Posts written by Thomas Lumley (2645)

The Herald headline is “Icecream cravings like cocaine: study”, and it’s another paper from our friends at the American Journal of Clinical Nutrition, but this time filtered through the British press.

Co-author of the study, Dr Kyle Burger, from the Oregon Research Institute, told The Daily Telegraph overeating high fat or high sugar foods appeared to change the way the brain responded to the food.

That’s not quite what the paper said he found.  The study looked at brain activity in people drinking a milkshake compared to a tasteless drink, and found that the milkshake effect was smaller in people who had eaten icecream frequently in past weeks.  But it wasn’t smaller in people who had eaten cookies, chocolate, hamburgers, cake, or chips, and there wasn’t any relationship to total dietary fat and sugar.   So, in fact, the study found that overeating high fat or high sugar foods did not appear to change the way the brain responded to the food.  As far as I can tell (and, unlike you, I can get the whole paper), the study does not report at all on the relationship between craving for icecream and the brain activity measurements.

So. The study found that brain response to a milkshake was weaker in people who eat a lot of icecream, but was unrelated to other high-fat or high-sugar foods, and that people who have cravings for icecream eat more icecream than people who don’t have cravings for it.

Stuff has a story on robot-assisted surgeries, claiming

Patients who have robot-assisted surgeries on their kidneys or prostate have shorter hospital stays and a lower risk of having a blood transfusion or dying – but the bill is significantly higher, a study found.

That’s not quite what the study found.  The abstract says

While robotic assisted and laparoscopic surgery are associated with fewer deaths, complications, transfusions and shorter length of hospital stay compared to open surgery, robotic assisted laparoscopic surgery is more costly than laparoscopic and open surgery.

The researchers used the Nationwide Inpatient Sample, a random subset of US hospital admissions, to compare three approaches to prostate and kidney surgery: laparoscopic (‘keyhole’) surgery by hand, robot-assisted laparoscopic surgery, and open (non-keyhole) surgery.  They found that laparoscopic surgery, whether robot-assisted or not, was safer than open surgery, but they didn’t report an advantage of robot over non-robot keyhole surgery, just an increase in cost.

Now, you might well be muttering about causation and correlation, and asking “How do we know the open surgeries weren’t just more difficult cases?” If you aren’t, then you can start now and I’ll let you pretend you were doing it all along.  Since surgeon experience makes a big difference, we should also worry whether it’s the most experienced surgeons who get the expensive and shiny new robots.

The researchers did try to cope with this problem using a technique called propensity scores.  Essentially, they tried to classify patients according to how likely they were to have an open surgery vs manual laparoscopic surgery vs a robot surgery, and match the patients so comparisons were done only between similar cases.  However, the researchers did say in the main body of the paper  “Results from unadjusted and propensity adjusted analyses were largely similar”, ie, the attempt to remove bias didn’t actually remove any bias.  The optimistic view is that this means there wasn’t any bias; the pessimistic realistic view is that it means the adjustment probably failed.

Surgical robots have been a bit of a disappointment.  It’s not that they don’t work, but they were supposed to have huge and dramatic advantages (over and above “ooh, shiny”) and these huge advantages don’t seem to have materialized.

The Herald

More than 25,000 animals including fish have died during research and teaching at the University of Otago since 2009, and the figures are expected to increase significantly when statistics are collated for 2011.

No, really?  Going from a two-year total to a three-year total will result in an increase?

We collect multiple years of data for two main reasons: to compute averages (reducing statistical noise) and to compute differences (to learn about trends). Computing totals across multiple years is rarely helpful.

The Otago Daily Times does a better job: it uses the multiple years of data to give trends

Animal use by different groups in New Zealand has dropped by 29% during a three-year period from a record high of 341,520 recorded in 2008.

and

The number of animals being used for research purposes by universities in 2010 has dropped 43% from a high of 123,739 in 2008.

The American Journal of Clinical Nutrition missed an opportunity when they put a meta-analysis of randomized studies of chocolate in the March issue rather than the February issue, but that hasn’t stopped the media coverage.  Stuff has a headline asking “Is chocolate really heart-healthy?”  There’s a law of journalism that states any headline ending in a question mark can be answered by the word “No”, and Stuff does not disappoint us here.

The newspaper story is actually pretty good — very good by the standards of nutrition journalism.  There are four key points, and the story makes three of them clearly and alludes to the fourth (more…)

During the Christmas news-deficiency season there was a story warning people to be careful around all the dangerous holiday accident risks.  Our Stat of the Week award went  to Sonia Pollak, who pointed out that 3040 accidents predicted for Christmas Day was quite a bit lower than the average for other days.

The story has returned in the past tense.  The ACC is reporting 3238 accident claims for Christmas Day (close to what they predicted, about 25% lower than an average day) and $850,000 in claims.  Even assuming that the initial claims relate only to medical treatment, that’s about half of the average daily expenditure  (medical treatment, $512 million/yr, hospital treatment – $199.3 million/yr)

We could prevent accidents and save money by being especially careful on Christmas Day.  But we could probably prevent more accidents and save more money by being especially careful on some other day, such as the anniversary of some significant historical accident.  I suggest June 14.

In a report on a cannabis-growing operation in Hamilton, the Herald gives us both new dodgy numbers and recycled dodgy numbers.

The new numbers (from the police)

• Between 50 and 100 plants
• a mature plant worth about $1000
• If 100 plants were allowed to grow to maturity, conservatively worth $400000 in income.

Simple arithmetical operations on these numbers show them to be not entirely consistent: if you can turn a crop that will at maturity be worth between $50,000 and $100,000 into $400,000 in income, you’re in the wrong business.

The recycled numbers are

A study involving the police and Environmental Science and Research released last month found that levels of THC, the primary psychoactive compound in cannabis, was more than four times stronger than in 1996 when ESR last tested it.

As we pointed out last time this statistic appeared, and will keep pointing out when it gets recycled in the future, what they actually found was that in 2004-2006 the maximum concentration was four times higher than the mean concentration in 1996.

The faster-than-light neutrinos found at CERN last year were always very likely to be some subtle experimental error, and the odds of this seem to have gone up. Professor Matt Strassler, who actually knows what he is talking about, has some blog posts on the topic, including a very nice graphic that will be familiar to statistics teachers.   The initial news reports last week said the FTL problem had gone away, but that appears not to be true.  What we’re waiting to find out is both what the estimated speed is, and how much uncertainty there is.  It could still be that the results are inconsistent with relativity, and that some other explanation is needed. Or it could be that the uncertainty is larger than the 60 nanosecond timing anomaly, so the results are consistent with relativity. Or, the experiment could have been sufficiently messed up not to tell us anything.   In any case, the real test will come when someone repeats the experiment.

(via: Chad Orzel)

Last week, The Australian reported on the politics of same-sex marriage in the West Island.  The story is a bit old (I picked it up from John Quiggin), but it’s such an impressive example that it’s still worth mentioning.

Roy Morgan Single Source survey data from the middle of last year shows that over a quarter of Australians aged 14 and over — 26.8 per cent — agreed with the blunt proposition “I believe homosexuality is immoral”.

That’s a fairly small minority, and although there is variation, it’s a minority everywhere in the country. But there is this:

In 80 of the 150 federal electorates, an above-average number of people support the proposition. 

So. In about half the electorates the proportion supporting the proposition is  above the national average (but still a minority), and in about half it is below the national average. That sure tells us a lot.

Thirty-three of these 80 are Labor seats. They take in a who’s who of the ALP.

Of the half that are above the national average, the proportion held by Labor is 41%, a little less than the 48% of all seats they hold.

Family First have collected some data from the Ministry of Education on ‘fees or donations’ to NZ state schools. You can read the Herald story, or just get the press release directly, without the light editing the newspaper provides.

The figure of $250 million per year is pretty impressive. NZ has about 750,000 children in school  (EducationCounts, school roll returns), so that’s maybe $330 per child, which is quite a bit for free education.

Unfortunately, Family First thought that wasn’t a big enough number and decided it would look better as a total over four years of about a billion dollars.  There’s obviously nothing special about four years here: school donations have been an issue for much longer than that, there wasn’t a big change in laws or policies four years ago, and no-one plans budgets over four year blocks, not schools, not government, not families.  It’s bad enough when reports don’t provide any context for numbers, without going out of their way to remove it. Presumably if Family First had been able to get ten years of data (and inflation-adjust it), we would see a headline total of $2.5 billion.

Interestingly, one of the paragraphs from the press release that didn’t make it into the story said that the Ministry couldn’t provide information on the proportion of parents who actually paid the donations.  This is interesting because the Herald has collected information of this sort itself in the past, as reported in this story and the Wikipedia article. As you might expect, the proportion paying was lower in the low-decile schools, but even in the high-decile schools there were quite a few who didn’t pay.

From the Herald (and via a Stat of the Summer submission), a story on students leaving NZ because of job fears.  According to the poll, students are more likely to be worried about getting a job when they finish than about their next rent bill or next decent meal. That’s what you’d hope, and it doesn’t really seem surprising.

We aren’t told whether the students planning to head overseas are the ones worried about getting a job here, or whether they’re the ones who could easily get a job in NZ but looking for higher pay or new surroundings.  In fact, it’s worse than that.  The story says

While a third planned on moving straight into career mode after study,…18 per cent were heading overseas – either for work or timeout.

That is, the 18% planning to head overseas includes all the ones aiming for time off and foreign travel.  This means there is absolutely no support in the story for the headline claim that students are leaving NZ because of job fears.

The other interesting part about the story is the attribution: “an online Colmar Brunton survey of more than 1000 students conducted on behalf of Student Job Search.”  That sounds like it’s one of the surveys from saywhat.co.nz, a site where people aged 15-30 can sign up to do fill in questionnaires for the chance to win prizes.   It’s hard to tell how accurate these polls are: they are a self-selected sample, but the respondents do have to give some personal information that could help adjust for the selection bias. The website uses age and gender as examples, which would be completely inadequate, but they may collect more useful information in the sign-up process (or, of course, by mining it from Facebook and the like).