Did the Census report on bike and walk commuting actually downplay the data?

Many other sources have noted that the Census bureau released a report on the rise of biking and walking to work over the last decade. 

Despite showing that bike comuting is up 255% over that time and that DC was 7th best it the nation, it might actually downplay the results here in DC. For it's start point it uses the 1-year data from the 2000 Census (1.2%) and for it's end data the 5-year data from the 2008-2012 American Community Survey (3.1%).

But if instead it had looked at the 2012 1-year data DC's bike commute rate would jump to 4.1%, making for a 340% increase and putting DC in a tie for 3rd (with Seattle). 

 

Comments

[Early Man]

The switch from the Decennial Census to the ACS for information on previous long-form items like commuting creates some problems reporting change over the decade. The long-form was a 1 in 6 sample whereas a single year of ACS is a 1 in 40 sample. Pooling 5 years of ACS data gives you something close to the precision of the long-form estimates. So this comparison is closer to apples to apples. If people start making arguments about decline or growth on the basis of 1 year ACS estimates they are going to be arguing mostly about sampling error.

[ACyclistInTheSuburbs]

Well, I think its a reasonable Bayesian prior that biking to work in DC increased between 2008 and 2012. It makes sense that Census used the five year pool to get equivalent sample size, but in terms of what I personally believe is true, I think WashCycle is correct - 4.1 is far more likely to be the actual correct number for DC in 2012 than 3.1 is. I cannot speak to the other cities though, so I am less sure of the rank.

[washcycle]

I get the sampling size argument, but a report issued in 2014, that includes a reference to 2008-2012 makes it seem like the endpoint on this comparison is much closer to today than to 2010. If they said that biking increased to 3.1% from 2000 to 2010 and then included a footnote as to where this data came from and why it was used, and perhaps a note that more recent - if less reliable - data exists that shows even higher mode share, I think that would be better.

[Grammar Geek]

Grammar geek says - "its" not "it's".

Strunk & White Rule #1 - "A common error is to write it's for its, or vice versa. the first is a contraction meaning 'it is.' The second is a possessive."

[washcycle]

Thanks. I know the difference.

It's called a "homophone substitution error". Basically, the theory is that your brain assigns a spelling to each word sound and then when a word is called, it defaults to the "primary" version of that sound. It takes much more effort to use one of the secondary versions. They aren't really sure why it happens, but they do know that the better you are at math, the more likely you are to make these kinds of errors.

People who are most bothered by it tend not to make the errors.

In the end, I've asked people to just realize that I'm not going to edit my own writing for these kinds of errors because I don't really have time to do so and to just go with it. You understood what I was saying, and that is all that really matters. Neither your third grade teacher nor mine is going to grade us.

[Crickey7]

Dammit, Jim, I'm a blog curator, not an editor!

[Will]

Just from an anecdotal case that we know is fairly significant, bikeshare didn't launch until September 2010, and didn't really hit its stride until Summer 2011, so if only 2011 and 2012 have meaningful bikeshare trips, and it's a 5 year average, it's certainly more likely to be the 4.1% figure. More importantly, if you look at the close-in neighborhoods, you find really high rates, as high as 13% for Adams Morgan if I'm remembering the last MWCOG survey on this.