An Estimation Of The Value Of Travel Time Saving Of Ho Chi Minh City’S People Based On Stated Preference Data_NguyenSyCuong_2018_00051000343
LITERATURE R E V IE W
Research shows that travel time savings account for approximately 80% of the benefits in transportation infrastructure projects (Welch & Williams, 1997; Fosgerau & Jensen, 2003) Recognizing the importance of Value of Travel Time Savings (VTTS), numerous studies have developed theories and methodologies for its estimation The UK pioneered VTTS estimation in 1965 through traveler choice modeling, although initially using a simple observational approach (Beesley) Subsequently, Quannby (1967) introduced formal statistical analysis based on traveler choices, leading to successful national-scale VTTS estimates using data by Rogers et al (1970) Today, two primary methods prevail for VTTS estimation: Revealed Preference (RP), which analyzes actual travel choices and utilizes Multi-Nominal Logit models, and Stated Preference (SP), which relies on survey-based data to assess travelers' hypothetical choices.
The SP survey dataset requires respondents to choose between two preferred modes based on specific factors such as travel time and travel cost Since the data involves binary mode choices, a Binary Logit Model is the appropriate analytical method to understand preferences This model helps identify how different factors influence decision-making in transportation mode selection, providing valuable insights for transportation planning and policy development.
Many researchers consider the Revealed Preference (RP) method to be the most scientific approach for analyzing traveler behavior under ideal conditions, with numerous successful estimations of Value of Travel Time Savings (VTTS) using this method, as demonstrated by Kato et al (2011) who estimated VTTS for road users in Japan based on a large RP dataset However, some scholars criticize the RP method for its limitations, such as potential bias in estimates due to limited attributes of available services, as highlighted by Calfee et al (2001) and Louviere et al (2000) When the necessary "trade-off" information is absent in observed data, valuable insights are lost To address these issues, the Stated Preference (SP) method was developed, and since the 1990s, it has been widely adopted by researchers to overcome the drawbacks of RP analysis.
The SP (Stated Preference) method has been validated through research by Louviere et al (1980), Hensher and Truong (1983), Bates (1984), and Wardman (1998), affirming its reliability in choice modeling Additionally, Ben-Akiva and Morikawa (1990) proposed a combined RP (Revealed Preference) and SP method to enhance accuracy in transportation studies More recently, Kato et al (2015) analyzed data from Japanese road users and found that, in some cases, the RP method can provide more accurate estimations than the SP method, highlighting the importance of context in transportation preference research.
Developed countries have established technical guidance for estimating Value of Travel Time Savings (VTTS) using stated preference (SP) or revealed preference (RP) methods, as highlighted by Andrew Daly et al (2011) However, most developing countries rely on less scientific approaches, such as estimates based on gross domestic product (GDP) or regional domestic product (RDP), due to technological and economic limitations, offering inexpensive and straightforward calculation methods but failing to reflect actual user transportation behavior Several researchers have used statistical data to estimate VTTS in Asian developing regions, including Western Visayas and Davao City in the Philippines (Roxas and Fillone, 2016; Ninomiya et al., 2017) and Yangon in Myanmar (Kato et al., 2010-2011) In Vietnam, a developing country, most VTTS estimations have also relied on GDP or GRDP figures Recently, Binh Phan Le et al (2016) successfully estimated VTTS for Hanoi using SP data, marking the first such study in Vietnam, with an estimated VTTS of 257.4 VND per minute—about 43.7% of the average wage rate Their findings indicated that higher-income individuals are willing to pay more for time savings With numerous transportation projects underway in Ho Chi Minh City, Vietnam's largest city, it is crucial to conduct VTTS estimations based on user behavior to enhance project effectiveness and validate previous estimates from Hanoi, thereby contributing valuable insights to the literature on VTTS estimation in developing countries.
M E T H O D O L O G Y
Likelihood íunction
The most widely used method to estimate parameters j3T, j3c, ớ, and k is the Maximum Likelihood Estimation (MLE) approach In this study, the likelihood of the entire sample is calculated as the product of individual observation likelihoods Therefore, the likelihood function is defined accordingly, serving as a cornerstone for parameter estimation and ensuring accurate statistical inference.
PT, j3c , Gk : The unknown parameters in the systematic function
PA n: The probability of individual n choosing mode A
The probability that individual n selects mode B is represented by \( P_{Bn} \), while \( y_{All} \) equals 1 if individual n chooses mode A and 0 if they choose mode B Conversely, \( y_{Bn} \) equals 1 when individual n opts for mode B and 0 when they select mode A These variables help model decision-making behavior, with each indicating the specific mode chosen by the individual, enabling analysis of mode preferences under different conditions.
The other way to vvrite this likelihood function is to make it more convenient by the logarithm íimction as follows:
L L { P t , p c , 9k) = £ ( y Ajl lo g PA n + y B n lo g PB „) (2.6) n=l
It is called that Log-Likelihood function Note that y Bn ^ l - ^ ^ a n d PBn =1 - P An, thereĩore, we have:
L L ( P T, P c , 8 k) = [ y Ajl log PA „ + (1 - y A„) lo g (l - PiJ t) ] (2.7)
To identify the maximum of the log-likelihood function (LL), we differentiate it with respect to each parameter—p_r, p_c, and 6k—and set these partial derivatives equal to zero This process involves solving the resulting equations to determine the optimal parameter estimates that maximize the likelihood By finding these critical points, we ensure the parameters p_r, p_c, and 6k provide the best fit to the data, adhering to essential principles of maximum likelihood estimation.
Which, if they exist, they must satisíy the necessary conditions that:
In first-order conditions, multiple solutions may exist, but only one corresponds to the maximum likelihood function Therefore, the asymptotic variance-covariance matrix of the estimates is determined based on this unique maximum likelihood solution, ensuring accurate inference and parameter estimation.
V2LL is the matrix of second derivatives of the log-likelihood function with respect to the model parameters, evaluated at their true values Each entry in this matrix corresponds to the second partial derivative calculated at these true parameters, with the element in the kth row and lth column representing the curvature of the log-likelihood concerning the respective parameters This information is crucial for understanding the precision of parameter estimates and plays a significant role in statistical inference, such as deriving standard errors and conducting hypothesis tests Properly analyzing the V2LL matrix enables researchers to assess the stability and reliability of their model estimations, making it a fundamental component in maximum likelihood estimation (MLE) and inferential statistics.
The consistent estimator of the true value: ô " õ 2L L ' N
Goodness of f i t
In order to evaluate the suitability of the model, a statistic called the
“likelihood ratio index” is used with binary choice model to measure how well the models fit the data:
LL(J3) is the value of the loir likelihood function at the estimated parameters
LL(0) is its value when all the parameters are set equal to zero
Statistical t e s t
In order to evaluate vvhether the parameters have statistical signiíicant or not, the t-test is used
Where: jBk is the estimate for kth parameter p ' k is the hypothesized value for the kth parameter
Sk is the Standard error of the estimate
D A T A
Overview
This thesis analyzes data from a 2016 survey conducted by the National Traffic Safety Committee, utilizing face-to-face interviews to gather information on demographics, daily activities, and transportation preferences The survey involved 50 trained interviewers who randomly selected and interviewed 1,620 households across 24 districts in Ho Chi Minh City, between November 24, 2016, and January 22, 2017 A total of 4,460 individuals from these households completed response forms, providing valuable insights into transportation mode choices in the region.
The design o f the questionnaire
Questionnaire was designed in Vietnamese and then translated into English This helps the respondents and interviewers avoid misunderstandings, increase the accuracy o f the survey results.
The survey consists o f 14 questions divided into five sections: household information, personal iníbrmation, traffíc safety, the most recent trip data and binary choice o f transportation mode.
The first segment of the survey focuses on collecting basic household information through three key questions Question 1 gathers specific address details and contact numbers for follow-up purposes Question 3 examines household demographics by recording the number of family members, employees, children, and elderly individuals Question 2 investigates the types and quantities of vehicles owned, including motorbikes, cars, trucks, taxis, coaches, and bicycles, with respondents providing detailed information about motorcycles such as the number of vehicles, ownership details, origin, type, brand, engine capacity, age, registration location, and purchase price, along with the gold price at the time of purchase to determine the vehicle’s current value adjusted for inflation This comprehensive data collection enables a thorough understanding of household characteristics and asset ownership.
The second part of the survey includes four questions (questions 4 to 7) directed at vehicle users, focusing on personal and behavioral information Question 4 comprises ten subcategories collecting data on age, sex, occupation, driver’s license status, vehicle type, and usage method (shared, owned, or family vehicle) It also assumes respondents choose among five vehicle types—truck, coach, car, taxi, bicycle, or motorcycle—and if a motorcycle is selected, its position within the family is noted Question 5 gathers information about previous motorcycle ownership, including duration of use, whether it was sold or transferred, and reasons for discontinuation Question 6 examines future motorcycle purchasing plans, with respondents providing their monthly income divided into four groups, and details on preferred brands and models if planning to buy For those not intending to purchase, the survey explores the influence of bank loan interest rates and manufacturer price reductions on their decision Question 7 targets organizational and government officials, asking about daily vehicle usage frequency and opinions on altering work schedules, such as shifting start times and reducing workdays from ten to two or nine days per two-week period.
The third section of the article focuses on traffic safety, encompassing questions about accident history and causes Respondents are asked to report the number of accidents they experienced over the past ten years and to detail the most serious incident, including when it occurred, the type of collision, responsible parties, and primary causes such as alcohol use, red-light running, speeding, careless crossing, braking errors, unsafe following distances, cell phone distraction, or other factors Post-accident treatment was also recorded to assess severity, categorized into four levels: no hospital visit (minor), a single medical examination (moderate), multiple visits (serious), and hospitalization (severe) Additionally, data on subsequent actions, like switching to different vehicle types, motorcycle usage frequency, and average distance traveled by motorcycle, were collected to analyze the impact of serious accidents on individual mobility behaviors.
Part 4 collected the most recent trip history data of the respondents or regular trips of another family member not attend during the intervievv All data such as departure, destination, departure time, arrival time, estimated distance and the used vehicle Respondents also provided information on the stops or transshipments and the cost of each mode If the respondent is a personal vehicle driver, more iníòrmation about the subject, the number of people in the vehicle, the parking space, the type of parking lot and the parking fee will be provided.
This section includes two questions designed to gather information on transportation modes among respondents over 15 years old Question 13 presents an assumption regarding factors influencing the choice among five vehicle options—train, car, motorbike, motorbike taxi, and bus—and requires respondents to select one based on variables such as gasoline price, parking fee, fare, travel time in comparison to the average trip duration of 30 minutes, and walking time to the terminal or bus station Question 14 offers 24 hypothetical scenarios, each detailing specific data on fare, gasoline price, parking fee, walking time to the station, and travel times relative to the last trip, to identify the factors influencing the selection between two vehicles (from the five options) The data collected primarily serve to estimate the value of time using the Social Preferencing (SP) method.
Table 3.1 An example of a choice in the SP experiment
Walking time to the station (minute) 0 15
Travel time compared to the present No change Double
Target area
This survey was conducted in whole Ho Chi Minh City: District 1, District 2, District 3, District 4, District 5, District 6, District 7, District 8, District 9, District
Ho Chi Minh City comprises districts such as District 11, District 12, Phu Nhuan, Tan Binh, Binh Thanh, Go Vap, Tan Phu, Binh Tan, Cu Chi, Binh Chanh, Thu Duc, Nha Be, Can Gio, and Hoc Mon According to the 2015 report by Vietnam's General Statistics Office, the city has a population of approximately 8,136,300 residents living within an area of 2,095.5 square kilometers, resulting in an average population density of about 4,025 people per square kilometer However, population distribution is highly uneven across districts, with core areas like District 1, District 2, and District 5 experiencing very high densities of up to 40,000 people per square kilometer, while other districts have significantly lower population densities.
(98 people/lkm2) in suburban such as Can Gio District, Hoc Mon District
Long An Province Binh Thanh
Characteristics of the collected data
The table 3.2 shows the population density and the number of households surveyed in each district.
Ho Chi Minh City, the most populous administrative unit in Vietnam, boasts a high population density of 4,025 people per km² as of 2015, ranking fourth nationally and being only slightly lower than Hanoi The city’s population is highly concentrated in the downtown area, which covers just 21.1% of the city's land but houses 81.7% of its residents, with districts like Binh Tan, Binh Thanh, Gò Vấp, and District 8 among the most crowded Districts such as District 5, District 11, District 4, and District 3 exhibit the highest densities, exceeding 37,000 people per km² In contrast, suburban districts have much lower population densities, averaging around 600 people per km², which is 16.7 times less than downtown areas and three times the citywide average The population distribution remains uneven within both inner urban and suburban districts, with densely populated districts experiencing densities up to 50,000 people per km², while districts like District 2 and District 9 have densities of approximately 2,300 people per km² The least densely populated suburban districts, such as Nha Be, Cù Chi, and Cần Giờ, have densities below 1,100 people per km² Rapid urbanization has significantly increased the urban population over the years, rising from 71.6% in 1995 to 83.3% in 2015, with most residents engaged in non-agricultural activities, and agricultural workers accounting for just 4.8% of the city's population by 2015.
According to Table 3.2, the number of households selected in each district is relatively similar Binh Chanh, Cu Chi, Thu Duc, Hoc Mon, Go Vap, Binh Tan, Tan Phu, Tan Binh, District 8, and District 12 each had the highest count, with 80 households.
In Can Gio District and Nha Be District, 40 households from each district participated in the survey In other districts, a total of 60 households from each district took part All respondents were chosen through a random selection process to ensure unbiased results.
Table 3.2 The population density and number of household
District N u m b er of household Population Density
Figure 3.2 illustrates the sex distribution of respondents in the collection database and the General Statistics Office of Vietnam’s 2015 data The collected data shows an uneven distribution, with females representing approximately 54.6% and males 45.4% Similarly, the General Statistics Office of Vietnam reports an imbalanced gender distribution in Ho Chi Minh City, with females constituting 52% and males 48%.
Figure 3.2 The proportion of male and female
When comparing the report of General Statistics Offĩce of Vietnam and the collected data, age pyramids (both rural and urban) are the same shape. m Male
Figure 3.3 The structure of Ho Chi Minh City’s population a, Collected data in survey b, Data of GSO of Vietnam
The age pyramid in Figure 3.3a shows a narrowing at the bottom due to declining birth rates over the past 20 years, indicating an aging population It also features a bulge in the working-age group caused by rural-to-urban migration, with a higher proportion of women than men, reflecting a common trend in Southeast Asia In cities like Ho Chi Minh City and Hanoi, this phenomenon is particularly prominent, driven by increased employment opportunities for women in sectors such as garments and services.
According to Vietnamese tradition, the eldest son typically inherits his parents' property in the countryside, which influences family dynamics and land distribution Additionally, modern family sizes are decreasing, with couples having fewer children, leading to a smaller number of men in urban areas who often remain in their hometowns to care for family traditions Figure 4.3b, depicting the age pyramid and sex distribution from the General Statistics Office of Vietnam, closely resembles Figure 3.3a, indicating that the collected data accurately reflects the population structure of Ho Chi Minh City.
Table 3.3 illustrates the occupational distribution of the collected data, with retail workers accounting for the largest share at 20.7%, followed by those in the manufacturing sector Notably, no respondents are employed in the hotel and restaurant industry Other occupational categories have similar participation rates, each under 5% Additionally, 30.5% of respondents are without occupation, primarily consisting of children and the elderly who are not part of the workforce.
Table 3.3 The career ííeld of respondents
STT C a re e r field Percentage (%) N um ber
Figure 3.4 illustrates the occupational distribution of respondents, with the majority working as officers, accounting for approximately 33.8% of the sample Self-employment is the second most common occupation, representing about 27.3%, while students make up around 15% of respondents.
Over 50% of the workforce is composed of individuals in these three primary occupation groups, highlighting their dominance in employment sectors In contrast, other roles—including retirees, unemployed individuals, military and police personnel, housewives, family support workers, business leaders, and part-time workers—represent a smaller proportion of the employed population.
S u p p o rt fo r fannly's b u ssin ess ° S elle r
Figure 3.4 The occupation of respondents
The data presented in Figure 3.5 illustrates respondents' after-tax income distribution across four main groups: low, below medium, above medium, and high Approximately 25.9% of respondents fall into the low-income category ( z ) Total travel time -1.15e-02 1.35e-03 8.501 < 2e-16 >fe * *
The initial model exhibited a low goodness of fit, likely due to respondent bias, prompting the removal of confusing cases such as those including the train option, which reduced the sample size to 8,883 observations Despite the smaller sample, the model's goodness of fit improved to 0.024, with most coefficients demonstrating high statistical significance As shown in Table 4.2, both travel time and travel cost have negative coefficients, indicating that increased travel duration and expenses decrease utility The results reveal that individuals aged 16-60, females, and those in the 13-25 income group are willing to pay more to save travel time Additionally, low-income groups, including officers, sellers, pedestrians, and bike users earning around 0 million, tend to be willing to spend more time traveling to reduce costs, which aligns with their financial constraints Overall, the signs and implications of the coefficients are consistent with those in the first model.
In the third model, the average respondent velocity is limited between 10 km/h and 30 km/h to eliminate impossible observations, and travel distances are constrained from 1 to 30 km, resulting in a reduced sample size of 6,015 observations The goodness of fit improves to 0.034, indicating that outliers previously impacted the results The analysis shows consistent signs for both travel time and travel cost coefficients: individuals aged 16–35, with approximately 0 million monthly income, who are sellers or officers, are willing to spend more travel time to save on travel costs Conversely, those with a monthly income of 13–25 million are willing to pay more money to reduce travel time, which aligns with expected economic behavior.
Table 4.3 Estimation results of the third model
Estimate Std Error z value Pr(> z )
Results o f the VTTS estimation
The estimated VTTS in three models are 762 VND per minute, 723.3 VND per minute, and 859.2 VND per minute, with an average of 781.5 VND per minute, which is approximately 0.83 times the average wage rate of residents in Ho Chi Minh City In comparison, the VTTS for a commuting trip in Hanoi is about 257.4 VND per minute, nearly three times lower than that of Ho Chi Minh City, reflecting Hanoi’s lower GRDP, which is approximately half of Ho Chi Minh City’s GRDP According to the HOUTRANS report, the average VTTS in Ho Chi Minh City is around 630 VND per minute, close to the estimates obtained in this study, indicating consistency Additionally, residents in Southern Vietnam tend to spend more on transportation than those in Northern Vietnam, supporting the reasonableness of the estimated VTTS values.
D iscussion
This study advances the literature on Value of Travel Time Savings (VTTS) estimation in developing countries and enhances understanding of VTTS specifically in Vietnam However, it has limitations, such as a low goodness-of-fit for the binary logit model, indicating the need for alternative VTTS estimation approaches Future research should consider employing advanced models like the mixed logit, revealed preference (RP), or combined stated preference/revealed preference (SP/RP) methods to identify the most accurate VTTS estimation for Ho Chi Minh City.
The Value of Travel Time Savings (VTTS) plays a critical role in cost-benefit analysis frameworks, especially in urban transportation planning Hanoi and Ho Chi Minh City, Vietnam's largest cities, are currently implementing numerous transportation projects to accommodate rapid development and growing demand Both cities utilize VTTS estimates derived from stated preference (SP) data; however, determining the most suitable SP data for VTTS estimation remains a challenge A comparative analysis reveals various advantages and disadvantages of existing SP datasets, providing valuable insights for designing more effective SP surveys This study offers practical recommendations for developing new SP data tailored for binary logit models in Vietnam, enhancing the accuracy of VTTS estimation for urban transportation projects.
First of all, all the personal attributes should be collected, such as age, gender, income and current mode Especially, the income iníòrmation should be written clearly by the respondent.
To ensure accurate results with the SP method, it is crucial to design the data clearly by assigning specific numbers to travel costs and travel times This approach reduces the burden on respondents, making it easier for them to provide precise estimates When respondents are overwhelmed with complex calculations, they may struggle to accurately determine their travel time and costs, which can lead to responses that do not truly reflect their maximum utility Clear and well-defined data design enhances the reliability and validity of the survey outcomes.
Finally, the binary mode choice should have only two vehicle option This can reduce the error component and raise the reliability of the study.
This study estimated the Value of Travel Time Savings (VTTS) for residents of Ho Chi Minh City using an empirical approach The data was collected in 2016 by the National Traffic Safety Committee through face-to-face interviews at respondents’ homes Participants completed binary choice tasks, selecting between options that varied in travel time and cost The VTTS was successfully estimated by applying a binary logit model, providing valuable insights into travelers' valuation of time savings in Ho Chi Minh City.
The study estimates the Value of Travel Time Savings (VTTS) at 781.5 VND per minute, indicating that residents of Ho Chi Minh City are more willing to pay to reduce their travel time compared to those in Hanoi Comparative analysis confirms that this VTTS estimate is reasonable and aligns with regional travel behavior trends These findings highlight the importance of considering city-specific willingness to pay in transportation planning and policy development.
This study enhances the existing literature on Travel Value of Time Savings (VTTS) estimation in Ho Chi Minh City through the use of the Stated Preference (SP) method It provides valuable insights into VTTS estimation specific to Vietnam, contributing to more accurate transportation planning Additionally, the research offers practical recommendations for designing SP data for binary logit models, improving the reliability and validity of VTTS assessments Overall, this study advances both academic understanding and practical application in transportation economics.
This study has certain limitations, including the use of Ihe SP data, which may not always produce superior results compared to the RP method, as highlighted by Kato et al (2015) Therefore, future research should consider incorporating alternative valuation techniques to enhance the robustness of findings.
MP method or combined SP/RP method This can help us raise the reliability o f the
■TTS Secondly, because time is limited, this study just uses the binary logit model
B seems that the error component maybe quite high because of the “bias” Thus, the
The RHE-R model should be evaluated alongside alternative models such as the mixed logit model to ensure accuracy The estimated P/TTS (Passenger Travel Time Saving) can then be integrated into a cost-benefit analysis framework for transportation projects This approach allows for a comprehensive assessment of the project's economic effects and overall viability.
[1] Bates, J J (1984) Values o f time from stated preference data Proceedings of Seminar H, PTRC Annual Meeting, Brighton.
[2] Beesley, M E (1965) The value of time spent in travelling: some new evidence
[3] Ben-Akiva, M and Morikawa, T (1990) Estimation of mode switching models from revealed preferences and stated intentions, Transportation Research Part A, Vol 24, No 6, pp 485-495.
[4] Calfee, J., Winston, c M and Stempski, R (2001) Econometric issues in estimating consumer preferences from stated preference data: A case study o f the value o f automobile travel time, The Revievv of Economics and Statistics, Vol 83,
[5] Daly, A., Tsang, F., & Rohr, c (2014) The value of small time savings for non-business travel Journal of Transport Economics and Policy (JTEP), 48(2), 205- 218.
[6] Fosgerau, M (2006) Investigating the distribution o f the value of travel time savings, Transportation Research Part B, Vol 40, No 8, pp 688-707.
[7] Gvvilliam, K.M (1997) The Value ofT im e in Economỉc Evaluation o f Transport Prọịects: Lessons from Recent Research Inírastructure Notes, Transport No OT-5,
[8] Hensher, D A and Truong, T p (1983) Valuation of Travel Time Savings from a Revealed Preference and a Direct Experimental Approach Research Report Paper No.269, School of Economic Studies, Macquarie University, Australia.
[9] HFIC (2017) Hfìc.vn, http://www.hfic.vn/chi-tiet-bai-viet/31420/100/Tin-Kinh- te-Tai-chinh/Phat-huy-hieu-qua-von-nha-nuoc-cho-cac-du-an-dan.
[10] JICA (2004) “The Study on Urban Transport M aster Plan and Feasỉbility Study in Ho Chi Minh Metropolitan Area ”.
[11] Kato, H., Inagi, A., Saito, N and Thun, p T T (2011) Feasibility analysis for the introduction of a bus rapid transit system in Yangon, Myanmar, Journal o f the Eastem Asia Society fo r Transportation Studies, Vol 9, pp 914-929.
[12] Kato, H., Sakashita, A and Tsuchiya, T (2015) Web-based versus paper- based survey data: An estimation of road users’ value of travel time savings,
Journal o f the Eastern Asia Society fo r Transportation Studies, Vol 11, pp 2502-
This study by Kato et al (2011) estimates the value of travel time savings for road users in Japan by analyzing large-scale household survey data The research provides valuable insights into how travelers value time savings, which can inform transportation planning and policy decisions The findings highlight the importance of understanding user preferences to improve traffic efficiency and enhance overall travel experiences Published in the Transportation Research Record, the article emphasizes the significance of using comprehensive survey data for accurate valuation in transportation studies.
[14] Le BINH, p., ABE, R., & KATO, H (2017) Estimating Value of Travel Time for Commuting Trips in Hanoi, Vietnam Journal o f the Eastern Asia Society fo r
[15] Louviere, J J., Henley, D H., Woodworth, G., Meyer, R J., Levin, I p., Stoner, J w , Curry, D and Anderson, D A (1980) Laboratory simulation versus revealed preference methods for estimating travel demand models, Transportation Research Record, No 794, pp 42-51.
[16] Louviere, J J., Hensher, D A and Swait, J D (2000) Stated Choice Methods: Analysis and Application, Cambridge University Press.
[17] Ninomiya, s Kato, H and Diaz, c E D (2017) Fare system of urban public transportation services in Davao City, Philippines, Presented at the Transportation Research Board 96th Annual Meeting, Washington D c (US), January 2017.
[18] Quarmby, D A (1967) Choice of travel mode for the journey to work: some findings Journal ofTransport Economics and Policy, 273-314.
[19] Rogers, K G., Townsend, G M & Metcalf, A E 1970 Plcinning fo r the
W orkJourney Local Government 0 R Unit Report C67.
[20] Roxas, N R and Fillone, A M (2016) Establishing value of time for the inter-island passenger transport of the Westem Visayas region, Philippines,
[21] Wardman, M A (1998) Comparison of revealed preference and stated preference models of travel behavior, Journaỉ o f Transport Economics and Policv,
Question 1: Current home address: No.: Road:
T ruck Ovvnership type: □ 1 Ovvnership □ 2 Installment purchase □ 3 Provid ed by y o u r co m p an y □ 4 EniỊaạe/borrovv
N u m b er o f light trucks (2.5 ton):
C oach car Ow nership typc: □ 1 O w nership □ 2 Installm cnt purchasc □ 3 Provid ed by y o u r c o m p an y □ 4 Engagc/borrovv
Private car Ow nership type: G 1 O w nership □ 2 Installm ent purchase □ 3 Provided by y o u r c o m p an y □ 4 Engaiỉe/boưovv
Taxi O w nership typc: □ 1 Ovvnership □ 2 Installment purchasc □ 3 P rovided by yo u r c o m p an y □ 4 EngajỊc/borrow
Bicyclc Ovvnership typc: □ 1 O w nership c 2 Installm cnt purchase □ 3 Proviđcd by y o u r co m p an y □ 4 E n g a g e /b o ư o w
Moto rbik e Ovvncrship C ylindcr capacity M a d e by Model Used timc Buying information
□ Provid ed by y o u r c om pany
• Rcgistration location: c H ochim inh City c Oth cr
• Registration location: c H o ch im in h City
□ Pro vided by y o u r c om pany
• P ric e: (mil V N D ) (tacls golci) c S cootc r m otorbike
□ Pro vidcd by yo u r c om pany
Question 3: Number o f people in your family:
• N um ber o f people are labor:
• Number of people >= 60 years old:
(For all members are using transportation vehicle)
• Driver license: □ l.M otorbike □ 2 Car □ 3 Truck
• Vehicles own (quantum): l.T r u c k : 2 Coach: 3 Car: 4 Taxi: i
• What order is your motorbike?: □ l.F irst □ 2 Second □ 3 Third
• If choosing “OW N” or “SHARE”, which kind o f vehicle do you usually use? c l.T ru c k □ 2 Coach □ 3 Car □ 4 Taxi c t
• If choosing “MOTORBIKE”, what order is your motorbike in your family?
Question 5: Information o f your previous motorbike:
• How long did you use it?: (years)
• Is it sold or gave for your family’s member?: □ 1 Sold □ 2
• Why do not you continue using i t ? :
□ 4 Do not have i Bicycle: 6 Motorbike:
3 Do not use family’s vchicle Bicycle □ 6 Motorbike
Question 6: About plan buying a new motorbike in next year (for people have income):
□ 1 < 6 milion VND □ 2 6-12 milion VND □ 3 13-25 milion VND □ 4 > 25 milion V N D 15
• In next year, do you have ap lan for buying a D l Y e s ! 2 No motorbike?
Brand o f new motorbike (ex.: Airblade 2014): l f the answer is “NO
• If the bank loans money for your new motorbike with market interest rate, will you buy with vvhich period below?
• If producer gives new price that less than current price, w ill you buy with w hich discount below ?
• If you buy this motorbike, what brand will you buy? (ex.: Airblade 2014):
Question 7: For govemment and private organization:
• How many times did you use motorbike or car per d a y ?
• Do you want to change the time to start working?
• Do you want to work in 9 days o f two weeks (rest in 2.5 days)?
Question 8: How many accidents have you had within the past 10 y e a rs ?
Question 9: Characteristic of the most serious accident:
• How long has the accident h a p p e n e d ? (years)
• What vvas the vehicle o f the person causing the accident?
□ 1 Drink alcohol □ 2 Pass the red light □ 3 High speed □ 4 Not careful in □ 5 Suddenly brake
□ 6 Too close to other vehicle
□ 7 ư se mobile phone when driving
• Who was caused the accident?
Question 9: The iníòrmation about your latest trip (today or yesterday) or the iníbrmation about the regular trip o f another family’s member who do not attend this survey.
Time/Purpose/Distance Sto p Other inforniation
No Name of the stop
(Only fo r members who are larger than 15 years old)
Question 13: The case below gives assumptions about the íactors that affect the choice one o f the five vehicles given Suppose this is your latest trip, you please choose a most appropriate vehicle among five vehicles below with particular assumptions:
No V ch ic les G a s o lin e p r i c e (V N D /k m )
A ssu m in g that the a vera g e tra v el tim e is 30 m inutes
W a l k i n g tim c to th e s ta tio n ( m in u tc )
Question 14: The numbered hypothetical cases below give the factors that affect the choice o f one o f the two vehicles given Suppose this is your latest trip, you please choose a most appropriate vehicle among fíve vehicles below with particular assumptions:
Travel tim e co m p a red to the present A h a lf Double
W a lk in g tim e to the station (minute) 0 2
Travcl tim e co m p a rcđ to the prcscnt No change A h a lf
W alk in g time to the station (minute) 5 15
Travel tirne com parcd to thc present Doublc No change
W a lk in g time to thc station (minute) 0 15
T ravcl time com pared to thc prcsent D ouble N o change
Travel time c o m p ared to the present A h a l f A half
W a lk in g time to thc station (minute) 0 15
Travel time co m p are d to the present N o change Double
W alk in g timc to thc station (minutc) 0 2
Travcl tim e co m p arcđ to thc prcscnt No changc Double
W a lk in g time to the station (minute) 15 15
Travel time co m p arcd to the prescnt A h alf No changc
\V alking time to the slation (minute) 0 5
Travel time co m p ared to the present Double A h alf
P arking fee (V N D ) 0 0 v / a lk in g timc to thc station (minute) 0 2
T ravel tim e co m p aređ to the present A h a lf A h a l f
W a lk in g tim e to th e station (minute) 0 5
Travcl time co m p are d to the present A h a l f Double
Travel tim e co m p arcd to thc prcscnt A h a l f No chanpc
W a lk in g timc to the station (minute) 15 5
T ravel time co m p ared to the prcscnt No change Double
W a lk in g time to the station (minute) 0 2
T ravel time co m p arcd to the prescnt Double Doublc
W a lk in g time to the station (minute) 0 2
Travel tim e c o m p a rc d to the prcsent A h a l f No change
W alking timc to the station (minute) 0 2
Travcl time com parcd to the prcsent N o ch a n g e A h a l f
W a lk in g time to the station (minute) 0 15
Travcl timc com pared to thc present N o ch a n g e N o c ha ngc
W a lk in g time to the station (minute) 0 2
Travel tim e com parcđ to thc present N o c h a n g e A h a l f
W alking time to the station (minute) 0 5
Travcl tim e com parcd to the present A h a l f N o changc
W a lk in g time to the station (minute) 0 5
T ravel tim e com parcd to the present D ouble A h a l f
T ravcl timc co m p ared to thc present D o u b le D oublc
W a lk in g timc to the station (minute) 0 5
T ravel time co m p ared to the present D o u b le D ouble
W a lk in g time to the station (minute) 0 5
Travel tim e com pared to the present N o ch a n g e N o cha nge