Fishing in the Bay

I recently came upon a piece in Slate Magazine by Steve Landsburg describing a very nice price of research. It concerns the financial costs to women of having childern. I thought this article (reproduced below the fold) might provide a nice class room example.

Read the rest of this entry »

On January 26, 2010 the Grattan Institute released a report (HERE) on measuring school performance. The main recommendation of the report is to replace measurement of average school performance with so-called value-added indices. The idea is very simple – to measure student progress as the primary outcome – and by employing an appropriate statistical model to extract that component of the improvement which can be attributed to the school.

Read the rest of this entry »

How do you detect election fraud? A recent article in the Washington Post describes a novel statistical idea. I wish I had thought of it.

Read the rest of this entry »

The graph below was provided by recent Nobel laureate Paul Krugman. It shows a geographic map of the swing from 2004 to 2008. Read the rest of this entry »

My colleague, Peter Cebon, has sent me the graphic below which shows a time series of temperature, CO2 and sea levels. It is actually a pretty good graphic. While it is true that one generally counsels against putting three plots, on different scales, on the same graph, I think that this one works pretty well, especially with the three axes at the right color coded to agree with the time series. The resolution is not that great so I suggest you download it in pdf form HERE.

On a recent ABC interview, a well known scientist made the following statement about global temperatures.

Actually, there has been cooling, if you take 1998 as your point of reference. If you take 2002 as your point of reference, then temperatures have plateaued. This is certainly not what you’d expect if carbon dioxide is driving temperature because carbon dioxide levels have been increasing but temperatures have actually been coming down over the last 10 years.

Is she right?

Read the rest of this entry »

Following on from the SuperCruncher’s post, I recently became aware of some interesting research on bias in job hiring practices. This is a field where a properly designed trial can reveal even unconscious bias, beyond any reasonable doubt.

Read the rest of this entry »

The inquest into Princess Diana’s death is a great wastes of public money. but in addition to fuel for the tabloids, it has supplied me with a nice little example of wrong statistical reasoning. The culprit is (of course!) an expert witness but not apparently an expert in thinking about uncertainty. The expert in question is a Dr. John Searle, a traffic engineer hired by the Ritz hotel which is owned by Dodi al Fayed’s father. In other words he has really been hired to support al Fayed’s conspiracy theory that Diana was murdered which requires him first to discredit the theory that Diana’s driver being pissed was the cause of the tragedy.

Read the rest of this entry »

Election polling is typically broken down by gender, age-group and city/country. It is well known folklore that the coalition enjoy their greatest support amongst women, country dwellers and older voters, but these effects have always been rather weak. At least weak compared to some other countries. In the US, voting intention is hugely dependent on ethnicity, black Americans being overwhelmingly democrat. I am unaware of a similar ethnic divide in Australia but then again I have never seen any data. I thought I would have a look at the last 11 years of Newspoll data and see if the demographic effects of gender, geography and age are becoming more or less apparent.

Read the rest of this entry »

Election polls typically ask respondents “who would you be most likely to vote for if an election were held this Saturday?” On the other hand, I always tell my MBA students that if they are trying to survey customer satisfaction they are better off asking for a response on a 1 to 5 or 1 to 7 scale, than a crude binary choice of satisfied/disatisfied. It seems to me that this applies equally to election polls.

Read the rest of this entry »