Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 51
51 ENDNOTES i In theory, this task could assume proportions considerably more the adults who completed a bicycle trip during their diary day (a total complex and sophisticated not pursued in this application for sev- of 138 throughout the seven county area), 86 of them (62%) were eral reasons. Some literature attempts to relate the amount of bicy- from the Minneapolis or St. Paul. cling to facility measurement, but with very limited accuracy. Other We also only included the population who reported having com- literature discusses possible ways of describing the bicycle envi- pleted any type of travel during their assigned travel diary day, a pro- ronment but does not relate this empirically to the amount of bicy- cedure consistent with other transportation-related research (177. cling in an empirical measurement. The team believed that a frame- Zahavi, Y. and J. M. Ryan, Stability of Travel Components Over work that required planners to evaluate the detailed bicycling Time. In Transportation Research Record 750, TRB, National environment of their area would be exceedingly complex and labor Research Council, Washington, D.C., 1980, p. 1926). Of the orig- intensive for the planners, and difficult for us to explain in an unam- inal 1,801 individuals, 148 individuals (8.2%) took no trips on the biguous way that would be easy to apply. Even conceptually sim- travel diary day and were thus excluded. This left us with an effec- ple measures such as miles of off-road bike paths proved hard to tive sample size of 1,653 subjects (91.8% of our original sample). define in a way that would allow comparability across different loca- The 148 subjects that were excluded were not significantly differ- tions. Local subtleties of design and location can be vitally impor- ent from subjects retained for analysis with respect to the likelihood tant in issues that might arise in different places, let alone place val- of living in a household with kids or living in Minneapolis. How- ues on them that could be used to develop an overall environment ever, compared with excluded subjects, included subjects were rating. more likely to be employed (83% vs. 41%, p < 0.001); more likely ii The amount of data needed for bicycle forecasting using these to have a college education (64% vs. 20%, p < 0.001) and more methods is actually much larger than for an auto forecasting model likely to be male (48% vs. 35%, p = 0.002). Included subjects were for two reasons. First, because so few people ride bikes in any short also less likely to live in households with an annual income less than period, relative to the number that drive cars. To have a large $50,000 (36% vs. 56%, p < 0.001) and less likely to be over 60 years enough sample of bicyclists to have some sense of their geographic of age (15% vs. 37%, p < 0.001). ix distribution would require a very large survey; of the roughly Such data was obtained from 2001 employment records of the Min- 10,000 adults in the Twin Cities Travel Behavior Inventory, only nesota Department of Employment and Economic Development. x about 200 rode a bike on their survey day, not nearly enough to esti- When measuring this dimension, it is important to measure the mate a regional model. Second, because the number of factors that diversity of different types of retail establishments while controlling might influence bicycling decisions is much larger than for car or for the potential disproportionate drawing power of larger establish- transit travel, which are typically just assumed to rely on travel time ments (e.g., a large clothing store offers high employment but little and monetary cost. By contrast, highly subjective factors such as diversity). The upper limit is set at businesses containing more than perceived safety and pleasantness of the riding environment could 200 employees and the number of employees for each area is tallied. play major roles in bicycling decisions. The final measure is the number of employees within the "neighbor- iii In addition, it may in fact be more useful than a more complex hood retail" subset within 1,600 meters of each home location. xi model that may be slightly more accurate, but that given the issues These include all businesses in the following NAICS categories: discussed may also have only the appearance of accuracy. Food and Beverage Stores (e.g., grocery, supermarket, conve- iv Based on an analysis of several sources (e.g. travel dairies such nience, meat, fish, specialty, alcohol) as the National Household Travel Survey, direct questionnaires Health and Personal Care (e.g., pharmacy, drug store) administered by the Bureau of Transportation Statistics), the team Clothing and Clothing Accessory Stores (e.g., shoe, jewelry, project's that approximately 3% of the U.S. population cycles one luggage) day per week and an estimated one percent of the U.S. population Sporting Goods, Hobby (e.g., needlepoint, musical instrument), receives their recommended weekly level of physical activity by Book, and Music Stores cycling. General Merchandise Stores (e.g., includes department stores) v A quick look at the data shows that 69% of our adult subjects within Miscellaneous Store Retailers (e.g., florists, novelty, used mer- Minneapolis and St. Paul have a retail location within 400 meters of chandise, pet, art, tobacco) their home. Food Services and Drinking Places (e.g., restaurants) vi xii Households were recruited to participate in the TBI using a strat- The principal reason from these breakpoints was to ensure adequate ified sampling design. Telephone interviews were used to collect distribution across each category. For example, 32 percent of the both household and individual socioeconomic and demographic households had a retail establishment within 200 m, 37 percent within data. Subsequent to the demographic interview, households were 400 m, 21 percent within 600, and 10 percent for the remaining. xiii assigned a travel day on which 24-hour travel diaries were com- A continuous measure assumes that for each additional meter of pleted for all household members five years or older. distance closer/farther there is a consistent incremental increase/ vii Home phone call interview information helped ensure the relia- decrease in the odds of bike or walk use. xiv bility of these self reported measures of walking and/or cycling. One potential disadvantage is that by subclassifying into cate- viii The sample was restricted to residents of Minneapolis and St. Paul gories, a strong homogeneity assumption is imposed. That is, the primarily because theses two cities--as opposed to the suburbs-- team assumes that the effects are the same for everyone within a had adequate representations of walking and cycling behavior. Of given category regardless of their individual proximity to a bike
OCR for page 52
52 trail. For example, the effect of living 400 meters from a bike trail xx As mentioned, this analysis only included residents from is no different than living 799 meters from a bike trail. However, Minneapolis and St. Paul. Most other communities within the given that the increments are within roughly 400 meter, there is rel- metropolitan area are more suburban in nature in terms of lower den- atively little difference, if any. sity, lower accessibility, and other related urban form features. xv Outside the downtown core of each city (for which there are very xxi To be technically correct, sampling weights should have been few respondents in the TBI), there is similar housing density spread employed. Given the secondary nature of the analysis and the fact across mostly all neighborhoods. that a sub-sample was selected, proper survey sampling weights xvi These 86 cyclists completed between 1 and 10 bike trips on the were not available. assigned travel day (mean = 2.9, SD = 1.79). For 73 of these cyclists xxii Our sample began with 42,750 records. Geocoding and remov- (85%) the total distance traveled by bicycle is also calculated, which ing records with missing or unreasonable data (e.g., homes with ranged from 0.74 km to 36.71 km (mean = 8.64, SD = 7.10). As zero bathrooms, zero square feet, or built before 1800) reduced expected, the proportion of bikers varied across levels of bike facil- our sample to 35,002. The relatively small number of records ity proximity, with more bikers living closer to bike trails and fewer removed still provided an even distribution of home sales across bikers living further from bike trails. Of interest, these distributions the metro area. xxiii differed depending on which measure of bicycle facility proximity Active open spaces are primarily used for recreation, and consist was used. In other words, the distribution of cyclists across cate- of neighborhood parks and some regional parks. Passive open gories of proximity to any bike facility was not statistically signifi- spaces are less accessible on foot. They include areas such as golf cant, nor was the distribution of bikers across categories of prox- courses, cemeteries, and large regional parks that are accessible imity to an off-street facility. However, the distribution of cyclists only through designated entrance points and often only by car. xxiv across categories of proximity to an on-street facility was statisti- Open space and bicycle variable names are prefixed by a c for cally significantly different, with increasing proportions of cyclists city and s for suburb. in the hypothesized direction (chi-square = 13.42; p = 0.004). xxv In Minneapolis, several of the streets in the downtown core have xvii Our definition of "walkers" did not include people who only bicycle lanes (although there are few home sales downtown). Most reported a walk trip from a different location (e.g., work or other). other on-street bicycle lanes are on busy commuting arterials or Individuals who only reported such walk trips are not included in around the University of Minnesota commercial district. On-street an effort to more cleanly identify correlations between the residen- lanes in St. Paul are a different story. They tend to be along a well tial environment and walking. maintained boulevard-type corridor (Summit Avenue) and the Mis- xviii This in turn can lead to an increase in the Type I error rate; that is, sissippi River corridor. These counteracting effects between Min- finding a statistically significant effect, when in fact there is none. neapolis and St. Paul may possibly cancel out one other. xix xxvi Some respondents may be pursuing walk trips from work or other The median sale prices in the city and suburbs for 2001 were types of locations. Only walk trips from home were considered. An $148,475 and $184,500, respectively. No significant relationship additional 137 people report having a completed a walk trip, how- was found between home prices in the city and proximity to on- ever, none of the walk trips they reported were from home. street bicycle lanes, so no effect is estimated in Table 27.