Blog
Bicycling beats the odds — National bike commuter rate holds steady
Despite predictions that the number of Americans biking to work would fall after gas prices returned to ‘normal’ in 2009, the percentage has held steady at 0.55 percent. The Bicycle Friendly Communities (BFC) among the 70 largest cities also held steady with a 1 percent increase, while non-BFCs increased their commuter share by 26 percent. BFC cities still have on average about twice the percentage of commuters as non-BFC cities.
Check out the bike commuter rates for the 70 largest US Cities for 2009 (PDF). Or see the online spreadsheet (new).
And see the bike commuter rates from 2000 to 2009.
NEW! See the commuter rates for Biking, Walking, Public Transportation, and Driving alone for 2000 to 2009.
The numbers on these tables are estimates produced by the US Census Bureau based on samples of the population. There can be a large margin of error, especially for the smaller cities. See below for a discussion of the limitations of the ACS methodology.
About the American Community Survey and its limitations
The Census Bureau has released American Community Survey data for over 6,600 geographic areas that meet the 65,000 population threshold. Population sizes are based on July 1, 2009 population estimates from the Census Bureau’s Population Estimates Program, for most legal geographic areas.
The American Community Survey is the country’s largest household survey with an annual sample size of about 3 million addresses. The survey uses questionnaires and interviews to gather information on demographic, economic, and housing characteristics, including journey to work information. Annual estimates are available for geographic areas with populations of 65,000 or more, although because of limited sample sizes, bike commuting estimates are not available for all of those locations. The ACS replaces the Census Long Form questionnaire, which was given to one in six Census-takers every ten years.
Using the decennial Census and the annual American Community Surveys, the U.S. Census bureau has asked the same question about means of travel to work over time, making it is one of the best sources for tracking trends in bicycling to work levels. However, there are several limitations to using the ACS as a measure of bicycling levels. Bicycling’s share of all trips is nearly three time large than bicycling’s share of commuter travel — the 2000 Census estimated that 0.34 percent of American workers usually bike to work, in contrast, the 2001 National Household Transportation Survey (NHTS) estimated that 0.9 percent of all trips were made by bicycle — therefore, the ACS bicycle commuter percentage should not be interpreted as equivalent to the proportion of all trips.
In addition, the ACS and the decennial Census undercount bicycle commuting levels. They ask for the principal mode of travel the worker usually used to get from home to work during the previous week.
Workers were asked to list only the means of transportation they used on the largest number of days in that week. This means that if the respondent rode a bicycle to work two days but drove three, they would not be counted as a cyclist. Likewise, workers were asked only for the means of transportation used for the longest distance during the trips. If someone biked one mile to a bus stop and rode the bus for two miles they would not be recorded as a bicyclist.
Finally, it is important to note that the results of the ACS (and the Census long form) are only estimates based on population samples. The ACS releases an estimate of the number of workers 16 years old and over and the estimated number of workers who used each mode (drive alone, car-pool, public transportation, walk, bike, taxi, or motorcycle). The estimates are used to calculate the share of workers using each mode. Along with these estimates, the ACS publishes the margin of error, a range within which they can be 90 percent confident that the true number falls. For example, the in 2008 ACS estimated with 90 percent confidence that the number of bicycle commuters in New York City was between 21,162 and 27,694. The given estimate is 24,428 with a +/- range of 3,266. For communities with low counts of bicycle commuters this range can be quite large, in few cases the +/- range can be as large as the estimated number of bike commuters. Plano, Texas, for example, has an estimate of 230 and a +/- range of 235, meaning the ACS estimates with 90 percent confidence that the actual number of bike commuters is between zero (since negative five is impossible) and 465. For simplicity’s sake, these tables use the mid-point in the range, but when reading the tables, keep in mind that all percentages are in fact just estimates, some with large margins of error.
Notes:
- As stated above, the ACS numbers are estimates – differences among years or cities may not be statistically significant.
- These numbers are based on the “principal city,” not the larger Metropolitan Statistical Area (MSA).
- Population numbers in the tables have been updated to reflect the Census Bureau population estimates.
Darren Flusche
League Policy Director
Flusche joined the League in April 2009 and has a B.A. in history from Syracuse University and a Masters of Public Administration with a concentration in public policy analysis from New York University.
Subscribe
September 28th, 2010 at 12:18 pm
While fuel prices game down, transit also saw service cuts. Some current bike commuters are likely former transit riders.
For shorter trips, it’s faster to ride a bike, even slowly on the sidewalk, than wait 30-60 minutes for a bus.
September 28th, 2010 at 2:36 pm
Is there a way to get this in spreadsheet form, for sorting?
September 28th, 2010 at 4:04 pm
When you live in the East Side of Milwaukee like I do, you find car ownership rather impractical, Only the most hardheaded motorist will keep a car here in spite of being ticketed and towed several times. I call this Milwaukee’s dirty little racket.
September 28th, 2010 at 4:38 pm
Darren, Good post!
I believe your population rankings on the LAB commuter sheet are off. Detroit is ranked 15th below cities with smaller estimated populations (e.g. Austin). Note that the ACS made a huge population correction for Detroit in 2009 and added 131,000 people. That may explain the rankings issue.
September 28th, 2010 at 4:46 pm
Hi Todd,
Yeah. ACS population estimates are inferior to the Census adjustments. I just used the 2008 ACS population rankings when I did it last year (for convenience), and left it again to avoid introducing new cities and excluding old cities. That’s why I made sure to label the column accordingly.
This does create an issue for Detroit. If I get ambitious next year, I might redo the entire list for every year since 2000 using 2010 official Census population estimates. I just didn’t have it in me today.
September 28th, 2010 at 5:54 pm
[...] League of America Bicyclists just released 2009 Census estimates for bicycle commuting, and I’m delighted to say Kansas City has jumped several spots above last place. Back in 2005 [...]
September 28th, 2010 at 9:33 pm
[...] the U.S. Census Bureau’s 2009 American Community Survey were released today. According to the League of American Bicyclists’ analysis of the data, national bike commuting rates held steady at 0.55%. From the LAB [...]
September 29th, 2010 at 10:31 am
[...] LAB’s blog also noted some limitations about the survey. In addition, the ACS and the decennial Census undercount bicycle commuting [...]
September 29th, 2010 at 12:04 pm
[...] the LAB’s Blog, analyst Darren Flusche notes that in the past year, cities not known for cycling amenities saw the [...]
September 29th, 2010 at 12:25 pm
[...] the LAB’s Blog, analyst Darren Flusche notes that in the past year, (Read [...]
September 30th, 2010 at 12:09 am
While still low, in Dallas we tripled our commuter rate! Let’s see if we can improve it next year again!
September 30th, 2010 at 10:59 am
Todd,
Good news. I was able to update the population rank on the sheet without losing any of the cities previously listed. Detroit is now at its proper place for population, #11.
October 5th, 2010 at 1:47 pm
Thanks for collecting the data for us. The links for the spreadsheets do not work for me.
By “Census adjustments” what do you mean precisely? Are you talking about some re-weighting scheme based on the demongraphic population estimates or the 2010 Decennial Census? I’m not sure whether the demographic estimates are closer to “truth” since the model is essentially additive. That is, they start with the 2000 Decennial Census estimate, add/subtract the number of births/deaths according to birth/death certificates, modify counts by tabbing mobility based on tax returns, and so on. So the 2009 estimate has a lot more error than say the 2001 estimate.
ACS is a weird animal. It samples households over the entire year and, for small population centers, aggregates data across years to form an estimate. Now in this case, it appears that only municipalities large enough for single-year estimates are used but you still have an issue that some people within these geographic areas were asked in the winter while others asked in the summer. This would increase the across-year variance of the estimate, IMO.
Anyway, if you think that the variances are important, you might want to report them with the 2008 and 2009 figures. Mind you, since you are calculating a share of commuters, the numerator and denomator are likely correlated and consequently the variance of the estimate is likely smaller than the number of bike commuter estimate used in your example. It also occurs to me that numbers might better measure certain advocacy goals than commuter share. But it is easy for me to create work for you!
October 5th, 2010 at 3:12 pm
Hi Geof,
When the 2008 ACS came out, they recommended that researchers use the Census Bureau population estimates (http://factfinder.census.gov/servlet/DatasetMainPageServlet?_program=PEP) rather than the ACS population estimates.Your comments about the estimates are right on. But despite the fact that they are still estimates, the population estimates in the tables now are better than what was there before.
You raise another interesting topic: season. As you note, the ACS surveys people year-round. They do so evenly across months. This means that season should impact any given city the same across years. That is, a roughly equal proportion of a city’s bike commuters will be surveyed in winter, summer, fall, and spring each year, making the numbers comparable across years. Now, season as a variable is more important in some cities than it is in others. Survey respondents in year-round temperate cities will more like say they rode a bike to work the previous week during all 12 months than cities with more extreme weather. What’s fascinating is the number of cold-climate cities that have high commuter shares nonetheless, according the ACS.
October 5th, 2010 at 11:30 pm
[...] League of American Bicyclists dove into the numbers and found bike commuting in the 70 largest cities (.pdf) in the United States climbed 35 percent [...]
October 6th, 2010 at 4:44 pm
[...] League of American Bicyclists dove into the numbers and found bike commuting in the 70 largest cities (.pdf) in the United States climbed 35 percent [...]
October 7th, 2010 at 1:39 pm
[...] League of American Bicyclists has crunched the numbers on the 70 largest US cities, and has concluded that nationwide bicycle commuting rates have held [...]
October 8th, 2010 at 11:02 am
I’m a Minneapolis cyclist who completed that questionnaire in January 2009. I almost never drive to work from March through November, but it was by a narrow margin that I biked more than half of the days for that week in January.
October 8th, 2010 at 12:40 pm
Could you please explain this statement? “The ACS replaces the Census Long Form questionnaire, which was given to one in six Census-takers every ten years.” Does this mean the 2010 census did not have a long form? or that the ACS attempts to update the long form in the intervening years between the census years? Also i recall (the one and only time i got a long form in 2000) the long form asking for commute mode for a week in APRIL and some of the comments implies the ACS was for a week in JANUARY. That also suppresses the bike commute rate.
October 8th, 2010 at 1:03 pm
The Long Form no longer exists. It has been replaced by the ACS.
Your memory serves. The Long Form asked about April, the ACS is rolling, year-round; it asks respondents about their primary mode of transportation “last week.”
October 11th, 2010 at 5:23 pm
The huge increase in bike commuting in New Orleans is happening for a variety of reasons.
First, the city is taking every opportunity to increase the bicycle infrastructure network. Where there was once just 7 miles of disconnected and poorly maintained bike paths and lanes, New Orleans now has over 30 miles and another 14 are coming online.
Second, the state of Louisiana recently passed a law mandating that cars must maintain a distance of no less than 3′ from cyclists. Passage of the law was followed by a massive awareness campaign which informed all while providing cyclists with a sense of safety and belonging on the road.
Third, concerns over high gas prices, growing traffic congestion and pollution are all driving more commuters to seek practical alternatives to driving cars in the city.
Forth, the city’s mass transit system was decimated by Hurricane Katrina and is just a shadow of it’s former self; yet, it is expected to serve the same city footprint, leading to infrequent schedules & inconvenient levels of service on much of the system.
October 22nd, 2010 at 8:52 am
[...] 10. Bicycling beats the odds — National bike commuter rate holds steady Despite predictions that the number of Americans biking to work would fall after gas prices returned to ‘normal’ in 2009, the percentage has held steady at 0.55 percent. The Bicycle Friendly Communities (BFC) among the 70 largest cities also held steady with a 1 percent increase, while non-BFCs increased their commuter share by 26 percent. BFC cities still have on average about twice the percentage of commuters as non-BFC cities. [...]
November 30th, 2010 at 11:16 am
[...] an automobile), make the roads safer, and help keep the air cleaner is a mystery to me. But with more cyclists on the road and more bicycle advocacy, this can change. Recent news from New York City is that some wrongfully [...]
March 9th, 2011 at 1:58 am
I like your website. Do you have mor informations which you can post? Thanks for the article.
May 25th, 2011 at 10:20 pm
Thanks for your post. I would like to say a health insurance brokerage also utilizes the benefit of coordinators of a group insurance policies. The health insurance professional is given a directory of benefits needed by anyone or a group coordinator. What any broker really does is hunt for individuals or coordinators which often best complement those needs. Then he gifts his advice and if both sides agree, this broker formulates an agreement between the two parties.
August 24th, 2011 at 6:56 pm
do it yourself debt relief…
Bikeleague.org Blog » Blog Archive » Bicycling beats the odds — National bike commuter rate holds steady…