key: cord-290921-dozqofrm authors: Tan, Limin; Ma, Changxi title: Choice behavior of commuters' rail transit mode during the COVID-19 pandemic based on logistic model date: 2020-09-19 journal: nan DOI: 10.1016/j.jtte.2020.07.002 sha: doc_id: 290921 cord_uid: dozqofrm To understand whether commuters will take rail transit during the COVID-19 pandemic, a logistic regression model was constructed from three aspects of personal attributes, travel attributes and perception of COVID-19 based on 559 valid questionnaires. The results show that: occupation, commuting tools before the COVID-19 pandemic, walking time from residence to the nearest subway station, the possibility of being infected in private car and the possibility of being infected in public transport have significant influence on the commuters’ choice of rail transit. Self-employed people and freelancers, commuters who used non-public transport before the COVID-19 pandemic, and commuters who take longer to walk from their residences to the nearest subway station are less likely to commute by rail transit during the COVID-19 pandemic. Commuters who think that the risk of being infected with the virus in public transport is higher have a lower probability of choosing rail transit. The confidence in bus/subway/taxi/taxi-hailing of commuters who do not choose to commute by rail transit during the COVID-19 pandemic is not high. The study of this paper can provide reference for the formulation of urban rail transit control measures during the COVID-19 pandemic, so as to formulate more perfect measures to ensure the safety of the returning workers. The dependent variable is whether to "choose rail transit", and " = 1" means to choose rail transit, Logit = ln ( /(1 − )) = + (1) where is the probability that = 1, is each explanatory variable, is a constant term, is the 144 regression coefficient of the model. perception of COVID-19. In order to improve the quality of the questionnaire data, this paper will inaccuracy of the data caused by the psychological defense of the respondents, the respondents were 152 informed that the data collected were for scientific research only at the beginning of the questionnaire. In 153 order to reduce the unconscious self-concealment of the respondents, the respondents are reminded to 154 make a rational choice before the questionnaire is started, and the questionnaire response time is set to 155 be no less than 3 min to ensure that the interviewees have enough time to think rationally. The specific 156 contents of this questionnaire are as follows. (1) Personal attributes: gender, age, education, occupation, personal monthly income. (2) Travel attributes: distance from residence to workplace, walking time from residence to the nearest 159 subway station, times of transfers required to get to workplace by rail transit, and commuting tools 160 before the COVID-19 pandemic. (3) Perception of COVID-19: the degree of understanding of the COVID-19 pandemic, the degree of concern about the COVID-19 pandemic, the degree of anxiety about the COVID-19 pandemic, the which is used to measure the respondents' attitude, view, evaluation or intention towards a certain thing. The scale level is set to 5 to make the options have a certain degree of discrimination, while ensuring 169 that the survey respondents will not be fatigued. indicating that most respondents believe that the risk of being infected in public transport and 248 taxi/taxi-hailing is high during the COVID-19 pandemic. The closer the value is to 100%, the more fitting the model is. In the model, the fitting accuracy of the 340 overall data is 93.4%, indicating that the model is good. In summary, the results of the model are 341 credible. 342 Table 4 The result of the test for the logistic model. It is speculated that commuters who distrust rail transit have low trust in bus and taxi or taxi-hailing. The results of the model can provide a reference for the formulation of urban transit management and Using logistic regression to estimate the influence The Latest News of Epidemic Prevention and Control in Suzhou Passengers of Beijing Metro Can Query the Carriage Full Load Rate in Real Time 479 and Less than 30% of Passengers Travel During the Morning Rush Hour Logistic Regression Models: Methods and Application The Influence of Weather Factors on Travel Characteristics of Urban Shared Bike Introductory Econometrics Analysis of influencing factors of traffic jam in Chongqing Bus operation in Chengdu Unconventional prevention strategies for urban public 514 transport inthe COVID-19 epidemic: taking Ningbo City as a case study Emergency response strategies for urban traffic in stages of public health security