OnPoint: Don't put words in our mouths, Rob
142 Responses
First ←Older Page 1 2 3 4 5 6 Newer→ Last
-
Good for you Keith.
-
You're on the money for attacking Labout about the racist ramifications of its data release. I don't know whether its dog-whistle was intentional or not, but I don't care. If they didn't know, then they're not qualified to comment.
Your attack on the statistical methodology is hysterical, however. The data sets are robust, and so the imputation of ethnicity is robust. 39.5% of homes, vs 9.5% of population. What is up for debate is the cause of that; in the absence of other data, it might be easily acceptable to say that there are no conclusions that can be drawn, but there is other evidence. Pretending otherwise doesn't strengthen your attacks on Labour.
-
Statistical model works for predicting lung cancer but with a disclaimer
"The tool is designed for people between the ages of 50 and 75, who smoked between 10 and 60 cigarettes daily for a period of 25 and 55 years. It is preceded by a disclaimer that reminds individuals that the tool is only a prediction based on statistics, and does not mean someone will or will not develop lung cancer."
www.mskcc.org/cancer-care/types/lung/screening/lung-screening-decision-tool -
Yes Keith. Thank you. What follows is my take on the quality of the research which doesn’t even mention Chinese ethnicity nor Bayesian estimation…
Having been overseas (Australia) for a month and coming back in the middle of this saga, I feel that Rob scored an own goal on this one. Maybe that’s the Herald and Talkback Radio dragging down Rob’s writing, but Rob gave them their ammunition. And I hold him responsible for that.
In an effort to catch up, I read what was available and used the original score and ignore criteria (available for download)
I’d score Rob’s work as reported in the original Herald
as:
Just Observing People? fail – we don’t even have individual level data just aggregate stitched together from different sources. No causal interpretation possible.
Yet Another Single Study? fail – doesn’t draw in enough context, and doesn't directly measure the variables it tries to talk about.
Might there be a different explanation you aren’t told of? fail – lacks context of all the other factors which influence house prices and claims one is of “big influence” based on no direct evidence.
Small (sample)? fail: 45% coverage claimed, but not a representative sample that we know of, and it is a short time scale if you are comparing historical trends
Original Information Unavailable? fail. Problems with open access for other researchers to analyze. See also my questions further down on “weighting”.
Headline exaggerated? fail – mentions many things which are not even measured directly: Did not directly measure “investor” or “speculator” status nor “residency” nor “citizenship”.
No independent comment? independent comment sought from Nick Smith +1
Higher risk? fail (see my comments on what was actually measured in terms of different influences on price) as no effects estimated at all
Public Relations Puff? yes if you substitute political for public relations
Half the picture: fails to address lots of other variables needed to model factors influencing house prices. Only then can you talk about the unique contribution of overseas buyers on price (and possible interactions).
Unjustified Advice? advice from Murray Horton to “follow Australia”. Sound advice? Well, Australia has the same story running there so I’m not sure the “problem” is “fixed” by following Australia. +1
Relevance unclear? fail. Didn’t even manage to be explicit about whether they are talking house prices or numbers of sales. Is it volume or value of sales? It pays to be clear about your dependent variable.
which gives it 2 out of 12 (one point each for Nick Smith and Murray Horton which seems ironic but there you go) although it’s hard to say for sure. Which is why I linked the article and gave my scoring. That’s an ignore from me.
PS: I note that in the original Herald Article under the table Auckland House Buyers by Ethnicity is the little phrase: “NOTE: weighted by volume of sales per agent”. Has anybody else asked Rob to explain what that means? As readers of public address may remember I’m big on matters of the correct use of weighting as it can hide a multitude of estimation issues. May we have the unweighted figures please?
-
Russell Brown, in reply to
Unjustified Advice? advice from Murray Horton to “follow Australia”. Sound advice? Well, Australia has the same story running there so I’m not sure the “problem” is “fixed” by following Australia.
To be fair, "following Australia" would achieve the major step of helping us understand whether we even have a problem.
-
The data sets are robust, and so the imputation of ethnicity is robust.
The data is not robust. It's for one company over a period of three months. It doesn't take variations between companies or seasonal effects into account.
in the absence of other data, it might be easily acceptable to say that there are no conclusions that can be drawn, but there is other evidence.
What is the other evidence?
-
steve black, in reply to
I absolutely believe we should be measuring it. I did give it a +1 for trying out the Australian idea of collecting information.
Hey, I’m a statistician. If we don’t count stuff I’m out of work.
That may be the only lasting value to dumping this on the media. Forcing a reluctant government to actually collect the stats needed to assess the situation rather than make stuff up. That’s an improvement. But I wish the work had been written in such a way as to be evidence based, and focus on how we can’t assess things properly at present because we lack the evidence.
-
I still think that Rob should have buried the data.
There was no useful way to use the data without creating problems such as these.
-
Steve Curtis, in reply to
The data is not robust. It's for one company over a period of three months. It doesn't take variations between companies or seasonal effects into account.
Good grief man you are grasping at straws: seasonal variations !!.
Imagine picking the slow season for real estate sales, that changes everything then. -
SHG, in reply to
The data is not robust. It's for one company over a period of three months. It doesn't take variations between companies or seasonal effects into account.
It's for one company <i>that specifically targets Chinese-speaking house buyers in its marketing</i>.
-
That was solid Keith, now is there any way we can get you and Rob on the same thread? He seems confident enough rebutting others more directly.
-
Steve Curtis, in reply to
that specifically targets Chinese-speaking house buyers in its marketing
So you have the lists of buyers from agencies ? Their web sites look all similar to me.
Were you meaning B&T were specifically targeting overseas clients.
-
steve black, in reply to
It’s for one company that specifically targets Chinese-speaking house buyers in its marketing
That’s something I’ve been wondering about and haven’t seen mention of so far in the context of how representative this data is of all market sales. To what extent is it usual practice in all real estate companies to target Chinese-speaking house buyers? All companies do it? Some do it, some don’t? Are there big differences in the spending on marketing this way by the real estate companies which do it? So many questions in terms of understanding the potential representativeness of the sample data compared to the sampling universe of all sales.
Oh, and just for fun, I thought I should include my suggested re-write of the original Herald headline having removed all the unnecessary words and those mentioning variables which aren’t actually measured: “People may influence Auckland’s property market”.
That’s whittled down from the original: “Leaked figures support claims that Chinese investors are a big influence on Auckland’s overheated property market.”
Yes I know my version probably wouldn’t have made the same media splash…
-
steve black, in reply to
Were you meaning B&T were specifically targeting overseas clients.
Are you highlighting the old switch-a-roo from “Chinese-speaking house buyers” to “overseas clients”? That seems to be a major conflation between two different concepts right back at the beginning of this saga A conflation leading to a conflagration.
-
SHG,
From the B&T website, where it talks about why you should choose B&T to sell your house instead of some other agency:
http://www.barfoot.co.nz/about-us/our-partners
Our partnership with HouGarden puts your property in front of the Chinese market. Your property is advertised on the website, giving you exposure to non-english speaking audiences./facepalm
-
Steve Curtis, in reply to
/facepalm
You do know of course Harcourts have their own Chinese website, as they are an international agency.
http://harcourts.cn/
Barfoots has to link with another Chinese company as its an Auckland only business. This Barfoots chinese connection doesnt make it any different from other agencies. They all go after the buyers no matter where they come from. -
Sue,
I'm so glad i don't read sunday papers, but i'm so sad to see words twisted like that. I think what Rob Salmond and the labour party have failed to do is listen.
Listen to people who are hurt & ashamed by a party that at it's roots is about people. They are so busy fighting to assert the rightness of what they are saying they didn't notice they've exposed some seriously ingrained racism in this country. Why are they not sad & embarrassed and apologetic about the hurt and pain they are causing all the many different asian communities in NZ.
Why is that not the lead on Rob's article, inside of an afterthought.
-
Steve Curtis, in reply to
hurt and pain they are causing all the many different asian communities in NZ.
Really ? You would know this how ? Speaking for all the asian communities sounds a bit of ' I know best'.
Many other countries have restrictions on no resident buyers and we kidding ourslelves by having a non discriminatory policy. Even China has restrictions on those who are non resident. -
I am so fucking pissed off. Rob says of me, Keith and CZ, “Their criticism was less about Labour’s intentions”?? Either Rob was lying or he didn’t bother reading my column (possible): because THAT WAS MY MAIN CRITICISM. Jesus, my blog didn’t even MENTION the effects of any racist backlash on the Chinese community.
Rob’s latest column is just barefaced partisan hackery. I know Russell reposted it to encourage generate debate, but I’m embarrassed that it might be seen as an endorsement of Rob’s independence. Russell, I think that perhaps for Speaker posts it’s a good idea to include a line about the author’s political party affiliations and employment.
Seriously, this is just blatant damage control for the Labour party. Rob*lies* about the debate, and is entirely focused on framing critics of the Labour party as CRAZY and IRRATIONAL while carefully singling out three Chinese critics for praise in order to avoid accusations of racism. He’s shitting on any non-Chinese person who supported us. Because obviously, if you’re not Chinese and say exactly the same things that me and Keith said, and openly supported our positions, you must be completely irrational.
Keith, this is what happens when we fight them on the stats instead of on the solidarity. It goes “Ah yes, much respect to the Chinese who are good at stats [ignores substance of everything the Chinese people were saying because they know nobody understands stats so you can say whatever you want about what the Chinese people were saying about the stats] everybody else is CRAZY.”
-
SHG, in reply to
They all go after the buyers no matter where they come from.
And without knowing whether or not all the agencies have Chinese-language websites, or partner with Chinese agencies, and how effective they are in doing so, and whether they all did exactly the same things in the 3-month period, we have no idea whether B&T's "Chinese name" buyer data means anything at all.
WHICH IS THE WHOLE POINT
-
David Hood, in reply to
Keith, could you elaborate on what you see as the seasonal effects on name frequency?
-
Sacha, in reply to
39.5% of homes, vs 9.5% of population.
What is up for debate is the cause of thatYou did read what Keith wrote above, right? Your second line I quoted is a pretty good summary. Well done.
-
Sacha, in reply to
I am so fucking pissed off.
And fair enough. I detest seeing good people and whole cultures shat on for the short-term gain of smug white men wearing suits. It's also profoundly unoriginal. Not in my name, you fuckers.
-
Sue, in reply to
Really ? You would know this how ? Speaking for all the asian communities sounds a bit of ' I know best'.
No the 'I know best' is anyone who has not noticed people who are hurt by all this. You must be living in a nice closed world if you haven't spotted people hurt by this and that those people are from a mixture of asian communities, of new immigrants and families who have lived in NZ for generations.
Yes NZ needs to start by registering the country of origin of house buyers, then if it's shown to have an impact (which people pretty much believe is true but believing is different from knowing) to do something about it to make owning your own house in the region you choose an affordable option for those who choose to live in New Zealand, either by birth or immigration. those 2 parts - a register and a possible future taxation/limits on overseas owners are not actually in question by anyone (well except john key and act) The problem is with what labour did, how they did it and who they hurt.
Why can't Labour step up and go yeah, we did this wrong. Our intentions were good but we did this wrong and the negative side effects are not worth this.
-
Thank you for writing that Keith
I had read Rob's rebuttal and thought my reading of your first statement must have been off
Not sure why Labour are refusing to get the message that Chinese names mined from data does not mean Chinese offshore investors are buying the country from under us "hard working Kiwis "What isn't being mentioned in the "overseas money coming in scandal " is that present owners are the ones cashing in
And if Labour were able to bring house prices down, all the present owners are going to be losers, which makes for an awful lot of unhappy voters, way more than renters wanting to be owners I would think
Sort of like the Power Company sell off, they haven't thought it through
Post your response…
This topic is closed.