posted
I've seen this paper referred to recently http://alcoholireland.ie/wp-content/uploads/2010/10/rand-report.pdf and I started reading it, before falling asleep. I'm going to have another go later, but it strikes me as very poor quality. My stats only went as far as first year undergraduate, but I know there are some proper liars (sorry, statisticians!) here that might be able to give a more considered view.
My take was that they're slightly struggling for truly relevant data. What might be useful would be a survey of all alcohol related A&E admissions where someone finds out from where and at what price alcohol was obtained - and maybe asks what the drinkers would have done had the cheapest stuff not had been so cheap (obviously, a bias correction would have to be applied to that).
By the way, when these meta-analyses are done, who judges if papers that don't support the preconceptions of the authors have rightfully been excluded? (Apologies for my preconception that the authors might have preconceptions, but I've spent enough time kicking around science depts to know that it's widely so)
posted
Alex Hopefully others have got more time, but 120 pages is too much for me to wade through.
Questions you can think of when looking at the data are: - Is the data sufficient (in sample size, relevance, timeliness) - Is it appropriate to the remit - Was it collected with independance - Have they (effectively) eliminated non-random sources of variation. - Are they correctly interpreting the results, or over-reaching to support an opinion
regards Ian
-------------------- Normal for Norfolk Posts: 7758 | From: Norwich | Registered: Apr 2004
| IP: Logged |
posted
Too long for me too, but based on the quality of "statistical" papers by non-specialists generally, I'd not hold out too much hope of it being any good. Worth noting that it's funded by a decidedly non-non-partisan organisation (the RAND corp...)
posted
Non-non - they're quite famous, aren't they - very right-wing, big on nuclear tactics during the cold war etc? That's not a judgemental comment btw (I'm aware I'm probably more right-wing than most on here, at least economically...)
quote:Originally posted by Alexander Lake: In what way are they non-partisan?
Very right wing. Donald Rumsfeld was a chairman. Origins with the US industrial-military lobby.
I certainly would not trust them on foreign policy, economics or defence.
But this was commissioned by the EU and out of European office of RAND. But they will have had a feel for the answers required by the DG who commissioned it.
Never ask a question to which you do not know the answer.
posted
That's a good point Wozza. I was slightly struggling trying to work out what answer a tea-party sponsored organisation would try to engineer, but couldn't see that it would be one way or another.
In terms of the the EU would like, even that's tricky - are they really the organisation of puritanism?
posted
I'm not sure that left wing or right wing map easily on to any particular set of views here. The traditional libertarian view would surely be to allow people to drink without any government interference - so support for a minimum price per unit should by tradition actually be a left-wing point of view. (I express no views myself here on what is right or wrong - indeed I'm not at all sure what mine are - so hopefully I steer clear of raising any party political views.)
And "the hoi polloi", Warren? Perhaps Sir Humphrey's views of the LSE were correct.
Posts: 2479 | From: Oxfordshire | Registered: Jul 2006
| IP: Logged |
quote:Originally posted by Druin Burch: I'm not sure that left wing or right wing map easily on to any particular set of views here. The traditional libertarian view would surely be to allow people to drink without any government interference - so support for a minimum price per unit should by tradition actually be a left-wing point of view. (I express no views myself here on what is right or wrong - indeed I'm not at all sure what mine are - so hopefully I steer clear of raising any party political views.)
And "the hoi polloi", Warren? Perhaps Sir Humphrey's views of the LSE were correct.
OK then "οἱ πολλοί" or "the polloi" or just "hoi polloi" = "the masses". Sorry for the tautology. [I learnt to read Greek through Maths not Classics like Sir Humphrey and Bernard ]
The RAND is big business / military-industrial linked so definitely not libertarian-linked.
Pro-MUP is one or several of
Puritanism / Religion People's Health Welfare People's Control Economic Development Anti-competitive Trade Protection
Anything else?
What do we think an organisation linked to "Donald Rumsfeld" would want?
That said I have not read the document. I will read it with an open mind but, as always, a suspicious mind. I like to try and guess the author's possible prejudices and angle before reading.
posted
Druin, you seem to be implying that libertarian and left wing are mutually exclusive. Or have I misunderstood?
Posts: 3096 | From: London | Registered: Aug 2007
| IP: Logged |
posted
I thought it a fairly bland document, as it happens. The questions about how valid it is don't seem to me to relate to statistical technique but epistemology - if you demonstrate a correlation between low price of alcohol and consumption, it doesn't mean that enforcing a minimum price reduces alcohol harms. That latter is an empirical rather than a political question - the politics come in when you start to consider if the restriction of freedom justifies the (possible) health benefits.
Phil - that's a question that fascinates me. My personal answer relates to my personal politics, so I'll share it with you offline only. I really do think that lack of party politics on this board contributes hugely to its high level of civility. But it does seem to me that traditionally the right wing viewpoint is associated more with being libertarian. "That government is best which governs least" is not an obvious description of a left-wing position. In that I'm describing historical beliefs as I see them, rather than articulating any political ideas of my own, would you disagree?
Warren - I have small Latin and less Greek, and I imagine you've more of both than I have, but the opportunity to tease you seemed irresistible.
Posts: 2479 | From: Oxfordshire | Registered: Jul 2006
| IP: Logged |
posted
I thought that these days, political positioning was a 2 dimension graph with libertarian vs authoritarian on y-axis and left/right on the x. This is as proposed by The Political Compass Organisation. It does appear that leftism is more associated with authoritarianism, but prob not best discussed here other than in relation to prospective legislation on MUP.
posted
For that we need a pentagram drawn in mature Domaine Dujac and some Hampshire sausages.
Posts: 2479 | From: Oxfordshire | Registered: Jul 2006
| IP: Logged |
I didn't read the whole trsct, it made me seethe with furious rage. I'll summarise.
Their data collection methods were highly partial and seemed to have been chosen to make sure the desired results came out.
There are several blatant falsehoods about the direction countries tsxes were going and bsrefaced lies about that correlation with drinking rated. The conclusion that more expensive booze causes fewer health and social problems is insultingly duplicitous.
Even ignoring those problems this is an example of marginal value epidemiology, and that's a load of bollocks. You cannot multiply up tiny risks to a population and expect it to have any validity. Moreover, the risk values are so low that even an impressive sounding increase in risk due to some factor doesn't make any practical difference. If you have a 0.00001% chance of your feet dropping off when you leave the house tomorrow, but that chance is 7x higher if you wear polyester, to all intents and purposes you are at no significant risk if you keep buying cheap socks.
When I reviewed epidemiology papers for academic journals i got loads of drivelly ones like that. I would normally write 'Crap' across the front page and send it back to the authors. Obviously that battle is largely lost, new age data generation can 'prove' anything, but that doesn't mean you should treat it like it has any vslud or meaning.
who's who
UBBFriend: Email this page to someone!
![[Wink]](wink.gif)
]
