Song? I only know it as Les Barker’s poem “Have you got any news of the iceberg?”
Milne’s response to having statistical errors pointed out was a classic example of the Dunning-Kruger effect in action. The original paper (Kruger & Dunning 1999) notes that those with insufficient competence in some field ipso facto cannot recognise and respond to corrections in that field, but instead may interpret any such corrections as irrelevant personal attacks. Kruger & Dunning give the example of students who “explain” their low test scores as indicating that their teacher hates them.
@Brent: Of course, media faith in statistics as a magic device to create information out of pure noise isn’t limited to newspapers – as seen in CSI and similar shows where “we used an algorithm to enhance the resolution of an image reflected in the victim’s eye, captured on a crappy security camera”.
The data is uninterpretable, because the numbers used do not express a single known quantity. To put it briefly: the data is rubbish. Too much noise, too little signal.
You defend publishing the data :: you defend publishing rubbish.
Any attempt to “analyse” such data cannot possibly be any better than interpreting chicken entrails, an utter waste of your writers’ (and then your readers’) time before you even begin.
Worse than that, it does actual harm, because the act of publication makes the numbers seem meaningful.
Add to that the complete lack of any mathematical skills on display in this “analysis”, as evidenced by
- the wild leap to claim that there is a “trend”, based on a statistically non-significant level of correlation,
- between two sets of numbers both of which are at best only indirect proxies for what you want them to measure,
- a correlation which even if it were significant still should not be claimed to demonstrate any causal relationship,
… and … I’m actually angry that something so uninformative, so devoid of solid foundation in fact, got published as “news”.
Yes; but now wait for the disgruntled middle classes not to see that. The really really annoying thing about this strategy is that, largely, it’s been working [in terms of maintaining support for National, rather than doing anything productive for NZ] – partly because the disgruntled middle-class subset of the audience is likely to experience confirmation bias; and the worse National do on the economy, the larger the ranks of the disgruntled.
Actually, that’s true of both the 2001 and 2011 surveys of children, so I should correct myself a little: the comparison seems to support “… a rise in overall recreational activity”, rather than specifically “… a rise in sports activity”.
By the way, what was up with the media angst yesterday (e.g. Paul Easton’s piece in the Dominion Post ) about soccer overtaking rugby in popularity among schoolkids (according to a survey from mid-2011)?
Firstly, why was this news?
As measured by NZ sports organisation membership in 2002-2003, soccer already outranked rugby in popularity in NZ 10 years ago.
NZ Golf Association 132063
Netball NZ 123069
NZ Soccer Inc (mens) 105000
NZ Cricket 102759
Touch New Zealand 94291
NZ Indoor Sports Incorporated 77065
Bowls New Zealand 60404
NZ Tennis 42312
NZ Hockey Federation 39574
Womens Golf NZ 36017
Yachting New Zealand 31338
Softball New Zealand 30320
Athletics New Zealand 29050
Basketball New Zealand 28911
NZ Rugby League 28215
(Source: Sport and Recreation New Zealand (SPARC) site <http://sparc.org.nz/research/clubs3.php>, accessed October 2005. Original URL redirects to new Sport NZ site, which does not seem to have the document.)
Secondly, the media focus on rankings seems to have missed a more interesting trend in the data. The children surveyed in 2011 seem to show a higher overall level of participation in sports than in the 2001 SPARC Active New Zealand survey (though it’s hard to be sure whether both surveys are measuring sports activity in comparable terms). For example, 79% of boys and 83-84% of girls listed the most popular options of “running/crosscountry” or “swimming”. By comparison, the 2001 survey found 73% of boys and 64% of girls were “active” in the sense of “taking part in at least 2.5 hours of sport or active leisure per week”. So it’s entirely possible that what we’re seeing is not any absolute fall in popularity of rugby, but a rise in overall sports activity (and apparently more of a rise among girls), possibly as a result of more inclusive sports offerings in schools and communities.
[Lowell’s] fantasies about the Martian canals […] were perhaps more grounded in observation
though of course, unbeknownst to Lowell, his observations were, at least in part, of the blood vessels in his own eyeballs.
Fuller description here (PDF).
Give away the news and sell the analysis?
Only works if readers believe that the analysis is useful … which, sadly, runs headlong into confirmation bias (i.e., people find it easiest to believe analysis that conforms to their existing opinions). Stripping (uninformed) opinion away to leave facts presented without bias is a large part of the “value added” by reputable news sources. (Not sure how much of NZ’s MSM still counts as “reputable” by this measure.)
Also, if there's any "public good" argument to be made at all, doesn't that imply providing the widest possible access to a full understanding of an issue?
Presumably "removing the one electoral seat threshold" means "winning one electorate, but getting less than 4% of the total vote, gets you that electorate seat", rather than "winning one electorate, but getting less than 4% of the total vote, gets you nothing".
But it still reduces proportionality. It would have been better to have the threshold for extra seats set lower, e.g. 1.5% -- if there must be a threshold at all.