Fun and Misunderstanding with Statistics: Travel-to-work data

With all its faults* the 2011 National Household Survey (not the census) came out with some headline material: how people travel to work.  
 
It was an exercise in ‘statistics for fun and misunderstanding’ – the problem that comes when a small percentage change in a big number is interpreted as signifying little change at all.  Or when an average over a region is interpreted as applying to all its parts.
Driving is up in Metro Vancouver, although it’s falling slightly as a mode share (from 67.3 percent to 65.9 percent). There were 39,245 more drivers in Metro than in 2006, but the City of Vancouver saw 3,050 fewer drivers as driving fell from 51.5 percent to 48 percent. 
If you design your urban region for driving – which is what we did for most of the last century – then that’s what you get.  When you start to provide alternatives, you see the dominant mode decline as other ways of getting around are adopted by those for whom it is practical.   Travel behaviour doesn’t change dramatically because it takes a long time for development to occur and habits to be formed, or to change. 
 
Much better news across the region was transit use: it went up  from 165,000 to 213,000 in Metro – or 48,000 more transit users.  That’s more than half the extra people travelling to work in the region and quite a lot more than the extra car drivers – an important indicator of change.  A majority of the newcomers to the mix are taking transit. 
 Just under 18,000 of these additional transit riders were in the City of Vancouver, 30,000 in the rest of Metro – again, another example of how urban form, density and alternatives are reflected in the stats.  The City also saw 59 percent of the extra walking in the region, though the rise was small: from 12.2 percent to 12.5 percent – equal to 2,700 more people of the nearly 37,000 who walk to work in Vancouver.  
 
Cycling also went up a little bit in Metro, from 1.7 percent to 1.8 percent, but almost all that increase was in the City – 82 percent of the extra cyclists in Metro Vancouver were in the City of Vancouver (2,440 of the 2,960 extra cyclists). 
So a headline which says “Number of cyclists who commute unchanged from 2006 to 2011” is missing an important aspect of the story by reporting only the average.  And a small percentage increase in transit use from 16.5 percent to 19.7 percent actually reflects a big jump in the number of transit users, also magnified in the City.  
Imagine if that growth had been in car traffic on an already congested arterial network. Unfortunately, some will draw the conlusion that the transit investment hasn’t paid off, when very clearly it is heading in the right direction.  Indeed,transit use is already maxxing out  in some parts of the system, which will only be exacerbated by ‘system optimization’ – cutbacks by any other name – if there is no new  investment.
Here’s the raw data below the fold, and a qualifier on the data sources.
 

*In 2006 20 percent of the population was asked a long set of questions, and the answers were multiplied by five to give the overall data set. As over 95 percent of the people who were asked, answered, it’s pretty safe to assume that, for largish areas like the City of Vancouver, it’s pretty accurate data.
 
In 2011 about 30 percent of the population were asked to answer an almost identical long set of questions, but only about 70percent answered those questions and some of them didn’t answer all the questions. It was a voluntary survey, so there’s no real way of telling how representative the sample who answered are of the entire population. (It’s missing everyone who hated the census, and probably almost everyone who hated what the Government did to the Census, for example). Still, Statistics Canada multiplied by five (or whatever they needed) to get to a data set supposed to represent everybody.