Women’s Issues in Transportation: Summary of the 4th International Conference, Volume 2: Technical Papers (2011)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

203 Gender Differences in Adolescent Travel to School Exploring the Links with Physical Activity and Health Kelly J. Clifton, Portland State University gulsah Akar, Ohio State University Andrea Livi Smith, University of Mary Washington Carolyn C. Voorhees, University of Maryland This paper investigates gender differences in the asso- ciations between adolescent mode choices and travel patterns for the trip to school and levels of physi- cal activity. Analysis relies on cross-sectional data collected from adolescents (N = 269) in Baltimore City for a school-based recruitment study of physi- cal activity and the built environment. Participants were recruited from two magnet high schools, com- prising a predominantly African-American sample (67%) with geographically disperse home locations. Data analyzed here for each individual include (a) a web-based survey that collects background informa- tion, attitudes, perceptions, and recall behavior, (b) week-long physical activity data collected by accel- erometers, (c) a week-long travel diary, (d) archived spatial data about the built environment around each student’s primary home address, and (e) height and weight measurements, used to calculate body mass index (BMI). Multinomial logit models of the primary mode to school were estimated separately for males and females. Results show key differences in the fac- tors associated with their travel choices. Aggregate lev- els of physical activity during the weekday for transit commuters and auto commuters were compared for males and females. Results show significantly higher levels of physical activity for both male and female students who commute by transit, although male physical activity levels were higher on average. There were no significant differences in BMI between transit and auto commuters by gender, however. With rising rates of childhood obesity, research-ers are increasingly alerted to potential ways to increase physical activity for youth through changes in their transportation patterns, including increasing walking to and from school (1, 2). Most of the research in this area focuses on elementary school- age children, and policies have been implemented to encourage more walking, such as Safe Routes to School and similar programs (3, 4). Considerably less attention has been given to adolescents, their travel choices, and health outcomes. Likewise, the role that gender plays in the behavior of this group is also ripe for investigation because efforts to encourage changes in transportation and physical activity behaviors may need to target the sexes differently. Teenagers are a compelling group to study. On the cusp of adulthood, teenagers have increasing autonomy and responsibility. The teenage years mark a rite of pas- sage for many Americans—obtaining a driver’s license— that has the potential to alter activity and travel patterns. Although increasingly independent, adolescents are still subject to parental control and consent. In turn, paren- tal consent and permissiveness may be influenced by the adolescent’s gender, having impacts on travel choices, including the decision to drive, walk, or take transit. These travel patterns may be associated with different levels of physical activity in that nonmotorized modes of travel are a direct source of physical activity and travel is often needed to access opportunities to be physically active. This paper investigates gender differences in the fac- tors associated with adolescent mode choices for the trip

204 WOMEN’S ISSuES IN TRANSPORTATION, VOLuME 2 to school and levels of physical activity. Analysis relies on cross-sectional data collected from adolescents in Baltimore City for a school-based recruitment study of physical activity and the built environment. Participants were recruited from two magnet high schools made up of a predominantly African-American sample (67%) with geographically disperse home locations. BAckground This paper brings together the growing literature on ado- lescent travel behavior and the established work on gen- der differences in transportation. There is a long thread of literature on gender differences in travel behavior (5, 6). Most studies focus on adult men and women, and many fewer studies center on the travel of children and teenagers in particular [see Weston for a review (7)]. given that adolescents are nearing adulthood, one might assume that some of the patterns observed among adults would also be present in teens, including the differences between the sexes. As it is for adults, the automobile is the dominant mode for most children’s travel. Weston found that children ages 13 to 15 are driven by their parents for a majority of their trips (7). The increasing number of children who are driven to school and elsewhere illustrates their depen- dence on others for much of their mobility, which many hypothesize contributes to the increasing rates of obesity among children. Thus, many programs are targeted at increasing the number of children who walk to school by making physical improvements to the transportation envi- ronment, organizing walking school buses, and introduc- ing programs that encourage walking to school. Despite the dependence on others for much of their travel, many children also travel independently to activi- ties. Clifton found that nearly 40% of teenagers travel alone on their trip directly after school (8). This inde- pendence in travel is dependent on the age and gender of children. Not surprisingly, the proportion of trips made alone increases with age (3, 8–10) because older children are more mature. But parents often have concerns about allowing female children to travel alone. Parent permission and constraint play an important part in children’s activities. Although teenagers have more independence than younger children, parental control plays a part in their activities and travel choices. Research has documented differences in the way parents treat their teenage children. Adolescent boys tend to be subject to fewer constraints and are often granted more autonomy than are adolescent girls (11–13). Parents tend to monitor their female children more closely than their male children (13, 14). This tighter rein on female children may help to explain differences in observed travel patterns, including travel alone, mode choices, and time of day. Research has shown gender differences in overall activ- ity participation and time use among teens (15). girls are more likely than boys to be involved in a wider array of extracurricular activities. Research has documented differences with respect to activities that affect physi- cal activity. Since 1991, the Youth Risk Behavior Sur- vey found lower levels of physical activity among young people in the united States and a significant decline in reported physical activity during the high school years (16). These marked trends are particularly striking for minority adolescents (17–19). Participation rates and time spent in physical activities are higher for boys than for girls (15, 20–23). This brief review highlights the complexity of adoles- cent travel choices. Travel to school is dependent largely on parent permissions, physical environment, activity participation, and gender. The mode to school among high school students is examined here to try to under- stand some of the factors associated with this decision in males and females and how this might be related to levels of physical activity. dAtA And methodS This section describes the data used in the analysis and the multinomial logit models (MNL) used to examine the factors associated with mode choice to school for males and females. Location This research takes place in the Baltimore, Maryland, metropolitan area and focuses on two schools in Balti- more City, located adjacent to one another and having a student body that resides in neighborhoods through- out Baltimore City and County. These schools can be accessed by light rail and bus transit, and students are required to find their own means of transportation to and from school. Students from both sample schools are predominantly African-American (70%–85%) and white (20%–25%) with total enrollments between 920 and 1,200 for grades 9 to 12. Data Cross-sectional data were collected from adolescents (N = 269) residing in the Baltimore City area for a 2006 study of physical activity and the built environment. Participants (ages 14 to 18) were recruited from two area magnet high schools located adjacent to one another. Magnet schools were chosen as the study sites because students are drawn from across the city and county, thus ensuring variation

205gENDER DIFFERENCES IN ADOLESCENT TRAVEL TO SCHOOL in the student participants’ home environments, which was an important consideration in the study design. Data analyzed here for each individual include (a) a web-based survey that collects background information, attitudes, perceptions and recall activity, and travel behavior, (b) objectively measured physical activity data collected by accelerometers for 1 week, (c) a week-long travel diary, (d) archived spatial data about the built environment around each students’ primary home address, and (e) height and weight measurements, used to calculate body mass index (BMI). The resulting sample was composed predominantly of African-Americans (67%), which reflected the demo- graphics of the school and of Baltimore as a whole. None of the students in the sample live within walking distance of the schools. The descriptive statistics of these data are shown in Table 1 and described in more detail below. The web survey asked students nearly 100 questions in an effort to gather information about the students, their households, attitudes, and behaviors. Relevant information useful to the examination of mode choices includes information about demographics, auto avail- ability, driver’s license status, primary mode of transpor- tation to and from school, and participation in in-school and out-of-school activities, including employment. In addition, a measure of how well students know their way around Baltimore City was constructed from self- reported assessments. A measure of parental permis- siveness to allow their teenager to travel independently by walking, cycling, or transit was constructed from a set of four questions about these modes with respect to parents’ views on teens traveling alone, with adults, and with friends and their knowledge of where and when they are going. Students were asked to wear accelerometers to cap- ture levels of physical activity. Following a standardized protocol, each monitor was initialized before placing it on a belt to be worn around each participant’s waist above the participant’s right hip. Students were asked to wear it all the time, except at night while sleeping and while bathing or swimming during the seven consecu- tive monitoring days. Activity counts were stored in 30-s time intervals. Students who failed to comply with mini- mal wear, had a monitor malfunction, or left fewer than 7 days of data (or nonusable data) were asked to wear the monitor again until usable data were collected. Acclerometry counts were summarized by quantify- ing the time (minutes) spent at different intensity lev- els. The thresholds for the activity intensities were less than 50.99 counts for sedentary activity, 51 to 578.99 counts for light activity, and 579 or more counts for moderate to vigorous physical activity. This threshold of 579 or more counts corresponds approximately to the lower bound for a 2.5-mph walk, representing an activity intensity level of three metabolic equivalents (METs). A 3-MET cut point to define the moderate to vigorous physical activities (MVPA) was used because of its use as the threshold for MVPA in previous stud- ies of youth. Accelerometer data reduction methods incorporated the following data processing issues sug- gested by the literature. Participants were also asked to record all of their trips made during the week in a travel diary. Information about the trips collected include day of week, departure and arrival time, mode, and destination (home, school, neighborhood, other). unlike the survey question, the diary allowed students to record not only their primary mode to school but also the secondary and tertiary modes used for these trips. Several measures of the built environment were com- puted around the students’ home address. Key measures relevant to this paper include pedestrian connectivity, population density, and mixed land use. These measures have been documented in the literature as correlated with transportation choices (24–27). Pedestrian connec- tivity is measured as a ratio of the number of intersec- tions (two- to six-way intersections) to total intersections (two- to six-way intersections plus cul-de-sacs) on the road network. Population density is computed from the u.S. census as the number of people per square mile living in the census block of the student’s home address. The mix of residential and commercial uses is measured by using the Herfindahl–Hirschmann index (HHI), which is computed as the sum of the squares of the percentages of each type of land use within a half-mile buffer of the student’s home. The values range from 0 to 10,000, with the latter value representing a buffer area with only one use (i.e., the higher the value, the lower the level of land use mix). Finally, BMI is a commonly used measure of obesity. Height and weight measures were taken from the stu- dents at the beginning of the study and measures were computed relative to their age and gender. Multinomial Logit Discrete choice models are developed to examine the var- ious factors associated with adolescent’s mode choice to school. Discrete choice models are based on the random utility theory, which assumes that the decision maker’s preference for an alternative can be captured by the value of an index, called utility. It is assumed that the decision maker chooses the alternative that yields the highest util- ity. The probability of any alternative i being selected from a choice set Cn is given by the following: (1) where U is the utility of the given alternative. Because the analyst will have imperfect information about an indi- P i Cn Pr( \ ) (Uin ≥Ujn)= ∀j ∈Cn

206 WOMEN’S ISSUES IN TRANSPORTATION, VOLUME 2 vidual’s utility level, uncertainty is introduced into the utility equation (28, 29). Equation 2 represents the util- ity (Uin) of alternative i in the choice set Cn for decision maker n. Vin is the systematic (or representative) compo- nent; ein is the random component. (2) In this study, the choice alternatives consist of three travel modes: (a) car driver, (b) car passenger, and (c) taking transit. The being a car driver alternative is avail- able only when the respondent’s household owns at least one car and the respondent has a driver’s license. To model this decision, MNL models are specified. The logit model arises from the assumption that the TABLE 1 Descriptive Statistics N Mean Min. Max. SD Personal and household characteristics Age Male 107 15.7 14 18 1.25 14 15.0% 15 41.1% 16 10.3% 17 24.3% 18 9.3% Female 162 15.7 14 18 1.19 14 19.1% 15 26.5% 16 19.8% 17 30.9% 18 3.7% Non-White Male 107 79.4% Female 162 84.0% Number of Cars Male 107 2.3 0 9 1.37 Female 161 2.2 0 10 1.36 Driver’s license Male 107 37.4% Female 162 34.6 Know Baltimore Male 107 3.9 1 5 1.07 Female 162 3.7 1 5 1.16 Parent permissiveness Male 107 14.0 0 16 2.48 Activities Employed Male 107 41.1% Female 162 44.4% Number of outside school activities Male 106 3.2 0 15 Female 162 2.0 0 17 Built environment Connectivity ratio Male 105 0.81 0.52 0.94 Female 162 0.82 0.46 0.95 Mixed use (HHI) Male 105 4497 561 9428 Female 162 4253 150 9370 Population density Male 105 14874 0 48353 Female 162 19491 0 179965 Transportation Mode to school Male 107 Drive 9.3% Get ride 43.0% Transit 43.0% Female 162 Drive 11.7% Get ride 37.7% Transit 50.6% Note: N = number; min. = minimum; max. = maximum; SD = standard deviation.

207GENDER DIFFERENCES IN ADOLESCENT TRAVEL TO SCHOOL difference of the error terms is logistically distributed. Under this assumption the choice probability for alterna- tive i is given by (3) Model Specification The choice set in this model is expressed as Cj = {car driver, car passenger, transit}. The variables included in the utility functions can be summarized as • Individual age; • Number of autos available at the household; • Whether the respondent has a job (dummy; 1, has a job; 0, otherwise); • Race (dummy; 1, nonwhite; 0, white); • Whether the person knows her or his way around Baltimore City (dummy; 1, yes; 0, otherwise); • Summary measure of parent permissiveness concern- ing walking, bicycling, and taking transit (variable range 0–16, with higher scores indicating more permissive); • Number of out-of-school activities respondent par- ticipates in; • Population density; • Connectivity ratio; and • HHI. On the basis of this specification, two models are esti- mated for males and females separately to better under- stand the factors associated with mode choice and the gender differences. RESULTS AND DISCUSSION This section provides the results of the analysis and dis- cusses the implications of these results for the male and female students. Mode Choice to School Results of the MNL models for primary mode to school for females and males are shown in Table 2. The primary modes modeled include auto drive, transit, and auto pas- senger (base case). Both models are statistically signifi- cant at the 99% confidence level, r < 0.001 for the c2 test. The log-likelihood ratio test clearly rejects the null TABLE 2 Primary Mode to High School Female Male Mode: Drive, Transit, Auto Passenger (base case) Coefficient t-Ratio Coefficient t-Ratio Drive Alternative specific constant 11.174 0.80 –77.083 –1.64 Age –0.544 –0.71 4.714 1.67 Race (1 = nonwhite) –0.719 –0.59 6.861 1.63 Number of cars in household 0.809 1.77 3.012 1.78 Knowledge of city 2.321 2.09 –5.311 –0.69 Have a job 0.340 0.34 11.074 1.24 Out of school activities 0.382 1.70 –0.532 –1.36 Parent permission –0.264 –1.25 –0.140 –0.48 Herfindahl–Hirshmann index 0.001 1.71 0.000 –0.82 Population density 0.000 1.48 0.000 0.31 Connectivity ratio –6.870 –1.46 –23.152 –1.52 Transit Alternative specific constant –2.722 –0.79 0.157 0.03 Age –0.337 –1.74 –0.357 –1.48 Race (1 = nonwhite) 2.241 3.39 2.372 2.87 Number of cars in household –0.366 –2.41 –0.262 –1.46 Knowledge of city –0.396 –0.87 –0.457 –0.72 Have a job 0.628 1.40 1.203 2.03 Out of school activities –0.139 –1.76 –0.056 –0.60 Parent permission 0.312 3.09 0.256 2.00 Herfindahl–Hirshmann index 0.000 –2.57 0.000 –0.86 Population density 0.000 –0.40 0.000 0.26 Connectivity ratio 5.248 2.07 1.071 0.34 Summary statistics: Female Male Number of observations 160 97 Log likelihood at optimate –98.50 –54.55 Log likelihood at no coefficient –175.77 –106.56 Pseudo R2 0.44 0.47 Adjusted R2 0.38 0.36

208 WOMEN’S ISSuES IN TRANSPORTATION, VOLuME 2 hypothesis that all independent variable coefficients are zero for both models. The segmented models show some key differences between the travel choices of boys and girls in their trip to school. Age is significant and positively correlated with driv- ing in the male model but insignificant and negatively correlated in the female model. The ability to drive is correlated with age (and why possession of a driver’s license was not included in the models), and thus the results for males make sense. Why the results are not similar for female students is less clear. Age is also significant and negatively correlated with taking transit for girls. As girls become older they are less likely to take transit relative to being a passenger in an automobile. Being nonwhite is significant in both models. The coefficient is positively associated with taking transit to school and has a similar magnitude for both males and females. This is consistent with other studies that demon- strate that African Americans and other ethnic minori- ties are more likely to use public transportation. In this case, explanations for this choice could be associated with lower household incomes relative to that of whites, leading to less access to automobiles. However, race was not significantly correlated with the number of autos in the household. In addition, the higher propensity for nonwhites to take transit to school may also reflect the spatial distributions of home locations that have better access to transit than do white students. However, the variables related to transit access (distance to rail and bus stops) were not significant in the models and were not included in the final versions. Access to automobiles was reflected in the models by the number of automobiles in the home. This variable was significant in both the male and the female mod- els, albeit with different effects. The more automobiles in the home, the more likely students are to drive to school. However the effect appears to be greater for male students than for females, as the coefficient for males is more than three times that for females. Like- wise, the number of vehicles is negatively associated with taking transit, but the variable is significant only for female students. An interesting result is the significance of the variable reflecting students’ knowledge of the city of Baltimore. This variable is significant, and the coefficient is positive in the female model for the choice of driving to school. Although this indicates that female students that drive to school feel that they have a better knowledge of the city, the direction of that association is unclear. Students who drive may develop a better, or at least different, knowl- edge of the city than those who take transit. The freedom to drive may allow for more exploration of the city than is possible for those limited to transit routes. However, it may be that students who know more about Baltimore City are given parental consent to drive or may feel more comfortable driving than those with less knowledge of the urban area. The fact that the variable is significant only for females raises questions about spatial knowl- edge and comfort level in traveling in places that may be unfamiliar. Having a job might influence the mode to school as a result of temporal or spatial constraints placed on a trip before or after school. Students who have jobs may also rely more on automobiles because of these constraints or because the job affords the resources to purchase a vehi- cle. Thus, one would expect this to be significant in these models. However, the variable was significant in the male model and only for taking transit to school. Surprisingly, it is positively correlated with the choice of transit. Participation in organized activities out of school may place scheduling constraints similar to holding a job. In addition, playing sports, taking music lessons, or other activity may necessitate carrying additional equipment, clothing, and other items. Therefore, travel to school by automobile may be more attractive, in that it provides higher levels of mobility and facilitates hauling gear. The results show a negative association between the number of out-of-school activities and taking transit to school among females. The coefficient was not significant for driving among females or among males. Although teens can have a great deal of autonomy in their activities and travel relative to younger children, parent permission remains an important determinant of children’s travel. Here the variable measures the level of parent permission for independent travel by walking, bicycling, or transit. As expected, the variable was sig- nificant for taking transit to school in both the male and the female models. Several built environment variables were included in the model because the literature has shown correlations between land use and urban form characteristics and mode choice, as discussed above. The mix of residential and commercial uses near the student’s home as measured by the HHI was significant in the female model only for driving and taking transit, although the magnitude of the coefficient is small. Population density was not sig- nificant in either model. Connectivity ratio, a measure of pedestrian networks, was positive and significant in the female model for taking transit to school. This suggests that the pedestrian network is an important facilitator of transit mode share for girls. Neighborhoods with higher connectivity may have better pedestrian access to transit or may be correlated with better transit service as a result of their location in the city. Why these built environment variables are significant only for the female models is puzzling and points to areas for more research. Some research shows differences in how men and women respond to, perceive, or hold atti- tudes about environmental characteristics in ways that

209gENDER DIFFERENCES IN ADOLESCENT TRAVEL TO SCHOOL may affect behavior. Clifton and Dill (30) show that dif- ferences in the built environment correlate with the walk- ing behavior of women and men. Studies show women appear to be more sensitive to safety and aesthetics than men (31, 32) Transit Use, Physical Activity, and Obesity There have been numerous calls for increasing children’s physical activity by facilitating walking to school. How- ever, the relationship between walking to school and measures of physical activity remains somewhat unclear. In this study, none of the students live within walking distance to the school; however, many walk as part of their transit commute trip. Here the relationship between measures of mode to school, physical activity, obesity, and gender is examined. The differences between male and female students in the factors associated with mode choice would suggest that there may be differences in their levels of physical activity. Figure 1 reveals these differences, significant to the 99% confidence level, between those who take transit and those who take automobiles (passenger and driver), in aggregate measures of physical activity dur- ing the weekday. The average weekday total (across 5 weekdays) min- utes spent in MVPA for males was 210 for those trav- eling by auto and 275 for those traveling by transit. For females, the comparable statistics are 168 min and 211 min, respectively. A significantly greater amount of time is spent in engaging in physical activity among those who take transit, for both genders, although males show greater amounts of physical activity overall than females. This trend holds for the average minutes spent per 1 weekday engaged in moderate physical activity. Students taking transit have higher levels of engage- ment in physical activity (54 min for males; 47 min for females) than those who take automobiles to school (43 min and 37 min, respectively). The measures of sedentary behavior are consistent with these findings, although the magnitude of the differences is smaller. Students who take transit to school spend fewer minutes in sedentary behaviors during weekdays than those who take auto- mobiles. Females spend more time in sedentary activity than males, however, across all categories of mode. Some have speculated that transit users will exhibit higher levels of physical activity because of walking to access and egress transit (33). To examine the walking trips associated with transit users (although not neces- sarily transit trips), the total number of walking trips made during weekdays (Monday through Friday) for all purposes, not just school travel, was calculated from the travel diary data. In Figure 1 the differences exhibited in mean total walking trips are significant to the 99% con- fidence level. Students who take transit to school under- take significantly more walking trips per week than those who travel by car. Although the total number of walk- ing trips for males and females is nearly the same (3.5 and 3.4, respectively) (Figure 2), there are clear gender differences in the number of trips by mode of transport to school. Females, who travel to school by automobile make only 1.4 trips on average during school days, com- pared with 2.7 by their male counterparts. Conversely, the most walking trips during the school week are made by females who take transit to school—amounting to 5.1 walking trips on average compared with 4.6 for males. The differences in walking rates between automo- bile and transit commuters are more straightforward to explain, albeit just speculation, than the gender differ- ences in walking. In addition to the hypothesis that transit users walk more as part of their transit trips, neighbor- hoods with good transit access may also be communi- ties with characteristics that support walking. Finally, students who use alternative modes may lack access to automobiles and thus use these modes more readily. 210 275 168 211 486 472 493 478 43 54 37 47 0 100 200 300 400 500 600 Auto Transit Auto Transit Male Female Primary Mode to School M in ut es MVPA total weekdays Sedentary weekday Moderate weekday FIGURE 1 Measures of weekday physical activity, by gender and mode to school.

210 WOMEN’S ISSuES IN TRANSPORTATION, VOLuME 2 One must be careful not to attribute causality to these correlations and must interpret them with caution. It is unlikely that the physical activity associated with taking transit—walking to and from stops—is responsible for all of these observed differences. What it does suggest is that transit may provide important mobility for youth in accessing opportunities to be physically active. Transit may allow independent travel that may facilitate partici- pation in activities, such as organized sports, informal recreation, or employment, that increase physical activity levels. Although travel by automobile is normally associ- ated with greater mobility, transit may provide greater levels of independent mobility for adolescents without access to a car, who cannot drive, or who cannot arrange for a ride because of adult scheduling constraints. The number of outside activities that female students partici- pate in, however, is negatively correlated with choosing transit, which confounds this logic. Comparing BMI for transit and automobile commut- ers does not yield significant results. BMI for males is 24.5 for those that commute to school by automobile and 24.7 for transit users; these results are not statistically signifi- cant. For females, BMI is 24.0 for automobile users and 25.0 for transit users. These differences are also statisti- cally insignificant. Obviously, body weight and obesity are more complex than just the level of physical activity. Diet, genetics, and physical activity all play a role. Similar to levels of physical activity, it is unlikely that differences in BMI could be attributed solely to transit use. concluSionS The analysis presented here shows some important gen- der differences in the variables associated with mode choice in travel to school. These range from personal and household characteristics, activity participation, parent permission, and the built environment. Race, age, the number of vehicles available, and parent permission affected both males and females, although at different magnitudes and often for different modes. Female stu- dents’ choices were associated with their knowledge of the city, number of out-of-school activities, and the built environment. Having a job was an important determi- nant of males’ transit use. The primary mode to school was also correlated with differences in physical activity levels and number of walking trips made during the week but not with measures of obesity. These findings suggest that students who take transit to school exhibit greater levels of overall physical activity although why this is the case cannot be discerned from these data. Consistent with other research, girls show lower levels of physical activity than boys, across all mode categories. Although males and females have simi- lar rates of walking, on average, females who take transit to school walk nearly four times as often as females who take an automobile to school. Although females who take transit tend to be making more walking trips than males, they are engaging in fewer minutes of physical activity, suggesting that the walking trips may be short or at a leisurely pace. Although the actual relationship to transit is unclear, these differences by mode are notable and should be examined more closely. Adolescents remain an interesting group for study because they have formative experiences and form pref- erences and individual and group identities that may hold as they enter into adulthood. The results of this paper raise more questions than they answer, suggesting that gender differences in adolescent travel can be a fruitful area for future investigation. referenceS 1. Centers for Disease Control and Prevention (CDC). Youth Risk Behavior Surveillance—united States, 2005. Mor- Av e ra ge W e e kl y W a lk T rip s 6 5 4 3 2 1 0 Male Auto Transit Total Female Male Female Male Female 2.7 1.4 4.6 5.4 3.5 3.4 FIGURE 2 Average total weekday walk trips by commute mode to school and gender.

211gENDER DIFFERENCES IN ADOLESCENT TRAVEL TO SCHOOL bidity and Mortality Weekly Report, Vol. 55, No. SS-5, 2006. 2. Special Report 282: Does the Built Environment Influence Physical Activity? Examining the Evidence. Transporta- tion Research Board of the National Academies, Washing- ton, D.C., and Institute of Medicine, 2005. 3. McDonald, N. 2005. Children’s Travel: Patterns and Influences. PhD dissertation. university of California, Berkeley, 2005. 4. McMillan, T. E. The Relative Influence of urban Form on a Child’s Travel Mode to School. Transportation Research Part A, Vol. 41, 2007, pp. 69–79. 5. Rosenbloom, S. Travel by Women. In Demographic Spe- cial Reports (1990 NPTS Publications Series). FHWA, Washington, D.C., 1995, pp. 2-1 to 2-57. 6. Rosenbloom, S. Trends in Women’s Travel Patterns. In Wom- en’s Travel Issues: Proceedings from the Second National Conference, October 1996, FHWA, u.S. Department of Transportation, Washington, D.C., 2000, pp. 16–34. http:// www.fhwa.dot.gov/ohim/womens/chap2.pdf. 7. Weston, L. M. What Helps and What Hinders the Inde- pendent Travel of Non-Driving Teens. PhD dissertation. university of Texas at Austin, 2005. 8. Clifton, K. J. Independent Mobility Among Teenagers: An Exploration of Travel to After-School Activities. In Trans- portation Research Record: Journal of the Transporta- tion Research Board, No. 1854, Transportation Research Board of the National Academies, Washington, D.C., 2003, pp. 74–80. 9. Stefan, K. J., and J. D. Hunt. Age-Based Analysis of Chil- dren in Calgary, Canada. Presented at 85th Annual Meet- ing of the Transportation Research Board, Washington, D.C., 2006. 10. Mackett, R. L., L. Lucas, J. Paskins, and J. Turbin. Health Benefits of Non-Car Travel by Children. Presented at Hertfordshire County Council Centre of Excellence Con- ference on School and Business Travel Plans, Hatfield, united Kingdom, Nov. 25, 2002. 11. Bumpus, M. F., A. C. Crouter, and S. M. McHale. Paren- tal Autonomy granting During Adolescence: Exploring gender Differences in Context. Developmental Psychol- ogy, Vol. 37, No. 2, 2001, pp. 163–173. 12. Hagan, J., J. H. Simpson, and A. R. gillis. Class in the Household: A Power-Control Theory of gender and Delinquency. American Journal of Sociology, Vol. 92, 1987, pp. 788–816. 13. Smetana, J. g., and C. Daddis. Domain-Specific Anteced- ents of Parental Psychological Control and Monitoring: The Role of Parenting Practices and Beliefs. Child Devel- opment, Vol. 73, No. 2, 2002, pp. 563–580. 14. Dishion, T. J., and R. J. McMahon. Parental Monitor- ing and the Prevention of Child and Adolescent Problem Behavior: A Conceptual and Empirical Formulation. Clin- ical Child and Family Psychology Review, Vol. 1, No. 1, 1998, pp, 61–75. 15. gibbons, J. L., M. Lynn, and D. A. Stiles. Cross-National gender Differences in Adolescents’ Preferences for Free- Time Activities. Cross-Cultural Research, Vol. 31, No. 1, 1997, pp. 55–69. 16. Troiano, R. P. Perspective: Physical Inactivity Among Young People. New England Journal of Medicine, Vol. 347, No. 10, 2002, pp. 706–707. 17. Strauss, R. S., and H. A. Pollack. Epidemic Increase in Childhood Overweight, 1986–1998. Journal of the Amer- ican Medical Association, Vol. 286, No. 22, 2001, pp. 2845–2848. 18. Ogden, C. L., K. M. Flegal, M. D. Carroll, and C. L. John- son. Prevalence and Trends in Overweight Among u.S. Chil- dren and Adolescents, 1999–2000. Journal of the American Medical Association, Vol. 288, 2002, pp. 1728–1732. 19. Centers for Disease Control and Prevention. Youth Risk Behavior Surveillance—united States 2003. Mortality and Morbidity Weekly Report, Vol. 53, No. 2, May 21, 2004, pp. 1–96. 20. Kohl, H. W. III, and K. E. Hobbs. Development of Physi- cal Activity Behavior Among Children and Adolescents. Pediatrics, Vol. 101, No. 3, 1998, pp. 549–554. 21. Larson R. W., and S. Verma. How Children and Ado- lescents Spend Time Across the World: Work, Play, and Developmental Opportunities. Psychological Bulletin, Vol. 125, No. 6, 1999, pp. 701–736. 22. Sallis, J. F., J. J. Prochaska, and W. C. Taylor. A Review of Correlates of Physical Activity of Children and Adoles- cents. Medicine and Science in Sports and Exercise, Vol. 32, No. 5, 2000, pp. 963–975. 23. Shann, M. H. Students’ use of Time Outside of School: A Case for After-School Programs for urban Middle School Youth. The Urban Review, Vol. 33, No. 4, 2001, pp. 339–355. 24. Cervero, R., and K. Kockelman. Travel Demand and the 3Ds: Density, Diversity, and Design. Transportation Research D, Vol. 2, No. 2, 1997, pp. 199–219. 25. Crane, R. The Influence of urban Form on Travel: An Interpretative Review. Journal of Planning Education and Research, Vol. 15, 2000, pp. 3–23. 26. Ewing, R., and R. Cervero. Travel and the Built Environ- ment. In Transportation Research Record: Journal of the Transportation Research Board, No. 1780, TRB, National Research Council, Washington, D.C., 2001, pp. 87–114. 27. Saelens, B. E., J. F. Sallis, and L. D. Frank. Environmental Correlates of Walking and Cycling: Findings from the Trans- portation, urban Design, and Planning Literatures. Annals of Behavioral Medicine, Vol. 25, No. 2, 2003, pp. 80–91. 28. Ben Akiva, M., and S. T. Lerman. Discrete Choice Analy- sis: Theory and Applications to Travel Demand. MIT Press, Cambridge, Mass., 1985. 29. Train, K. Discrete Choice Methods with Simulation. Cam- bridge university Press, 2002. 30. Clifton, K. J., and J. Dill. Women’s Travel Behavior and Land use: Will New Styles of Neighborhoods Lead to More Women Walking? In Conference Proceedings 35:

212 WOMEN’S ISSuES IN TRANSPORTATION, VOLuME 2 Research on Women’s Issues in Transportation: Report of a Conference; Volume 2: Technical Papers, Transporta- tion Research Board of the National Academies, Washing- ton, D.C., 2005, pp. 89–99. 31. Bianco, M., and C. Lawson. Trip Chaining, Childcare, and Personal Safety. In Women’s Travel Issues: Proceed- ings from the Second National Conference, October 1996, FHWA, u.S. Department of Transportation, Washington, D.C., pp. 121–143. http://www.fhwa.dot.gov/ohim/wom ens/chap8.pdf. 32. Livi Smith, A. D, C. V. Burnier, and K. J. Clifton. Wom- en’s Walking Behavior and Health: Exploring the Links. Presented at 86th Annual Meeting of the Transportation Research Board, Washington, D.C., 2007. 33. Wener, R. A Morning Stroll. Environment and Behavior, Vol. 39, No. 1, 2007, pp. 62–74.