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.

154 A Commitment to Continue? Comparing Women and Men Commuters Who Choose Transit over Driving Alone Jane Gould and Jiangping Zhou, University of California, Los Angeles This study tracks results from an employer-sponsored travel reduction program to explore whether there are gender-related differences in how men and women select and use public transit. Although women ride public tran- sit more than men, it is not clear that this effect would be found in a study of middle-income women with full-time jobs. Their responsibilities and roles might lead these women to favor the flexibility and convenience of an automobile. The study follows 381 commuters, 144 men and 237 women, who chose to give up their drive-alone commuting for a 3-month period in Southern California. More than two-thirds of both the men and the women remained as transit riders when the time came to make a commitment to continue with transit. The analysis found few differences between men and women. However, the women who chose to participate in the transit experi- ment in the first place had distinct characteristics: they had fewer children at home, were from select age groups, and had smaller households. The study provides insight for future social marketing experiments in transporta- tion and provides results for public transit providers who wish to attract busy commuters who have family and household responsibilities. As more women have entered the workforce full time, the mode share for public transportation has declined. There is interest in reversing this trend, but also widespread recognition that public transit may not be conducive for workers who have multiple family and household responsibilities. This study reports on an employer-sponsored travel reduction program. Men and women who voluntarily chose to switch from drive-alone commuting to public transit were longitudinally tracked during a 3-month period. Their willingness and ability to ride public transit are examined at two separate points in time. There are a number of reasons that the decision to switch from a drive-alone mode to public transit might be associated with gender. The paper begins with speculations on some of these factors (see Table 1) and then returns to them in reviewing the literature. overvieW of The TransiT experimenT A social experiment was conducted at the University of California, Los Angeles (UCLA), to recruit full-time single-occupancy drivers and provide them with free commuter transit passes. Although the university had an employee transit mode share of nearly 14% (1), it was felt that more aggressive transportation promotions were needed during a time of rising gasoline prices. A 3-month experiment was conducted between June and September 2008. Participants received a commuter transit pass at no cost if they turned in their parking permit. At the end of the 3-month trial, employees could regain their park- ing allocation or elect alternative transportation. Regu- lar bus riders then received a transit fare subsidy of 50%. The experiment was found to be an unusually successful demonstration of transportation demand management (TDM). At the end of the 3-month period, only 30% of the men and women who enrolled decided to return to drive-alone commuting.

155COMPARiNG WOMEN AND MEN COMMUTERS WHO CHOOSE TRANSiT in July 2008 gasoline prices reached nearly $4.60 a gallon in California, and most employees were attracted to the program so that they could mitigate the fuel costs, as well as the cost of monthly parking, which was about $70. However, many participants cited a secondary reason for enrolling: they wanted to explore new travel modes without risking their parking allocation. previous research on Women’s choice of TransiT for commuTinG Studies of “single-occupancy vehicle (SOv) conversions” and voluntary travel behavior programs have not focused on gender issues, per se (2–4). The outcomes are generally favorable and show that the perception of public transpor- tation is improved. However, the literature incorporating self-selection in travel choices underscores that decisions depend on abilities, needs, and preferences. There is a dif- ferent stream of literature suggesting that participation in TDM and car-reduction schemes will generally be biased against participation by working women (5). Historically, women took more transit trips than men (6), and employed women were more likely to ride tran- sit, although even by the late 1970s it was observed that gender differences in the use of public transportation were not large (7). There are underlying and interrelated labor factors related to transit use: women workers live closer to their jobs, have lower salaries, and are more likely to work part-time. However, during the past 30 years there have been “complex social changes” in women’s work roles. During this period, the proportion of women using public transit and carpooling has declined, while the proportion of women depending on private cars has increased (8). There is recent evidence from urban met- ropolitan areas of a convergence of men’s and women’s modal share for transit (9, 10). But, overall, there are still important gender differences. Specifically, analysis of data for the period between 1985 and 2005 indicates that across all travel modes, aver- age commute lengths and travel time have increased for both men and women (8). There continues to be a gender gap in both distance and travel time, and commutes for women are shorter. The patterns of transit ridership vary by race: there has been a large decline in ridership among black female commuters, but their commute distance and commute times have increased, signaling a spatial mis- match between jobs and housing. For other women riders, the commute length on transit has increased slightly and the duration (time) has declined. Their transit mode share is less than 10%. Crane observes that a shift in commute trips from transit to car has occurred because women pre- fer cars, or need them for work, or live in areas with a poor transit level of service. Some factors that discourage women from using tran- sit more are that they cannot trip chain or engage in “serve-other” trips. Trip chaining allows busy working women to balance their domestic and child care obliga- tions and makes personal vehicles a more convenient and efficient choice. The need to make transportation choices congruent with the fulfillment of household duties is evi- denced by multilinked travel trips (11–13). in principle, the spatial distribution of schools, shopping, leisure, and medical facilities should be factors that influence the fre- quency and need for trip chaining. However, McDonald found that higher-density neighborhoods are not neces- sarily associated with a reduction in vehicle miles trav- eled among households with children (14); one reason is that many urban areas are not necessarily safer for children to traverse alone. Safety has an important link to women’s travel choices for a separate and distinct reason. The characteristics of neighborhoods where transit boarding takes place are closely related to the issue of security and crime (15). Concerns about safety are associated with both waiting for transit and actual time in transit. There are complex social reasons that give rise to these perceptions, delin- eated in a recent study by Loukaitou-Sideris et al. (16). For female commuters who take transit, one might pre- dict that security concerns would be more acute among those who work at night, among those who live in poorer neighborhoods, and those in neighborhoods with less frequent service. Rosenbloom and Burns have conducted one of the most extensive studies of gendered issues among employee commuters (17). They found that women commuters were more favorable to TDM strategies than men, but were also less likely to be able to switch to TABLE 1 Comparison of Transit Versus Car Travel Characteristic Transit vs. Car Potential impact on Household Time to travel Usually longer and often less frequent Transit rider may spend more time in travel and less time with transit service family. Also reduces family time together in car. Flexibility of travel Reduced flexibility to pick departure Transit rider may be less able to get home or to school “on demand.” Trip chaining Particularly hard for groceries, packages Need to take additional trips for certain errands, shopping, and leisure. Little ability to engage in “serve-other” trips. Exercise Transit provides more opportunity to walk Transit riders may spend less time at gym and reduce other exercise regimes, leaving more time for other activities.

156 WOMEN’S ISSUES IN TRANSPORTATION, VOLUME 2 alternative modes because driving helped them serve family needs and responsibilities. Their data show that having an automobile is important to employed women, particularly those with lower incomes. One conclusion is that “having a car is a necessity and not a luxury for most working women and their families, given current land use, housing, employment, and service patterns.” Specific to TDM programs and transit passes, they note that transit passes cannot compensate for time lost to travel on longer commutes, added hours or child care or elder care expenses, the lack of current transit service, and associated security issues. It has been suggested that many decisions about using TDM are local because of wide variations in service levels and availability. In the Los Angeles area, 60% of Metro riders are currently women (18). This percentage is for all travel, not just commute trips. Among studies that have focused on Los Angeles commuters, Sarmiento examined panel data for Southern California commuters (19). She found that gender had a role in predicting the frequency and types of car trips and carpool use in the region. She did not ask specifically about transit versus carpool use. In another Southern California study, Novaco and Col- lier found in a telephone survey, that women with lon- ger commutes (>20 mi) perceived more stress, and those who said they found commuting stressful were also more inclined to try rail or carpooling (20). RESEARCH METHODOLOGIES AND PROCEDURES The social experiment took place during a 3-month period, from June 2008 through August 2008. A longi- tudinal data set was created based on recordings at three different times. At the initial data capture in June an online survey was used to gather the commuting address for each respondent and some basic demographic infor- mation. In August 2008 a follow-up was done with an online survey that asked respondents about their experi- ence and attitudes while riding transit. In October 2008 the database for the study was updated and finalized by using parking records to classify each participant’s sub- sequent parking decision. This study is organized around reporting the results from these three periods. First, the characteristics of those who signed up for the trial are described, and with the use of geographic information system (GIS) tools, their travel distances are reported. Also reported are the demographic and household characteristics of those who self-selected for the transit experiment. In the second part, the opinions and attitudes these riders expressed toward transit are presented. When this survey was administered, riders were still engaged in the social experiment and had not made a final mode deci- sion. The online survey had a response rate of 48%. In the third and final part, the outcome is reported: whether or not respondents returned to drive-alone com- muting. The dependent variable was measured as the decision (yes–no) to regain employee parking. Letting p denote the probability of an employee regaining parking, the binary logit model investigates the sensitivity of this response to a posited vector of explanatory variables X by the following specification: where a = intercept, b = vector of parameters, and e = error term. Two different logit regression models were run because the sample sizes were different and doing so avoided hav- ing too many missing values. The first logit used geo- graphic variables, keyed from the respondents’ home address. Density was classified by using a GIS overlay of 2006 census tract information for Los Angeles County, and the coding for distance was accomplished by using a shortest path function in TransCAD. There are also data on the number of direct bus routes serving the respon- dents’ home address. The second logit model, which relies on self-report and attitudinal preferences, is based on the survey data only. There were data for approximately 185 respon- dents, one-half of the study group. Their demographic characteristics and personal ratings of transit are used as predictors. These ratings of transit are essentially quali- tative but are measured on conventional scales. In the final section of the data analysis, an alternative method was explored. A brief textual analysis of open-ended statements was performed to explore what riders “said” when they completed the open-ended portions of their questionnaires. MEASUREMENT AT TIME 1: WHO SIGNED UP AND WHAT ARE THEIR GEOGRAPHIC CHARACTERISTICS? During a 3-week period, 381 commuters (n = 381), 144 men and 237 women, signed up for the social experi- ment. Sixty-two percent (62%) were female. This rate corresponds with the underlying population. Accord- ing to a 2006 workforce study, 64.4% of the university workforce was female (21). The same report notes that only 33% of women were in senior management. In Table 2 the demographic characteristics of partici- pants are presented. Although none of the overall rela- tionships are statistically significant at the p £ .05 level, log logit π π π α β ε( )= − ⎛ ⎝ ⎜⎜⎜ ⎞ ⎠ ⎟⎟⎟⎟= + ′ + 1 X

157COMPARING WOMEN AND MEN COMMUTERS WHO CHOOSE TRANSIT there are suggestions that gender differences did influ- ence who volunteered to give up their SOV commute. First, there was no statistical difference in income levels, and on the basis of the university workforce report, it appears that the majority of participants, both men and women, came from administrative jobs and hospital sup- port functions (21). However, women at both ends of the age spectrum, those between 26 and 35 years of age and those over age 55, were more likely than men to elect the mode change. Likewise three-fourths of the women who chose to participate did not have any children at home, com- pared with just two-thirds of the males. Nearly two- thirds of the transit riders (both genders) did not have children under 16 at home and their households were small. Women who had fewer child-raising responsibili- ties were more likely to preselect as transit riders. When respondents signed up, they self-reported their level of transit experience, so a travel packet could be customized. The men who enrolled indicated somewhat more transit experience than the women. Sixteen percent of the men and 31% of the women said that they had no experience riding transit, and 19% of the men and only 10% of the women said that they had a great deal of previous experience (c2 = 6.696, degrees of freedom = 3, p = .082). About one-third of both the men and the women asked for assistance with transit routing, sched- ules, and maps. During the initial sign-up, each person was individu- ally matched with the transit provider that had the clos- est routes and best served the person’s home address. Approximately 50% rode the Big Blue Bus or Culver City Bus; 37% used Los Angeles Metro buses; 8% rode on Los Angeles Department of Transportation vehicles an average distance of 15 to 20 mi; and 5% came on City of Santa Clarita Transit, a coach-style bus with a 35-mi one-way commute. Figure 1 maps the addresses of participants and the residential “hot spots.” TABLE 2 Comparison of Male and Female Participants in Social Experiment (Survey Data) Variable Male (%) Female (%) c2 Degrees of Freedom Significance Household Size (n = 62) (n = 122) 1.72 3 .632 1 34 27 2 26 34 3 21 18 ≥4 19 21 Children age £ 16 yrs (n = 62) (n = 121) 4.546 3 .208 0 66 74 1 15 16 2 18 7 3 1 3 Age (n = 62) (n =122) 8.99 4 .061 20–25 7 3 26–35 26 35 36–45 29 20 46–55 29 20 55+ 10 23 Number of cars in household (n = 61) (n = 122) 4.157 3 .245 1 44 33 2 43 44 3 7 16 ≥4 6 7 Salary band (thousands) (n = 114) (n = 190) 2.760 3 .430 0–30 4 3 30–60 49 58 60–100 40 33 >100 7 6 FIGURE 1 Direct bus routes to UCLA and hot- spot locations for take-a-vacation participants.

158 WOMEN’S ISSUES IN TRANSPORTATION, VOLUME 2 The median length from home to work was found to be 10 mi, by applying the shortest path function in TransCAD. Additional travel information is reported in Table 3. Women did not live closer to work than their male counterparts, and men and women had equal length trips. By using self-reports from survey data, it was found that the average time to get to the bus stop from home was only 8 min, but the range was large. Some respondents drove to a park-and-ride, but the majority walked from their residences. The self- reported time to get from the bus stop to the office was 6.2 min. The average time spent riding the bus was 42 min, but with large deviations in travel time. Respon- dents were not asked about the amount of time spent waiting for the bus, because it was assumed that regu- lar commuters learned to time their departure times to the transit schedule. MEASUREMENT AT TIME 2: ATTITUDES AND OPINIONS WHILE PARTICIPATING IN TRANSIT This section reports on participants’ attitudes and opin- ions about 8 to 10 weeks after they began the transit experiment. The riders had not decided yet whether they would continue with transit, return to SOV travel, or find another alternative. Although the increase in gas prices was the primary reason people said they enrolled, participants cited other rationales: they wanted to try transit, it was more sus- tainable, and driving was stressful (Table 4). There were no gender-related differences, and both men and women strongly agreed that transit was more relaxing than driv- ing. Furthermore, neither group wished to vanpool or carpool instead. There was also some agreement that the buses are too crowded and that transit provides a good way to get more exercise. Even on a critical variable about travel flexibility, “did respondents mind the fixed sched- ule of transit,” the mean score was 3.2 on the 5-point scale. Likewise, with regard to the statement “Transit in Los Angeles has come a long way,” the response tipped toward favorable, and the mean score was 2.8. A separate set of questions asked respondents to rate the place where they waited for transit (the bus stop), in regard to its safety, cleanliness, and comfort. There were no significant differences by gender. Only 17 survey respondents, in total, said their bus stop was unsafe. MEASUREMENT AT TIME 3: THE DECISION TO RETURN TO DRIVING ALONE (QUANTITATIVE AND QUALITATIVE ANALYSIS) The substantive interest was whether participants returned to drive-alone commuting after their “free” transit pass expired. This outcome was measured a full 6 weeks after the experiment ended, in case a vacation or payroll-timing delayed the sign-up. Overall, only 114 of the social experiment participants requested parking again (114/381 = 30%). In the experiment, there were 144 men, and the number that requested parking and returned as drivers was 41 (41/144 = 29%). There were 237 women, and 73 regained parking (n = 73/237 = 31%). A chi squared test indicated no significant gender differences (c2 = .232, not significant at the .05 level) . There were also no significant differences among transit carriers (e.g., Metro versus Los Angeles Department of Transportation). TABLE 3 Actual Travel Distance and Self-Reported Travel Times, by Gender Variable Men Women t-statistic Significance Travel distance in miles (mean) 10.2 (8.58) 10.0 (7.57) .308 .758 Self-report travel time to bus stop 7.8 (6.6) 9.32 (8.3) –1.238 .217 Self-report travel time on bus(s) 42.02 (23.2) 42.03 (20.4) .005 .996 Self-report travel time from bus to work or office 6.0 (3.9) 6.4 (4.6) –.643 .547 Note: Standard deviation is given in parentheses. TABLE 4 Attitudes About Transit and Driving Attitude Male (n = 62) Female (n = 122) t-statistic Significance Transit is more relaxing than driving 1.1 1.1 1.24 .214 I get more physical exercise when I take transit 2.2 2.1 .632 .181 I might consider carpooling as a commuter option 3.9 3.9 .001 .999 I hate to be tied to the fixed schedule of transit 3.2 3.2 .041 .886 I might consider vanpooling a commuter option 4.3 4.2 .015 .901 The buses are too crowded 2.7 2.6 .287 .787 Transit in Los Angeles has come a long way 2.8 2.8 .194 .845 I still need to drive 4.1 4.1 .018 .894 Note: Five-point scale, where 1 = strongly agree, and 5 = strongly disagree.

159COMPARiNG WOMEN AND MEN COMMUTERS WHO CHOOSE TRANSiT predicTinG The TransiT–sov choice Many factors were reviewed, and a smaller set was then identified, which it was expected might predict whether users continued to commute via transit or switched back to driving, particularly geographic influences (22). if women were more constrained by their household duties, then they would be less likely to keep the transit pass, particularly if it involved longer travel times or if they lived further from work. it was also posited that neighborhood density might be a factor because women who lived in denser urban areas might have less need to trip chain for shopping and errands. There were also data on the number of direct bus routes that served the home-to-work commute (for most riders, it was only one). A binary logit regression with this set of geographic variables is presented in Table 5. initially 73% of the cases for men were correctly classified. in the final model, the geographic variables, such as distance, number of bus routes, and residential density, did not show any impact. No additional cases were classified. Among women, 67% of the cases were correctly predicted both before and with the model. The geographic variables were not shown to be additional predictors of mode choice for either men or women. By using the online survey data from the Time 2 mea- surement (Aug. 2008), a separate logit analysis based on behavioral and sociodemographic inputs was undertaken. The indicators included the respondent’s age (chrono- logical), children under age 16 (0/1), household size (one member to four members), and salary (lowest = 1). Also tested in this model were ratings about the bus stop safety and comfort, the attitudinal variables (1 = strongly agree, 5 = strongly disagree), and whether respondents said they continued to commute by car during the trial. Table 6 reports the outcome. For this analysis there were not many men (n = 57) after cases with missing variables were excluded. The model for men classified 63% of the cases with no information and improved to a classification of 83% on the basis of almost a single variable: whether the TABLE 5 Binary Logit of Commute Choice with Geographic Variables Men (n = 132) Women (n = 217) variable B Exp b Significance B Exp b Significance Constant –1.110 .329 .081 –.382 .682 .471 Commute distance –.010 1.010 .759 –.026 .975 .370 Number of bus routes .101 1.106 .667 –.139 .870 .506 Density high (d3) –.248 .780 .637 –.097 .908 .807 Density medium (d2) –.031 .970 .948 .205 1.228 .561 Cox and Snell R2 .003 .009 Nagelkerke R2 .005 .013 Note: 0 = transit, 1 = car. TABLE 6 Binary and Logit Model of Commute Choice with Demographic Attitude Variables Men (n = 57 ) Women (n = 117) variable B Exp b Significance B Exp b Significance Constant 5.618 275.3 0.219 –1.463 0.232 0.497 Pay level 1a –2.938 0.053 0.121 1.621 5.057 0.371 Pay level 2 –3.208 0.04 0.065 2.131 8.421 0.099 Pay level 3 (highest) — 2.046 7.738 0.109 Statement: Bus stop is safe –0.089 0.914 0.93 –0.329 0.72 0.498 Statement: Bus stop is pedestrian friendly 0.052 1.054 0.956 –0.092 0.912 0.859 Household size (1) –1.101 0.332 0.471 0.622 1.864 0.476 Household size (2) –0.798 0.45 0.59 –0.167 0.846 0.829 Household size (largest) 1.372 3.943 0.367 –1.03 0.357 0.213 Any children <16 (binary) –0.406 0.666 0.724 0.156 1.169 0.819 Age (chronological) 0.41 1.507 0.387 0.39 1.477 0.122 Statement: Get more physical exercise 0.727 2.066 0.067 –0.266 0.766 0.177 Statement: Hate ties to fixed schedule 0.398 1.489 0.259 –0.149 0.862 0.43 Statement: Drive to UCLA = more than twice a week –1.12 0.326 0.015 –362 0.696 0.052 Statement: Buses are too crowded –0.276 0.388 0.588 0.003 1.003 0.989 Statement: LA transit has come a long way 0.138 1.148 0.696 –0.144 0.866 0.553 Cox and Snell R2 0.37 0.21 Nagelkerge R2 0.5 0.28 Note: 0 = transit, 1 = car, — = no cases. a Number of cases < 7.

160 WOMEN’S iSSUES iN TRANSPORTATiON, vOLUME 2 respondent drove to work at least twice a week throughout the trial. For women, there was more modest improvement, from a baseline classification of 61% to 66%. Gender-related variables, such as household size and the presence of children, were not decisive factors. Even the statement “i hate to be tied to the fixed schedule of transit” was not a good predictor. Men tended to cite more the exercise benefits of taking transit, and salary level was somewhat associated with the transit benefit, particularly for women. The most important result is the influence of continuing to drive a vehicle to work during the 12-week trial. These respondents, both male and female, were far more likely to leave transit at the end of the trial. Another way of stating it is that they never invested in transit to begin with. One can only speculate on why some participants con- tinued to drive throughout the trial. Perhaps they were less invested in the trial to begin with, and after they signed up, never made the necessary schedule changes or accommodations to successfully use transit. That is a psy- chological interpretation of their behavior. A more objec- tive approach is to say that there was some unmeasured variable—these participants had needs or preferences that could not be accommodated during the transit trial, and what these were was not captured in the surveys. A third interpretation, an economic one, is that these participants were never fully vested in the trial; perhaps they were really on vacation away from work for much of the summer and did not commute at all, or perhaps they had settled on a carpool as an alternative. QualiTaTive analysis of The TransiT– sov decision Because the binary logit analyses could not identify why these few participants continued to use their automo- biles during the trial, the written, open-ended survey responses were examined for clues. Perhaps respondents would explain, using their own words, some concept or idea, to explain their dependence on driving. A small textual analysis was carried out of written responses of 41 participants who wrote in a comment and responded that they frequently drove to campus during the experiment. Through this short textual analysis, it was observed that the most frequently cited reason by women was that their work schedule did not suit the bus schedule, particularly for weekends and evenings, or that they never “got into” transit because they used a different mode or traveled infrequently. For men, the most frequent reason was that they needed a car at work or during the daytime. The written responses are reported in Table 7. The qualitative analysis suggested that this could be a promising avenue for future research: men and women voice, in their own words, the characteristics of transit rid- ing. in this short analysis, women were more likely to cite atmosphere or onboard conditions, and men mentioned time and schedules. The qualitative analysis was expanded to include the written comments of all respondents who par- ticipated in the social experiment and used the open-ended question. A textual analysis was used to identify keywords, but only those that did not have ambiguous meaning (e.g., “flexibility” or “time” were excluded because they are con- text-dependent and would have required human coders, e.g., during this “time” i take the bus, the “time” i wait, i take a “flexible” bus, my schedule is “flexible”). The state- ments were assigned on the basis of the keywords, and only the first, or primary, reason was coded per statement. it appears that the descriptors used by men and women do vary (see Table 8). Women actually mentioned the savings from the program more frequently than men and also mentioned concerns that might be captured in bus- stop safety; namely traveling during the fall, when it was colder, wetter, or darker. A few women mentioned a car- pool as an alternative to transit, but men did not. Most of the statements by the men would have involved coding in context for terms such as “travel time,” “flexibility,” and “schedules.” summary and conclusions This study tracked 381 commuters, 144 men and 237 women, who chose to give up their drive-alone commut- ing for a 3-month period, in return for a free transit pass. Many of these individuals used the pass only for com- TABLE 7 Frequency of Mention of Keywords Used Just by Those Who Drove Frequently Men (n = 12) Women (n = 29) Keyword Number Percent Number Percent Not financially viable 2 17 3 10 My work schedule π bus schedule 2 17 7 24 Mobility needs; parking 4 34 3 10 Reliability of bus; travel time 3 25 3 10 Bus seating; comfort on bus 1 8 4 14 Used a different mode; part time 6 21 Don’t know; keep parking spot; other 4 14 Time change in fall (gets darker) 1 3

161COMPARiNG WOMEN AND MEN COMMUTERS WHO CHOOSE TRANSiT mute trips, but others tried it for other types of travel. Although there are differences, what this population shares in common is that they all owned vehicles and for transit companies, they are so-called choice commuters. Because of gender-related responsibilities, different patterns were expected in how men and women might choose, and then use, the free transit pass. Although women ride buses more than men, it is not clear that this effect would be found in this California study of middle- income women with full-time jobs. That is because their many responsibilities and roles might lead them to favor the flexibility and convenience of an automobile. in this study, few differences between men and women were actually found. Much of this can be attributed to self-selection. When the respondents initially signed up, they had several days to complete the enrollment process for the driving-transit exchange. During this time, they were likely to look up the bus routes, times, and sched- ules and examine how feasible it was for them to use transit. it is probable that both the men and women who opted into the program had successfully prescreened themselves and were able to make adjustments, say, in travel time or departure–arrival schedules. At Stage 1, there were no statistically significant differences between the demographics of the men and women who enrolled. The groups were similarly matched on salary, household income, and so on. The men and women were also simi- lar in geographic factors, such as transit time and travel distances. Even those with longer travel times did not ultimately return to SOv commuting. in the second part of this analysis, survey data were used to explore whether the male and female respondents reported different attitudes and opinions about using transit. Both men and women rated the transit program favorably and said that taking the bus reduced stress. Neither men nor women were more likely to view where they waited for the bus as dangerous or unsafe. in the third part of this study, two separate binary logit regressions were used to explore factors that might explain whether a respondent returned to single-occupancy driving or stayed with transit after the experimental period. Overall, just 29% of the men and 31% of the women returned to the drive-alone mode. None of the geographic factors, such as density or travel time, were good predictors of the decision to return to SOv commuting. According to the self-reported survey data, neither were factors such as the presence of chil- dren in the household, household size, or perceived travel time. The logit presented an interesting result, nonetheless, which is that the 13% of respondents who continued to drive during the trial were among those who then tended to reclaim parking at the end of trial. it is not known whether the reason is that they could not make the accommodation or would not make the accommodation. A few gender-related differences were perceived pri- marily in the open-ended analysis of write-in data. Women commented more about the crowding and discomforts of taking transit, and there was a hint that they were con- cerned about bus stop safety. Men focused more on their car and how they used it. There is a caution to overinter- preting the qualitative results, because some people may have a propensity to write more if English is their first language, and others will skip write-ins entirely. There has been speculation in transportation studies that women and men are becoming more alike in their travel for certain markets and for certain trip character- istics (23). This study shows that similar travel choices are made by male and female commuters who are fairly matched on salary, household size, geographical loca- tion, travel distance to work, and other characteristics. The key in this study, is that participants self-selected a new travel mode (transit), and those that self-selected had smaller households, less need for a car during the day, and few or no children under age 16. Perhaps these factors were “transit enablers.” Surprisingly, once people preselected, distance and travel time were not important factors in their final choice of commute mode. This study lends strong credence to a social experiment approach to transit marketing. More than two-thirds of both the men and the women remained as transit riders when the time came to make a commitment to continue. references 1. South Coast Air Quality Management District (SCAQMD). UCLA Compliance Survey. Transportation. 2008. TABLE 8 Qualitative Analysis of Keywords Used in Open-Ended Comments About Mode Choice Decision in the Larger Sample Male (n = 41) Female (n = 73) Keyword Number Percent Number Percent Safety; darker; time change; crime 4 10 13 18 Children; son; daughter; kids; pregnant; elder care 3 7 6 8 Carpool 0 0 5 7 Household chores; errands; extra trips 0 0 0 0 Gas price; savings; economic; cost 5 12 18 25 Other words 29 70 31 42

162 WOMEN’S iSSUES iN TRANSPORTATiON, vOLUME 2 2. Cooper, C. Successfully Changing individual Travel Behav- ior Applying Community Based Social Marketing to Travel Choice. Presented at 88th Annual Meeting of the Transpor- tation Research Board, Washington, D.C., 2009. 3. Fujii, S., and R. Kitamura. What Does a 1-Month Free Bus Ticket Do to Habitual Drivers? An Experimental Analysis of Habit and Attitude Change. Transportation, vol. 30, No. 2, 2003, pp. 81–95. 4. Taylor, M. voluntary Travel Behavior Change Programs in Australia: The Carrot Rather Than the Stick in Travel Demand Management. Journal of Sustainable Transpor- tation, vol. 1, 2007, pp. 173–192. 5. Rosenbloom, S., and E. Burns. Gender Differences in Com- muter Travel in Tucson: implications for Travel Demand Management Programs. in Transportation Research Record 1404, TRB, National Research Council, Washing- ton, D.C., 1993, pp. 82–90. 6. Wachs, M. Men, Women and Wheels: The Historic Basis of Differences in Travel Patterns. in Transportation Research Record 1135, TRB, National Research Council, Washington, D.C., 1987, pp. 10–16. 7. Madden, J., and M. White. Women’s Work Trips, An Empirical and Theoretical Overview. Women’s Travel Issues: Conference Proceedings and Papers, U.S. Depart- ment of Transportation, 1978, pp. 201–242. 8. Crane, R. is There a Quiet Revolution in Women’s Travel? Revisiting the Gender Gap in Commuting. Journal of the American Planning Association, vol. 73, No. 3, 2007, pp. 298–315. 9. Pisarski, A. Commuting in America III. Transportation Research Board of the National Academies, Washington, D.C., 2006. 10. Pucher, J., and J. Renne. Socioeconomics of Urban Travel. Transportation Quarterly, vol. 57, No. 3, 2003. 11. McGuckin, N., and E. Murikami. Examining Trip Chain- ing Behavior: Comparison of Men and Women. in Trans- portation Research Record: Journal of the Transportation Research Board, No. 1693, TRB, National Research Council, Washington, D.C., 1999, pp. 79–85. 12. McGuckin, N., and N. Srinivasan. Journey to Work: Trends in the United States and Its Major Metropolitan Areas. FHWA-EP-03-058. FHWA, U.S. Department of Transportation, 2003. 13. McGuckin, N., Z. Zmud, and Y. Nakamoto. Trip-Chain- ing Trends in the United States: Understanding Travel Behavior for Policy Making. in Transportation Research Record: Journal of the Transportation Research Board, No. 1917, Transportation Research Board of the National Academies, Washington, D.C., 2005, pp. 199–204. 14. McDonald, N. Does Residential Density Affect the Travel “Gender Gap”? in Conference Proceedings 35: Research on Women’s Issues in Transportation: Report of a Con- ference; Volume 2: Technical Papers, Transportation Research Board of the National Academies, Washington, D.C., 2005, pp. 68–75. 15. Loukaitou-Sideris, A. is it Safe to Walk Here? Design and Policy Responses to Women’s Fear of victimization in Public Places. in Conference Proceedings 35: Research on Women’s Issues in Transportation: Report of a Con- ference; Volume 2: Technical Papers, Transportation Research Board of the National Academies, Washington, D.C., pp. 2004, pp. 102–112. 16. Loukaitou-Sideris, A., A. Bornstein, C. Fink, L. Samuels, and S. Gerami. How to Ease Women’s Fear of Transporta- tion Environment: Case Studies and Best Practices. Report 09-01. Mineta Transportation institute, 2009. 17. Rosenbloom, S., and E. Burns. Why Working Women Drive Alone: implications for Travel Reduction Pro- grams. in Transportation Research Record 1459, TRB, National Research Council, Washington, D.C., 1994, pp. 39–45. 18. FY 2002 On-Board Bus Survey. Metro, Los Angeles County Metropolitan Transportation Authority, Oct. 2002. 19. Sarmiento, S. Household, Gender and Travel. 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. 36–52. http://www.fhwa.dot.gov/ohim/womens/chap3.pdf. 20. Novaco, R., and C. Collier. Commuting Stress: Rideshar- ing and Gender: Analyses from 1993 State-of-the-Com- mute Study in Southern California. in Transportation Research Record 1433, TRB, National Research Council, Washington, D.C., 1994, pp. 170–176. 21. UCLA Staff Affirmative Action Office. Campus Wide Work- force Demographic Profile. http://www.upte.org/dwc/saa doc-demographicdata-2005.pdf. Accessed March, 2009. 22. van Wee, B. Self-Selection: A Key to a Better Understand- ing of Location Choices, Travel Behavior, and Transport Externalities. Transport Reviews, vol. 29, No. 3, 2009, pp. 279–292. 23. Gossen, R., and C. Purvis. Activities, Time, and Travel. Changes in Women’s Travel Time Expenditures 1990– 2000. in Conference Proceedings 35: Research on Wom- en’s Issues in Transportation: Report of a Conference; Volume 2: Technical Papers, Transportation Research Board of the National Academies, Washington, D.C., 2005, pp. 21–29.