One of the problems with lookingfor patterns in Big Data is that there’s a lot of things that look like patterns. It’s easy to see things that aren’t there.
In a dramatic example, Phil McCarthy feeds a random polygon generator into an automatic face recogniser, and makes random changes to the polygons to improve the recognition score.
(via Ben Goldacre and Prosthetic Knowledge)
As a former Australian, I need to point out that the natural denominator for comparisons between Aus and NZ should include sheep as well as people.
On this metric, Australia has one medal per 4.5 million population, and one gold medal per 50 million population. New Zealand has one medal per 4.4 million population, and one gold medal per 11.8 million population.
Looks like New Zealand is still ahead, even if you include sheep in the population. On the other hand, Jamaica leaves us both in the dust.
Nerdnite is a global collection of groups that get together for presentations on interestingly nerdy topics somewhere where you can get beer. Wellington has had a group for quite a while, and Auckland has its first evening next Tuesday, August 14th, starting about 6:30pm, at Nectar in Kingsland
The presenters will be Siouxsie Wiles (bioluminescent superbugs), Shay Brazier (renewable energy), and me.
Lost receipts are costing Australian taxpayers about A$7.3 billion (NZ$9.4b) in total, or about A$1,000 each, according to a Commonwealth Bank survey.
The story in The Australian goes on to mention that Commonwealth Bank is introducing a product to help, so this is basically an advertising press release. I can’t find out whether the survey is a real survey or some sort of bogus poll (there’s nothing on the Commonwealth Bank media releases page, for example), but there’s clearly something strange about the figures. If you divde $7.3 billion by $1000, you get 7.3 million. If you do the same calculations for the time spent looking for receipts, you get about the same figure. But there are about 12 million Australians who lodge individual tax returns (14.6 million tax returns, 84.7% for individuals), so these figures don’t seem to add up.
[Update (28/8): the media release is up now, but it doesn’t clarify much. The description suggests this is a bogus poll with reweighting to Census totals, but that doesn’t explain the discrepancy with actual tax returns]
The multi-year aggregate headline number is back again. The Herald tells us
John Ivil is the $300 million man – that’s the amount of money he and his team of public servants have saved the taxpayer in two years.
You might think that that’s $300 million per year, or since ‘two years’ was mentioned, $150 million per year. In fact, if you read on, the savings are totals over contracts that run for five to seven years, so we’re looking at perhaps $50 million per year. Still well worth saving, but less than the headline number suggests.
Each week, we would like to invite readers of Stats Chat to submit nominations for our Stat of the Week competition and be in with the chance to win an iTunes voucher.
Here’s how it works:
Next Monday at midday we’ll announce the winner of this week’s Stat of the Week competition, and start a new one.
A big gold medal to Statistics New Zealand for its daily tally of Olympic medals by population (the most usual table is number of medals won), with the numbers sliced and diced different ways. Whichever way you slice and dice, the Aussies will still be cross.
The lead in an otherwise reasonable story about a large employee survey in the Herald today is
Just one-third of New Zealand employees are currently working to their full potential.
If you go and look at the report (the first PDF link on this page), you find that the survey says it’s a stratified random sample, matched on organisation size, and then goes on to say that 93% of respondents “were from private organisations employing 50 or more people”. At little work with StatsNZ’s business demography tables shows that about 57% of NZ employees work for organisations employing 50 or more people, and when you remove the public-sector employees from the numerator you get down to 42%. The survey must have specifically targeted people working for large private organisations. Which is fine, as long as you know that and don’t just say “NZ employees”.
Also, the link between “working to their full potential” and what was actually measured is not all that tight. The 33% is the proportion of respondents who are “engaged”, which means responding in the top two categories of a five-point scale on all eight questions targeting “job engagement” and “organisational engagement”.
Although it’s harder to interpret actual numerical values, since the company seems to use consistent methodology, changes since the last survey really are interpretable (bearing in mind a margin of error for change of around 3%). And if you bear in mind that the survey was done by people who are trying to sell managers their services, and read the report with an skeptical eye to what was actually measured, it might even be useful.
Stuff has a story based on a real and useful poll, but summarised with a dreadful graph. You will have heard statisticians ranting about pie charts and may have wondered whether their medications need to be adjusted. Here’s why we rant.
Notice that the pie isn’t round; it’s an ellipse. Presumably we’re supposed to imagine it being tilted away at some angle (in contrast to the table, the headline, and the legend, which are aligned with the page. Also notice that the wedges have numbers on them — that’s often a sign that the graph can’t be interpreted by itself. The red wedge looks a lot smaller than the blue wedge.
Sammie Jia draws my attention to a story in Nature News about drug-testing for atheletes. The story mentions the impressive performance of swimmer Ye Shiwen in the women’s 400m individual medley, and suggests that tests should focus on atheletes who have unusually good performances or strong improvements.
I’m surprised this isn’t already being done. There are at least three good arguments for it. Firstly, if doping works, the rate will be higher among athletes who have improved dramatically, so more cheats will be found with lower testing costs. Secondly, the successes of performance-enhancing drugs are the cases it’s most useful to catch, either if you think the point is deterrence in young atheletes or if you are doing it for some sort of abstract ideal of fair competition. And finally, athletes who make dramatic performance gains honestly deserve to have any doubts removed.
In contrast to many other approaches to targeting tests, this one is very hard to game. The whole point of doping is to get otherwise-impossible improvements in performance, so there isn’t any useful way to avoid attention. Targeting performance improvements would be even more efficient in circumstances where taking and storing samples is cheaper than analyzing them: it would be possible to store a larger collection of samples and retrospectively test performers who had done surprisingly well. That, more or less, is how outcome-dependent sampling is used in medical research: we take blood samples from everyone, but focus the expensive assays on people who, possibly decades later, stand out for good or bad health.
Recent comments