not particularly informative.
Reference to ecological correlations
Since I haven't read the analysis methodology of the studies in question I don't know what they have done. What I do know is that rather a lot of international economic and social comparisons suffer from this. It seems to be accepted practice to interpret analysis based ecological correlations as if they tell you something useful. The only redeeming thing is that if both sides of the debate are using the same approach then this OECD work acts as another refutation of some beliefs (eg "trickle down") by people working in the same analysis paradigm. Much of macroeconomics and international comparisons seem to be based on "ecological correlations" and relatively decoupled from microeconomics. I agree that you are right to feel uneasy about that.
-
Hard News: Housing, hope and ideology, in reply to Sacha,
Apartments, or...
Flatmates. Why hasn't this come up in discussions about there being too many bedrooms in the houses they have available? Did they not go to University? Did they live at home until they entered Parliament?
-
Yes John Key's numbers don't stack up. But before you even get into the details of his two numbers, I think this is another case of "asset sales arithmetic".
The first thing I thought of with John Key's numbers this morning is: over what time period? Spending $12M per year? per election term?
And comparing this to what exactly? Building $500M in buildings? This means you haven't given away the money. You have invested it in buildings. You now have an asset.
Remember those? State Assets which were sold to get cash although Keith Ng (among others) demonstrated that on traditional economic analysis it was worth more the the country to keep them?
I sense a triumph of ideology over good management practice, and backed by fluffy numbers and even fluffier thinking.
-
The bogus polls which relate to the debate have given us a textbook example of why all self selected polls are not likely to be representative of the general population -- no matter how big the sample size. And if one of them were representative of the general population which one might it be? We have no way of knowing.
http://www.statschat.org.nz/2014/08/28/bogus-clicky-polls/
Note also that based on the variation in results, Thomas Lumley says the accuracy equates to a properly conducted random sample of about 40 people. That my be just a quick back of the envelope calculation, but I think you get the idea of just how bogus self selected polls can be.