New Census Report on Bicycle Commuting

[Harry Meatmotor]

One interesting thing to note about the ACS (from which that report is written) is that the ACS is run year-round, and there’s a good chance that the cycling to work population is underrepresented regionally. This is a difficult thing to deal with in the statistics because they represent such a small portion of the population, too; properly weighting 0.6% of a population is dicey. Unless you’re competing in the freezing saddles competition, when the survey asks you “did you bike to work last week?” in February and you answer “i haven’t been on a bike all month!”, the ACS is going to assume you rarely bike to work, year on year. Also, comparing the Census numbers (only captured in April) to the ACS is dicey for behaviors with seasonal/regional variability, especially with small populations. I hope somebody at national bike challenge is gearing up to produce some yummy stats from all the data they’re collecting!

tl;dr – those ACS numbers are probably a little low; actual numbers probably higher.

[cvcalhoun]

I also notice that Washington, DC went from 1.2% bicycle commuting in 2000 to 3.1% in 2008-2012. That’s way above the national average.

[KLizotte]

^^Yes, I agree with Harry that the ACS must undercount biking since it is probably not capturing all of the immigrants and low income residents that bike for transport. Many of these folks are undocumented or have undocumented people in their homes so do not tend to answer surveys, they are too darn tired to answer surveys from working multiple jobs, and there is also the issue of language barriers.

Go through a poor neighborhood and you will be amazed by the number of Target quality bikes parked in the area.

[dbb]

My demographer coworkers say that the ACS samples about 300K addresses per month (validated by the ACS website that says “1. How large is the sample? – The full implementation of the ACS, which began in 2005, originally sampled approximately 2.9 million housing unit addresses annually stateside. In 2011 the ACS sampled approximately 3.3 million housing unit addresses – this corresponds to an increase in the targeted annual sample size of 3.54 million addresses that began with the June 2011 ACS sample. The 2012 ACS sample is approximately 3.54 million housing unit addresses. The PRCS samples approximately 36,000 housing unit addresses each year in Puerto Rico.)

The entire survey is at http://www.census.gov/acs/www/Downloads/questionnaires/2012/Quest12.pdf

I will attempt to get the guys that have a relationship with Census to tweak the commute question. No promises they will do more than listen to me.

[Greenbelt]

Biggest problem I’ve heard mentioned with transit surveys is that they classify multimodal transport based on the leg with the longest distance. That’s fair in a way, but it undercounts the numbers of people who use their bikes to get partway to work by a lot.

For example, people who take Metro to Cabi, or bike to Metro, would be classified as Metro if the distance was longer (and people answered the question correctly).

So it’s wrong to cite their results as the percentage of people who biked to work. Instead, it’s the percentage of people who used their bikes for the longest leg of their commute. All those shorter legs by bikes don’t reckon into the statistics, but it’s a lot of people.

I think the question would give much more informative results if the respondents were allowed to check multiple boxes. Please ask your demographer friends about that!

[UnknownCyclist]

Not using a bicycle on the longest segment of one’s commute only indicates an unsophisticated understanding of Rule 5.

[cvcalhoun]

One of the issues, I suspect, is that allowing the checking of multiple boxes might seriously overstate the people who were using walking or bicycling as an alternative to driving. It would mean, for example, that someone who parked next door to his or her office building, and then walked over, would get counted as a walker. That would result in virtually 100% of commuters being walkers–and the statistic being meaningless.

@Greenbelt 84703 wrote:

Biggest problem I’ve heard mentioned with transit surveys is that they classify multimodal transport based on the leg with the longest distance. That’s fair in a way, but it undercounts the numbers of people who use their bikes to get partway to work by a lot.

For example, people who take Metro to Cabi, or bike to Metro, would be classified as Metro if the distance was longer (and people answered the question correctly).

So it’s wrong to cite their results as the percentage of people who biked to work. Instead, it’s the percentage of people who used their bikes for the longest leg of their commute. All those shorter legs by bikes don’t reckon into the statistics, but it’s a lot of people.

I think the question would give much more informative results if the respondents were allowed to check multiple boxes. Please ask your demographer friends about that!

[Harry Meatmotor]

In related news:

http://www.bikerumor.com/2014/05/08/new-strava-metro-feature-shows-government-how-cyclists-are-using-city-streets/

MOAR DATAs!

[Greenbelt]

@cvcalhoun 84726 wrote:

One of the issues, I suspect, is that allowing the checking of multiple boxes might seriously overstate the people who were using walking or bicycling as an alternative to driving. It would mean, for example, that someone who parked next door to his or her office building, and then walked over, would get counted as a walker. That would result in virtually 100% of commuters being walkers–and the statistic being meaningless.

Easy to handle, either by adding multi-modal question boxes (car, transit) or (bus, bike) or by just asking for information on commute legs longer than 1/2 mile. A metro ride followed by a 10 block walk counts as multimodal to me, but not just walking in from the parking lot or a block or two from the bus stop.

[lordofthemark]

@Greenbelt 84762 wrote:

Easy to handle, either by adding multi-modal question boxes (car, transit) or (bus, bike) or by just asking for information on commute legs longer than 1/2 mile. A metro ride followed by a 10 block walk counts as multimodal to me, but not just walking in from the parking lot or a block or two from the bus stop.

The number of questions on the ACS is limited (plenty of non-transport questions) to keep the total length manageable (to keep the response rate up) and IIUC there is fierce competition for that real estate. Could address that with a larger sample, and asking different questions to different respondents, but that would cost more $$.

[lordofthemark]

@KLizotte 84699 wrote:

^^Yes, I agree with Harry that the ACS must undercount biking since it is probably not capturing all of the immigrants and low income residents that bike for transport. Many of these folks are undocumented or have undocumented people in their homes so do not tend to answer surveys, they are too darn tired to answer surveys from working multiple jobs, and there is also the issue of language barriers.

Go through a poor neighborhood and you will be amazed by the number of Target quality bikes parked in the area.

You might want to take a look at this

http://usa.streetsblog.org/2014/05/08/low-income-americans-walk-and-bike-to-work-the-most/

which confirms what you have said, and what I have observed in NoVa. That low income people in fact bike a great deal, as do those who have not graduated high school, and hispanics generally. OTOH bike commuting is also higher among high income people and people with graduate/professional degrees.

Where is weakest is in the middle – folks intermediate in education and income, and among african americans generally.

I suspect that has something to do with status issues – until you get high on the income/education scale, adult biking for transportation is seen as a low status activity done by people who can’t afford a car, acquisition of which is a mark of status and adulthood. People with higher incomes and education either are more secure in their social status, have other ways to express it (not excluding $10,000 carbon fiber road bikes) or don’t care about social status (which some would say is just another way to say they are VERY secure in their social status)

I would also suggest that presents a challenge in expanding transportation cycling. It may be easier to gradually grow it down the income scale (IE from affluent professionals to somewhat less affluent college grads without advanced degrees) than to get more working class cyclists. Though thats frustrating because cycling can be such a good solution to the transportation problem for folks with less money.

I am not sure why there are so many more hispanic cyclists than blacks, but guess part of that is that many more low income hispanics live in suburban areas where there are few other good non-auto options, part is that at least some hispanics cannot get drivers licenses because of immigration issues, but I think it may also be cultural issues.

[brendan]

@lordofthemark 84790 wrote:

You might want to take a look at this

http://usa.streetsblog.org/2014/05/08/low-income-americans-walk-and-bike-to-work-the-most/

Love that the article is written by a friend I met group camping the Friday night before my first Seagull century.

B

[dbb]

I spoke with my economist coworkers about the ACS and they suggested a special tab that looked at the data by calendar quarter or monthly. They indicated that the Census did that after Katrina in the Gulf because the conditions changed so much after Aug 29 when the storm hit they really ended up with two sample sets.

I will continue to engage on this to see what I can coax out of Census.

[wheelswings]

Not sure if folks saw this piece in Vox. It includes some interesting charts, e.g. comparing ped and cyclist death rates in the US and other countries, and looking at the world’s top cities by cycling share of commuters.
http://www.vox.com/2014/5/16/5720334/the-bike-to-work-community-is-growing-fast-but-its-far-from

[chris_s]

Apparently the Census is considering REMOVING this question from the ACS! Data about how we get around is really important for biking and walking advocacy. Removing this question would be extremely detrimental to advocacy efforts for biking & walking. As a wise man once said, if you don’t count it, it doesn’t count.

Detailed information on taking action is available here from the Alliance for Biking & Walking.

[Author]

Posts