In epidemiological research Wing (1981) found that among people with high-functioning autism or Asperger syndrome there were as many as fifteen times as many males as females. On the other hand, when she looked at people with learning difficulties as well as autism the ratio of boys to girls was closer to 2:1. This would suggest that, while females are less likely to develop autism, when they do they are more severely impaired.

Attwood (2000), Ehlers and Gillberg (1993) and Wing (1981) have all speculated that many girls with Asperger syndrome are never referred for diagnosis, and so are simply missing from statistics. This might be because the diagnostic criteria for Asperger syndrome are based on the behavioural characteristics of boys, who are often more noticeably “different” or disruptive than girls with the same underlying deficits. Girls with Asperger syndrome may be better at masking their difficulties in order to fit in with their peers, and in general have a more even profile of social skills.

So, as usual, it’s “not quite that simple,” apparently. Here’s the link to that page:

http://www.autism.org.uk/about-autism/autism-and-asperger-syndrome-an-introduction/gender-and-autism/autism-why-do-more-boys-than-girls-develop-it.aspx

That sort of anti-Piratism should be a thing of the dim and unenlightened past: the link between the near-extinction of pirates and the rise in global temperatures is a thing that transcends statistical analysis.

One need only look at what has happened since our brave pirate brethren in Somalia began their careers as modenr privateers: global warming stopped (about 15 years ago), thus enabling us to rejoice at the failure of the 2000 prediction (by David Viner of CRU) that “Children just aren’t going to know what snow is” (implying a global absence of snow… and probably of pirates too). That statement was reported in the Independent of March 20th 2000.

Thankfully, UK kiddies have been enjoying abundant snow thanks to the unsung efforts of our Somali chums.

TL;DR: climate models are statistically less robust – and follow more of a “do what’s required to get the answer we want” paradigm – than any economic model of which I am aware. Aaargh, me hearties.

Generally, a phenomenon related to population will, in equilbrium, grow broadly in line with population. (e.g., long-run behaviour of the labour force). Things that grow in line with aggregate expenditure will – again, as a measure of tendency rather than in every observation – grow at “m-n-beta” (the rate of growth of the money supply, less population growth, less technological change). which is the theoretical equilibrium rate of growth of nominal quantities.

If you have two quantities that are both growing exponentially (even at divergent rates), you will always find an affine transformation of one quantity that will generate a large correlation coefficient (and therefore a relatively large coefficient of determination).

If the two series are growing linearly, the transformation is trivial (multiply either series by the ratio of the slopes) and the coefficient of determination will approach unity… it won’t actually BE 1 unless both series exhibit (a) zero; or (b) perfectly correlated; noise around their respective linear growth paths.

This is why trained statisticians (and econometricians – of which I am one) test data for stationarity, make sure that regressors are of the same order of integration as the dependent variable (so that the regression has a stable steady state), and if required take differences (or even log differences) of data to reduce the regression to one involving only stationary quantities.

TL;DR: statistics is sufficiently complex that a layman or autodidact is likely to get it wrong in ways that aren’t amusing to anybody who finished high school with decent marks in maths.

Is your argument based merely on profound ignorance, or is it the bad faith indicated by your link to whale.to? For example, someone that knew what they were talking about would be aware that the introduction of antibiotics effective against tuberculosis in the early 50s drastically reduced death rates from that disease. Likewise, the availability of effective antibiotics and improved supportive care reduced the death rates from diseases like pertussis and measles without having much of any effect on their incidence. Plenty of children who would have died in earlier years ended up permanently disabled instead (one of my great-aunts among them).

OLS is efficient and consistent (and hence unbiased – hence BLUE) only when the Gauss-Markov conditions hold, and they will NEVER hold for anything growing geometrically… at the very least there is a problem with autocorrelation, and almost certainly heteroscedasticity.

And as we all know, the coefficient of determination (R^2) is the square of the Pearson correlation coefficient – whichin turn  is cov(x,y)/var(x).var(y) (why can’t we do LaTeX up in this bitch, yo?); using R^2 as a goodness of fit measure assumes that OLS is BLUE for the problem under consideration.

It is highly unlikely that two I(1) or I(2) series will have stable variances.

Model that shit with ARIMA and the correlation will disappear, I bet.

People who are dead do not eat and make no choice.

People who choose not to eat are alive.

It does appear that the act of choosing keeps us alive Or…. :)

MrEnergyCzar

I like starting that estimation pretty high, even though it’s wrong most of the time, just because I think it stops me from underestimating people and missing opportunities. And it helps me keep a better attitude. But I know what you mean.

Sometimes people produce trend charts like this with closely matched lines, and then point to the correlation and blame one of the trends for the other.  Such as the autism/vaccination correlation mentioned above.  Or the “violence in schools/video game sales” correlation that is currently all the rage.  Just because two trends have similar curves doesn’t mean one is causing the other.

See also this chart implying that global warming is caused by a lack of pirates:  http://www.venganza.org/about/open-letter/

Today’s science doesn’t really blame things like food additives for autism, but when you’re a parent of an autistic child and searching for answers, it’s certainly one more thing to try.