The impact of COVID-19 pandemic on travel preferences by students in Padang City

. The use of public transportation after implementing the Level 3 Community Activity Restrictions (PPKM) policy during the COVID-19 pandemic has become a challenge for students. This study aims to determine the effect of applying PPKM rules in using public transportation among students using descriptive statistics compared to conditions before the pandemic and determining the probability of high school students in Padang using public transportation during PPKM rules. Data were obtained from 425 respondents using a web-based questionnaire survey. The pandemic had an impact on reducing student allowances per week by 26% because school lasted three days out of 6 days per week; there was a decrease in the frequency of using public transportation 4-9 times/more per week by 22.4% The opposite occurs in the use of private vehicles. Travel with a frequency of >7 times/week increased by 31.4% during the pandemic. The resulting utility model is expressed in multiple linear regression equations. The utility model of using public transportation for high school students in Padang: Y(AUK-AUO) = 0.577 + (-0.039) Δtime + (- 0.072) Δcost +0.848 health protocols applied; (R 2 =0.26). Furthermore, the travel fare difference attribute was the most influential attribute of students' probability to use conventional public transportation.


Introduction
Use The COVID-19 pandemic has impacted various aspects of life, including the decline in the movement of public transportation modes.During the early days of the Covid-19 pandemic in Indonesia, in March-April 2020, the Indonesian government implemented a Large-scale Social Restriction (PSBB) policy.The policy, which was followed up by massive socialization to the community to work from home, study from home, and worship from home, and the closure of tourist sites had limited the movement of people outside their homes.Restrictions placed on public transportation are related to the number of passengers and operating hours.Public transportation in the DKI Jakarta area decreased significantly during the COVID-19 pandemic.TransJakarta services during April 2020 (until April 15th,2020) decreased by around 83,000 people daily [1].
The impact of the COVID-19 pandemic on a significant decrease in the use of public transportation has also occurred in cities around the world.A large-scale survey conducted with individual participants in Japan found the most significant reported change as a decline in the use of public transit (36%), with a resulting increase in car trips (29%) and walking and cycling (27%) [2].The pandemic has caused 19% of Indian bus operators to lose 90% of their passengers and the remaining 91% without passengers at all [3].The decline in Paris was about 80% compared to before the pandemic [4].
Starting June 1, 2020, the Indonesian government implemented the Large-Scale Social Restrictions Transition (PSBB Transisi) or new normal habit policy, whereby public transport operators are only allowed to load with 50% occupancy.The results of Kurniati and Valentino's research during the implementation of this rule, the frequency of use of public transportation before the pandemic was classified as frequent/often with an index value of 60.86%; during a pandemic, it fell to 44.48% with a rare interpretation.The type of public transportation often used is online transportation (motorbike and car) by 62.1% [5].
The Indonesian government implemented Enforcement of Community Activity Restrictions (PPKM); PPKM consists of level 1 (low incident), level 2 (medium incident), level 3 (high incident), and level 4 (very high incident) levels up to four-level PPKM consisting of level 1 (low incident), level 2 (medium incident), level 3 (high incident), and level 4 (very high incidence).When the research survey was carried out (after October 18 th, 2021), the city of Padang implemented PPKM level 3 [6].
The rules also explain the rules for riding public transportation.There is a slight difference from the previous rules regarding the passenger capacity of public transportation in the PPKM Level 3 area.The rules for public transportation (public vehicles, mass transit, taxis (conventional and online), and rental/rental vehicles.
The rules regarding limited face-to-face learning (PTM) are regulated in a Joint Decree (SKB) of 4 Ministers regarding implementing learning during the COVID-19 Pandemic.Limited PTM can only be done at schools in the PPKM area level 1-3.Limited PTM in schools is carried out through two phases: a transition period that lasts two months from the start of little faceto-face learning at school and a new habit period that begins after the transition period is complete [6].
Head of DKI Jakarta Transportation Agency (Dishub) Syafrin Liputo revealed an increase in community mobility while implementing the Community Activity Restrictions (PPKM) policy.The volume of motorized vehicle traffic increased on January 11-31 or during the PPKM period compared to October 12-November 1 2020, during the transitional period of Large-Scale Social Restrictions (PSBB).The increase in citizen mobility is based on two things, namely, the volume of motorized vehicles and users of public transportation."The volume of motor vehicle traffic has increased by 12.18%" [7].
The result of research by Yosritzal et al. (2022) shows that the probability of using conventional public transportation decreases when travel time and travel fares increase compared to online transportation.The attribute that significantly influences people's probability of using conventional public transportation is travel fares [8].
Several existing studies and news have attracted researchers' interest in knowing the movements of the community, especially school students when PPKM level 3 was implemented in the city of Padang.

Data collection
The data collection was carried out through online questionnaire interviews.The school helped distribute questionnaires through social media.Students who become respondents are users of conventional or online public transportation.Determination of the number of samples using a simple random sampling technique.Based on the Slovin formula [9] with a population of 43,521 high school students [10] and an error of 10%, the minimum sample size is 100 people.
Questionnaires are created in Google Docs forms.Questionnaire questions are designed into two (2) parts.The characteristic of the questions for the two parts of the questionnaire is closed questions.The first part is a question regarding the socioeconomic and travel characteristics of respondents before and during the pandemic.The socioeconomic characteristics consist of gender, vehicle ownership, and pocket money before and during the pandemic.
The questions of travel characteristics are reasons for using modes, frequency of mode use, the purpose of the trip using the mode, travel time to the destination, and the waiting time to get public transport.
In section 2, the questions are designed using a stated preference technique.Respondents were offered alternative (Alt) trips with the attributes of cost difference, travel time difference, and application of health protocols using conventional and online public transportation.Each attribute consists of two (2) levels, and there are eight (8) alternative questions posed to the respondents [11].
The travel time attribute level is the difference in travel time (Δtime) using conventional and online public transportation by 10 minutes and 20 minutes.The difference in travel fares (Δcost) with level 1 is Rp.5,000, and level 2 is Rp.7,500.The level of application of health protocols consists of implementing and not implementing health protocols.Respondents will then choose to use a point rating technique with a four-point semantic scale, namely: 1 = very unwilling to use conventional public transportation; 2 = unwilling to use conventional public transportation; 3 = willing to use conventional public transportation; 4 = very willing to use conventional public transportation.Table 1 displays the part 3 questionnaire.

Data processing
Characteristics and travel data of respondents were processed with descriptive statistics.Stated Preference data processing uses the binomial logit method to transform the semantic scale into a numerical scale.The multiple linear regression equation is mathematically expressed by: Where Y is the respondent response, a is a constant; b 1 , b 2 , b 3 is the regression coefficient value, and X 1 (Δtime), X 2 (Δcost) and X 3 (health protocols) is the independent variable.
To resulting probability value of each attribute can be calculated from equation 2 [12]. (2)

Result and discussion
The survey results obtained data from 425 respondents.Respondents can be classified based on the transportation used for movement before and during the pandemic, as shown in Table 2.

Respondents characteristics
The personal characteristics of respondents, as shown in Table 3.Of the survey respondents, 64% were female, 77% owned 1-2 motorbikes, and 52.9% did not own a private car.Table 4 shows an increase in the number of students (21%) who experienced a reduction in pocket money to <Rp 50,000 during the pandemic.The use of public transportation to travel to the house during the pandemic, for the type of paratransit, Trans-Padang Bus, and online taxis, has increased while other public transportation has decreased.For other travel destinations, there has been an increase in all types of public transportation, Table 5.The use of public transportation during the pandemic has increased by 1-3 times/week, while the remaining three frequency classes have decreased, table 6.The frequency of students traveling using public transportation has decreased during the pandemic.The opposite occurs in the use of private vehicles.Travel with a frequency of >7 times/week increased by 31.4% during the pandemic (table 7).Table 8 displays the waiting time to get public transportation based on the type of public transportation before and during the COVID-19 pandemic.In general, the waiting time for public transportation is longer.Data shows that waiting times of more than 5 minutes are greater during the pandemic than before.The most significant change in waiting time for public transportation occurred in paratransit.The waiting time to get paratransit less than 5 minutes was reduced by 16.9%.There was no significant change (more than 10%) in waiting times for other types of public transport before and during the COVID-19 pandemic for other types of public transport.

Preferences for using public transport
The result of processing the stated preference questionnaire data is the utility model, a function of the difference between conventional public (AUK) and online public transportation (AUO).
The results of the model using all respondents are Y(AUK-AUO) = 0.385 + (-0.031)Δtime + (-0.058)Δcost +0.712 health protocols; R 2 value is 0.199.The results of data processing using Microsoft Excel show the results of Pvalue with the conclusion that travel time, travel costs, and the application of health protocols affect students' preference for using public transportation.
The utility model is based on respondents who used public transportation before and during the pandemic are Y(AUK-AUO) = 0.577 + (-0.039)Δtime + (-0.072)Δcost +0.848 health protocols; R 2 value is 0.267.Based on the results of the P-value test, the attributes of travel time, travel costs, and health protocols have a significant effect on the preferences of students who use public transportation.
Probability analysis of changes in independent variables by performing sensitivity analysis.Sensitivity analysis of travel cost difference, i.e., when the difference in travel cost is zero, transportation modes' users prefer AUK over AUO with a probability of 0.640 or 64% (Fig. 1).The more significant the difference in travel fares, the probability of students using the AUK mode of transportation will also be more negligible.
Sensitivity analysis of travel time attributes is when the difference in travel time is zero, users of transportation modes prefer conventional public transportation (AUK) compared to online public transport (AUO) with a probability of 0.64 or 64%.The more significant the difference in travel time, the probability of students using the AUK mode of transportation is negligible.Sensitivity analysis of health protocols applied, namely that the probability of users of transportation modes choosing AUK is greater when AUK applies a health protocol with a percentage of 65%.So, the probability of students using AUK is more significant when the mode of transportation applies the health protocol.

Conclusions
The pandemic had an impact on reducing student allowances per week by 26% because school lasted three days out of 6 days per week; there was a decrease in the frequency of using public transportation 4-9 times/more per week by 22.4% The opposite occurs in the use of private vehicles.Travel with a frequency of >7 times/week increased by 31.4% during the pandemic.The purpose of using public transportation increased for trips to the house.The highest percentage change in waiting time to get public transportation occurred in paratransit, with a waiting time of <5 minutes, which was 16.9%.There is no change more significant than 10% in the waiting time to get public transportation before and during the COVID-19 pandemic for other types of public transport for all the waiting time intervals.words.
The resulting utility model is expressed in multiple linear regression equations.The utility model of using public transportation for high school students in Padang after the implementation of the PPKM level 3 policy was obtained as follows: Y(AUK-AUO) = 0.577 + (-0.039)Δtime + (-0.072)Δcost +0.848 health protocols applied; (R 2 =0.26).Furthermore, the travel fare difference attribute was the most influential attribute of students' probability to use conventional public transportation.

Table 3 .
Personal Respondents Data.

Table 5 .
The use of public transport based on the type and purpose of the trip.

Table 6 .
Frequency of use of public transport.

Table 7 .
Frequency of use of private vehicle.

Table 8 .
The waiting time to get public transport.