Great post Lloyd. As an architectural grad student specializing in sustainable design, this is an eternal struggle. The profession tends to want to function in the same way and solve our ecological problems with technology. I beg to ask, what is more efficient than not having to heat or cool 1000 extra square feet. Effective design has the ability to make smaller spaces more livable. While it may not be as sexy as the new energy technologies, a simple solution is to give people better "small" options. We must break from the traditional thought that equates consumption with wealth.

A similar, but longer term and potentially better correlated, trend is the popularity of central heating compared to obesity.

A friend of mine (a former head of a UK academic statistical advisory group) showed this relationship to me, and it's pretty startling.

please... I can't stand these correlations that are made out of two numbers that happen to increase at a similar rate when scaled to fit.

Im sure that if you compared global temperature increases, they are responsible for bigger houses. Or maybe house sizes are increasing because of pollution increases? It makes sense... people want more buffer between them selves and the bad air.

Im sure there is an indirect relation ship, but that like saying your friends with kevin bacon through six-degrees of separation.

Not to rain on the parade, but this graph is basically meaningless.

I could come up with nearly any correlation I want just by changing the (completely unrelated) y axes.

I'm not denying that America's desire for bigger and bigger things probably extends to our waistlines as well, just that a correlation is implied here by how well the graphs match up that is founded merely in scaling coefficients.

Perhaps that's why the graph author said "Keep in mind that this is pure amusement and not terribly scientific any any way."

Whereas I'm thinking, "Hey, if I want to lose weight, I should build a small house!" :D

I agree with Mike, this is a completely useless graph, quite a shock to look at initially, but overall, meaningless. I could have just as well correlated the rise in obesity with a rise in the consumption of organic goods. If we use this type of tactic, there's absolutely no reason we can't expect a backlash.

Correlation is not causality. I do think that this graph says _something_ though.

http://en.wikipedia.org/wiki/Correlation_implies_causation_(logical_fallacy)

Wow, a 200 square foot increase? That's insane in the membrane!

Besides the whole correlation-doesn't-prove-causality thing, the big reason why this graph does not mean much is that the two axes are scaled differently.

What the creator of the graph did is simply take two data sets with linear trends (housing size, obesity rate), and by carefully choosing the y-axes, scaled their slopes such that two trends appear to overlap.

It is interesting, though, that both of these things have monotonically increased for about a century...

A better graph might show the growth of people and houses as a percentage.

maybe that would show if there is a correct correlation such as for every 1% increase in house size the average person grows .5% or something like that.