Stats Chat

The US Supreme Court declined to review seven same-sex marriage decisions today. The StatsChat-relevant aspect is the flurry of maps this prompted:

I think the New York Times (via Twitter) is my favorite version: the square statebins use geography just as an index to make states easier to find, and (in contrast to the last statebins I linked to) they’ve moved Alaska to the right place

(more…)

I’ve written before about how it’s hard to get accurate estimates of the size of small subpopulations, even with large, well-designed surveys.

Via the Herald

Mr Key said that was an emerging issue for New Zealand. “If I was to spell out to New Zealanders the exact number of people looking to leave and be foreign fighters, it would be larger, I think, than New Zealanders would expect that number to be.”

If the government really knows the ‘exact number’, there must have been a lot more domestic surveillance than we’ve been told about.

New Zealanders probably don’t have any very well formed expectations for that number, since we have basically no information to go on. My guess would be along the lines of “Not very many, but people are strange,  so probably some.” I’d be surprised if it were less than 10 or more than 1000.

From Richard Vale, at University of Canterbury

What’s this? You are claiming that we can use Bayesian statistics to predict Game of Thrones?
Probably not, no.

But we can try?
Yes!

…

Further coverage, and his full article

(via Dion O’Neale)

One of the problems with electoral maps is the ‘one cow, one vote’ effect: rural electorates are physically bigger, and so take up more of the map. When you combine that with the winner-take-all impact of simple colour schemes, it can look as though National won basically everything instead of just missing out on a majority.

Using a design by Chris McDowall that I linked earlier this year, David Friggens has mapped out the party votes across the country with equal area given to each electorate.  These maps show where the votes for each major party came from

He also has maps for the minor parties, some of which have very localised support.

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:

• Anyone may add a comment on this post to nominate their Stat of the Week candidate before midday Friday October 10 2014.
• Statistics can be bad, exemplary or fascinating.
• The statistic must be in the NZ media during the period of October 4 – 10 2014 inclusive.
• Quote the statistic, when and where it was published and tell us why it should be our Stat of the Week.

Next Monday at midday we’ll announce the winner of this week’s Stat of the Week competition, and start a new one.

(more…)

• Two more programs to simulate various sorts of colour blindness, so you don’t make your graphs completely useless: Sim Daltonism and Chromatic Vision Simulator

• From the New York Times: Can Big Data Tell Us What Clinical Trials Don’t? (complies with Betteridge’s Law) (via Roger Peng) [and more from KQED, via Mark Ferris]

• Interactive picture of NZ electricity generation, usage, and price (via JJ Eldridge)

Earlier this week, the Herald had a story about antibiotics and childhood obesity. No-one involved really covers themselves in glory here, but probably the reporter comes out best.

The headline (Antibiotics ‘link to child obesity’) with appropriate claim quotes, and the lead

Children who are given antibiotics in the first two years of their lives are at greater risk of becoming obese in early childhood, a study suggests.

are fine, though there’s a sub-headline “Study finds some treatments kill bacteria that affect rate at which toddlers gain fat” that manages to be wrong in every respect.

The story goes on

The United States research is based on the medical records of 65,000 children between 2001-2013.

Researchers found infants who were given broad-spectrum antibiotics at least four times in their first two years were 16 per cent more likely to be obese by the age of 5.

That is in the research paper, but it’s much less impressive in context

The circles are the observed ratios of rates of obesity, the lines give the margin of error. You can see that, compared to no exposure, one exposure seems bad, two or three not so bad, and four or more worse.  They’ve looked at a lot of comparisons that don’t show a clear pattern, and picked out the biggest number.

There are other quibbles, for example, ‘broad spectrum’ doesn’t mean what it usually means, it means ‘everything except penicillin and amoxicillin’, but the biggest problem is confounding.  There are lots of things related to obesity and to antibiotic prescription, and it wouldn’t be at all surprising for one of them to explain this relationship.

More importantly, there’s no way this explains any meaningful fraction of the increase in childhood obesity.  Only about 3% of children were in the ‘4+ broad spectrum exposures’, so even if the ratio of 1.16 was true, the antibiotics would only be responsible for about half a percentage point of the obesity rate.  Even less of the the obesity increase would be explained, since antibiotics aren’t actually completely new.  Differences between countries also don’t seem to fit this as an explanation. For example, South Korea used to have a serious antibiotic overuse problem, which was reduced by new regulations in 2000. They don’t have much of a childhood obesity problem.

Still, there’s no reason the result couldn’t be true to some extent.  Antibiotics do affect gut bacteria, and gut bacteria are important in metabolising food. It might be true, but it’s definitely being oversold, and there are more urgent reasons not to over-use broad spectrum antibiotics.

The other part of the story that’s disappointing for the light it casts on science communication is some of the response

Some New Zealand experts are sceptical. Fight the Obesity Epidemic founder Dr Robyn Toomath noted that the latest study was funded by the American Beverage Foundation for a Healthy America, founded by the soft-drink industry.

“This is the industry buying crap science,” she said. “People who are poor get sick and get more antibiotics. They are more likely to be fat and a lot of other things as well.”

I’m in favour of scientists commenting on public issues, and I don’t see anything wrong with advocacy, but I think public allegations of intellectual dishonesty need a bit more detailed backup than this. You’d almost get the impression that the research hadn’t looked at number of doctors visits or any indicators of poverty.

Team Ratings for the Grand Final

The basic method is described on my Department home page. I have made some changes to the methodology this year, including shrinking the ratings between seasons.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 200 matches played, 117 of which were correctly predicted, a success rate of 58.5%.

Here are the predictions for last week’s games.

Predictions for the Grand Final

Here are the predictions for the Grand Final. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

Team Ratings for Round 9

The basic method is described on my Department home page. I have made some changes to the methodology this year, including shrinking the ratings between seasons.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 32 matches played, 24 of which were correctly predicted, a success rate of 75%.

Here are the predictions for last week’s games.

Predictions for Round 9

Here are the predictions for Round 9. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

Team Ratings for Round 8

Here are the team ratings prior to Round 8, along with the ratings at the start of the season. I have created a brief description of the method I use for predicting rugby games. Go to my Department home page to see this.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 54 matches played, 34 of which were correctly predicted, a success rate of 63%.

Here are the predictions for last week’s games.

Predictions for Round 8

Here are the predictions for Round 8. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.