## less-than-average illustration

Posted in Books, Statistics with tags , , , , on June 7, 2013 by xi'an

This Sunday morning, I was reading the NYT when I came upon this picture illustrating Stephanie Koontz’s tribune “when  numbers mislead“, a rather dull summary of an unreferenced and impossible-to-google paper called “The Trouble With Averages” [what about Quételet?!] dealing with the impact of marriage(s) on happiness (and vice-versa). [Funny enough, Andrew was telling me about this economist from Warwick working on happiness just yesterday night!] I however wanted mostly to point out how a-statistical this picture was, from a meaningless Venn diagram (reminding me of Templeton!) to the notion of opposing average and variation, to outliers standing in the wrong place (intersection of whatever!). [I do not think it is relevant to stress the innumeracy revealed by the column and its title!]

## new significance (out)

Posted in Statistics, University life with tags , , , on July 8, 2012 by xi'an

I have just received the latest issue of significance (June 2012) and there are plenty of interesting articles in it (with no horror story as in the latest issue!). From the cover story about finding emperor penguin colonies on satellite images via guano stains (large scale!, with a terrific and terrifying extract from Mawson’s journal) to “moral maps” à la Quételet, to teaching statistics as seen by the young statisticians section (of the RSS), to Tony O’Hagan favourite formul

$var(X) = E[var(X|Y)]+var(E[X|Y])$

(where he curiously fails to mention Pythagoras, which is how I justify the formula to my students), to the inappropriateness of using hand X-rays to determine whether Indonesian smugglers are under age or not. The less convincing section is obviously the “controversy” one, where the authors make a mechanistic proposal to bypass the drawbacks of p-values and Type I error, without contemplating the ultimate uses of tests…. Very pleasant read (I could have kept for the looong flight to Australia…)

## Quételet’s average (wo)man

Posted in Statistics, Travel, University life with tags , , , , on October 15, 2010 by xi'an

In the EuroStar last morning, I got attracted by my neighbour headline’s about average and went to pick a spare copy at the head of the train:

The article is awfully written, mixing nonsensical figures (like “the average British woman, who is 40, enjoys an extra half-day work each week”!!!) with a few remarks about the Office for National Statistics (ONS). This is supposed to celebrate the first World Statistics Day but there is very little to celebrate if this is the vision of statistics this day promotes. Are we back to Quételet‘s quest for the “average man”?!

## The “average” Australian

Posted in Statistics, Travel with tags , , on September 19, 2009 by xi'an

In the July-Sept. issue of Australian Geographic, there was a story about the Facing Australia project. This is interesting from a statistical point of view since the self-description of this experiment is

“Facing Australia creates composite male and female portraits based on current Australian Bureau of Statistics (ABS) census data. This census data is examined, establishing the proportionate age and ethnic profile of a chosen community. Individuals from the selected community are invited to participate by having their portrait taken. These images are carefully layered to create a composite portrait of the census-determined male and female.”

Thus, those pictures are somehow representative of an “average” member of the community, in the spirit of the ideal man of the Belgian 19th Century statistician Adolphe Quételet. The Facing Australia website does not detail how the pictures are added, as in the Melbourne Football club example below. I presume they use a recentring (plus  a rescaling?) algorithm to keep the eyes focussed as well as the features of the face recognisable. Otherwise, the outcome would be too blury to make sense. The website does not either explain on which statistical basis the 50 women and the 50 men are chosen within communities. (In the case of the Melbourne Football club, all players were added.)In the end, I do wonder about the statistical relevance of the exercise even though I quite appreciate the sociological and artistic aspects. The faces obtained at different locations are after all very close (see, e.g., the series of Charles Stunt University students) and do not deliver a message about the specificity of the community thus “averaged”. In addition, some features do not add to a “human” result (as already known at the time of Quételet). For instance, adding female and male photographs or 20 year old and 70 year old photographs does not produce anything meaningfull… And the processing algorithms must erase some variability like height or bulk (even though the above picture shows a range of shoulder widths in the football players).