Passenger's perception about city buses in the aftermath of COVID 19: Experience from Guwahati city, India

Abstract

The COVID-19 epidemic has affected public transport in various cities worldwide, including Guwahati city in India, where a 59.09% decline in the ridership of city bus services occurred in autumn 2021. The present work aims to investigate the impact of COVID-19 on travelers' perception and identify the parameters associated with the user's dissatisfaction for improving bus service in Guwahati city. In addition, the study also estimated the revenue loss of bus services due to this pandemic. This was carried out based on survey data from travelers in pre-covid (January 2019) situation, unlocked with 50% seating capacity (November 2020), and unlocked with 100% seat capacity plus some limited amount of standee (January 2021) situation. Results showed an increase in the number of users' priorities in post covid compared to pre covid conditions. Factors such as comfort, fairness, reliability & convenience, and safety are the priority areas of users' pre covid. Whereas bus service attribute such as vehicle condition & hygiene emerged as an additional priority of users post covid. Unfortunately, the bus service in the city performed poorly in terms of hygiene, leading to the shifting of the user to other modes of transport. The poor condition of the vehicles is another major factor that made the users to discard the bus service. This led to huge financial losses of the public transit authorities. The cost analysis result showed that the city bus operators faced a financial loss due to the COVID19 outbreak. The average wage of the drivers reduced from 780 Rs./day to 339 Rs./day in the first unlock phase and 476 Rs./day in the second unlock phase. The present study discusses the introduction of the new route in the city, regular cleaning, regular servicing of the buses, and other recommendations to improve the bus service in Guwahati city.

Graphical abstract

1. Introduction

In December 2019, the outbreak of Coronavirus disease, also known as COVID-19 began in Wuhan, China, and spread rapidly across many countries. The outbreak of the disease was declared as a pandemic by the World Health Organization in March 2020, with China, Italy, Spain, India, and the United States being the most hit. While other epidemics such as Ebola, SARS, or MERS occurred in previous decades, the global economy and society were not confronted with a significant and detrimental impact as caused by covid 19 [1,2]. Many countries have taken various measures to discourage physical interaction to delay the spread of the virus. These steps can be defined as ‘social distancing’ and are particularly effective for diseases such as COVID-19 that are transferred by coughing or sneezing or even exhale of air droplet by the infected person [3]. As a result of this outbreak, many people are temporarily unemployed or have a job from home. This decreased travel demand led to huge reductions in car traffic and public transport in many countries [4,5]. The COVID-19 epidemic has resulted in a worldwide drop in transportation ridership. Urban travel worldwide has fallen, but not consistently for all modes: as survey-based data demonstrate, public transit has been the most affected among the several other modes of urban transportation [6]. This is, of course, just a temporary situation, and we would assume that the demand for outside the home and travel will increase again when the regulation is lifted. At present, the social distancing norms are started to be relaxed, but recent reports of different COVID-19 variants have increased concern that could lead to new waves of social distance in the near future. Moreover, when social distancing laws no longer apply impacting engagement in events and travel, there can still be concerns of social interaction because of the fear of the COVID-19. The government-imposed quarantine measures have significantly reduced travel opportunities across the world and within the country also [7]. Eventually, it will lead to more use of the private vehicles, which will further deteriorate the prevailing traffic conditions. And it is expected that the ridership of the public transport will further reduce with the occurrence of a new wave of COVID-19.

Therefore, it felt necessary to investigate the loss of ridership due to the Covid-19 so that necessary steps could be taken to increase the ridership of the public transport. To increase the ridership, it is required for the policy makers to understand what users are expecting from the transit service and whether there are any differences in the perception of the users between pre-covid and post covid situation. Thus, for successful policy creation and management of transport systems, a clear understanding of the people's travel habits and perceptions of public transit modes during COVID-19 outbreak are required so that necessary steps could be taken to improve the service quality of public transit with respect to this recent Corona situation. The objectives of this study are set as below:

• 1To estimate the ridership loss and revenue loss of the city bus service due to the COVID-19 situation.
• 2To investigate the users' perception of public transport service in post covid situation and compare it with the pre-covid situation.
• 3Understanding the importance of different parameters associated with the dissatisfaction of the users.

2. Literature review

The speedy transmission of the COVID-19 virus, which in several weeks has become an international epidemic, has now been due to the hypermobility in our present way of living, globalization, and Wuhan's connectedness and accessibility [8]. The COVID-19 pandemic has since evolved quickly, ranging from a huge fall in air travel to a never before seen increases in teleworking and its implications on lifestyles and travel around the world [1,8]. Public transportation are most affected mode of transport among the other transportation modes [6]. In some cases, this was accompanied by a reduced service supply and aggravated by the perception that public transport is a more risky option than private or personal transport, due to the closer contact between the individuals in vehicles and as well as stations [9].

Millions of people utilize public transport vehicles every day; they often carry passengers above their capacity, particularly during peak time in morning and evening. This could help to transmit infections among users of public transport [9]. A number of articles have found a strong correlation between early stage mobility data for human beings and the spread of viruses [7] and travel restrictions have played an important role in slowing or minimizing virus spread through various cities. In the course of COVID-19, epidemiologists encourage social distance, which means they should be roughly six feet (or 2 m) or more from one other, like they did in past epidemics and pandemics. Clearly, such measure contradicts the idea of public transport. In view of the fact discussed above, several countries have advised its citizens to avoid public transportation as much as possible. The policy regarding the public transportation in different countries are given in Table 1 :

Table 1.

The policy regarding the public transportation in different countries in Covid period.

Country	Policy
United Kingdom	Avoiding public transportation completely whenever possible and using of alternative modes were encouraged, permitted the use of public transport for essential travel only [9]. In the later stage physical distancing is maintained in the bus stops by marking 2 m spot for passenger to wait. In London many major city-center roads was made closed for the cars to provide more space for pedestrians, cyclists and buses to support the recovery process [10].
Australia	Buses and trains were made to run with a considerably lower number passengers than the seating capacity [9]. In the later stage bus and trains were made to run with seating capacity by mandating face mask. Hand sanitizer were made available at all bus and train stations [10].
China	During the initial phase, Wuhan metro was shut down for two months, Subway system in Beijing was strictly monitored for non compliance of the face masks. Bus capacity is lowered to only 50% in certain places in China, allowing all bus seats to be used while cameras monitor compliance with their capacity [9].
India	Completely closed the public transport system during the first Phase of Covid 19 [11] Just after the lockdown the public transport resumes with a lower occupancy and mandatory use of face mask. Gradually all it was tried to bring the public transport back to normal by Establish protocols and enforcement measures [12].
Spain	Completely closed during initial phase, in some of the cities public transport resumes after the initial lockdown with 30–50% occupancy, mandatory use of face mask. To offset these capacity restrictions, the city is installing emergency bike lanes and develop an inexpensive bicycle model to be manufactured locally as part of the city's economic recovery plan [10]

These policies regarding the public transport have greatly caused serious financial loses of the public transport. For example, Munawar et al. [13] found that the transport sector in the Australia is facing a serious financial loss as the mode share of public transport has dropped nearly 80%. In public transit, people are at highest risk of infectious diseases if they settle or stand closer [14]. These vehicles can be a considerable source of micro-organisms if passengers are coughing and sneezing without covering their mouths [14]. According to epidemiology research, social distancing, internal cleaning and sanitation of public transport vehicles are one of the usual measures provided by the authorities to stop the spread of the micro-organism in the public transport vehicles. In many cities public transport vehicles are disinfected regularly and social distancing norms are followed by allowing lesser number of passengers. It has been found that, the spread of the virus could be controlled with an obligatory usage of masks at public space [9] and this has been reflected in the policy making of public transport in different countries during the Covid-19 outbreak. While it has been found that the use of crowded public transport vehicles may be associated with the acquisition of infectious diseases, these findings are not supportive of the effectiveness of the suspension of mass urban transport systems as a counter-pandemic in reducing or slowing the spread. This is because the risk of spread associated with public transport is relevantly lower than the risk associated with household exposure [15]. Moreover, the spread of the virus could be considerably reduced using the face mask as evident from the previous literatures. However, it is quite expected that choice riders of the public transport modes will shift to other modes preferably private vehicles if no measures have been taken to improve the service quality of public transport [16].

Covid 19 has also impacted the transport sector in India and the public transport are the most affected one which in terms affected a larger number of populations. This is increasingly important in rising economies like India, where a major proportion of society is captive of public transport. Most of the public transport users are found to prefer private vehicles over the public transport mode [4,17] and the recovery of users base of public transport will take longer time due to the fear of Covid 19 [18]. A move from public transport to other private transport modes, as the danger of infection is apparent higher, would worsen current problems of congestion, pollution and casualties within the urban areas due to large imbalances in demand and supplies of urban transport. Since the COVID-19 outbreak, limited findings of research [4,[17], [18], [19]] on the passenger transport system have been available to date and limit the possibility for a complete policymaking process. Therefore, it is necessary to understand users’ needs and expectation from the public transport in the current scenario. During the early stages of the outbreak, Zhang et al. [20] reported the reactions of people in Japan. Some other research focused on behavior changes and related interventions [7]. It can therefore be noted that only a few attempts have been made to monitor behavioral changes since the COVID-19 outbreak, in particular in the area of travel related activities.

From above discussion it is quite evident that after the pandemic also the ridership of public transport will continue to decrease unless serious measures are taken to drastically change the policies for public transport modes. Moreover, no studies till date have compared the pre-covid perception of the users with post-covid situation. This is important as this study will open a different dimension for the policy makers to explore how users’ perceptions are changing in this situation. It will help to adopt suitable policies to mitigate the situation.

3. Study area

Guwahati is selected as a study area for this work. The population of Guwahati is 957,352 as per 2011 Census. In Guwahati, among the registered vehicles, the numbers of two wheelers are found to be highest with a proportion of 52.5% followed by car with a proportion of 27.8% [21]. The proportion of the buses is found to be lowest with a percentage share of 2.7%. Moreover, the average annual growth rate of car and motorized two wheelers are 12.06% and 11.19% while the growth rate of buses is only 2.03% [21]. With the increase in the population of the city, travel demand of the city has increased enormously.

A passenger occupancy survey was carried out on October 2019 (pre-covid) using video photographic survey in different locations (mainly bus stops) of Guwahati city. The bus stops/locations were selected based on the stratified random sampling. In Guwahati, currently buses run on a total no of 16 bus routes. All the prominent bus stops along these routes were listed first.. In this way, a total number of 87 prominent bus stops were identified. From each bus route randomly 2 bus stops were selected for actual survey. The common bus stops across different routes were considered only single time So, a total number of 32 bus stops were considered for the survey. The survey was conducted for a period of 7 days from 7AM to 7PM. Only representative sample (15 min of survey data in each hour) is taken during the occupancy survey. From every location, the occupancy data for all the vehicle types were extracted from the videos taken. This data roughly gives an idea about the average occupancy of different class of motorized vehicles. A similar type of survey was also conducted on November 2021 to have an idea about the occupancy in post-covid situation. The details of the bus routes along with the selected bus stops for the survey is showing in Fig. 1 and Table 2 .

Fig. 1.

City bus routes in Guwahati.

Table 2.

Bus stops considered for the survey along different routes in Guwahati.

Bus Route number	From (Bus stop no)	To (Bus stop no)	Via (Bus stop no)	Bus stops/points considered for survey
1	26	5	25, 21, 20, 19, 34, 10, 9, 8, 7, 6	8, 21
2	18	5	19, 30, 34, 10, 12, 7, 6, 5	18, 34
3	28	5	19, 30, 34, 12, 8, 7	28, 12
4	5	24	7, 8, 12, 10, 34, 32, 33	32, 33
5	5	21	7, 8, 10, 34, 31, 20	5, 31
6	24	3	26, 28, 32, 34, 10, 9, 8, 7	9, 3
7	5	33	7, 12, 16, 22, 23	7, 22
8	4	33	7, 8, 9, 10, 34, 31, 19, 23	4, 23
12	33	17	24, 29, 4, 7, 8, 11, 13, 15	11, 17
18	4	26	26, 29	26, 29
19	19	5	16, 15, 13, 8, 7	13, 16
21	8	24	11, 13, 15, 19, 20, 23	20, 24
23	5	24	7, 8, 10, 34, 30, 19, 23	10, 30
24	8	1	2	1, 2
27	4	27	6, 7, 13, 21, 25	25, 27
29	5	24	6, 7, 14, 13, 19, 23	14, 19

The percentage share of different motorized modes in terms of total passengers carried in pre and post covid situation is shown in Fig. 2 . Among the motorized modes, 22% of the passengers use city buses, 29% use motorized two-wheeler (MTW), 31% use car and 17% use intermediate public transport (IPT) services. The mode share of the city buses stood at 9% after the COVID-19 relaxation and the mode share of private car and MTWs have considerable increased. It indicates a massive ridership decline of 59.09% for the city buses. For this reason, traffic congestion has become a part of the city post COVID situation. Therefore, it is essential to improve the service quality of city buses to attract users.

Fig. 2.

share of different motorized modes in terms of total passengers' carried in pre-covid and post-covid situation.

Guwahati is a non-metro city. Apart from the IPT modes, city buses are only the public transportation mode available in the city. Therefore, city buses play an important role in transporting a large number of people within the city. A total road length of 171.6 km is covered by city bus route network in the city. Currently a total number of 753 buses are operated in the municipal area (based on the data collected from regional transport office). The ridership of the city buses is continuously reducing and recent break of the COVID-19 has tremendously affected the ridership base of the city buses. Due to spread of the COVID-19, Indian Govt. announced a complete lockdown for 21 days starting from 25th March 2020 [11]. But this lockdown was extended till 7th June 2020 due to the further spreading of the virus. The lockdown measures were started to be relaxed in a phased manner from 8th June 2020 [22]. And the city bus operation in Guwahati started from 16th August 2020 with only 50% seating capacity and with mandating all the COVID-19 protocols (social distancing, mandatory use of face masks etc.). The bus service resumed after nearly 144 days of complete halt. During this time bus operators faced a huge financial loss. After repeated objections from different transport authorities, finally, Assam Govt. allowed to carry passenger up to full seating capacity plus some restricted quantity of standee passenger (one third of seating capacity) on 9th October 2020 [23]. But after this measure also, users are reluctant to use the buses. In maximum of the cases buses remains empty. Therefore, it felt necessary to investigate how users are perceiving the bus service after the COVID-19 situation so that necessary measures could be taken to make the bus service attractive for the users.

4. Methodology

4.1. Data collection

Data were collected from both the primary and secondary sources. The data requirement for the study is given in Table 3 .

Table 3.

Data requirements for the study.

Data Category	Parameters to be Collected	Data Type	Data Collection Technique
Revenue Losses of the buses	Average daily revenue collection of buses	Secondary source	regional transport offices, and several private transport operators
	Wages of drivers, ticket collectors and helpers	Primary Survey	By interviewing staffs of the buses
	Fuel price and maintenance cost	Secondary source	regional transport offices, and several private transport operators
Users' Perception data	–	Primary Survey	Questionnaire Survey

4.1.1. Data collection corresponding to revenue loss of the bus

Average daily revenue collection of the buses was collected by secondary sources (regional transport offices, and several private transport operators). Wages of the staffs, fuel prices and maintenance cost contributed to the expenditure of the buses. In the city buses in Guwahati apart from the driver, one ticket collector and one helper are assigned. The wages of these three categories of the workers varies. Generally, the drivers are provided with highest wage followed by the ticket collector. The wages of the helpers are found to be considerably low. The data related to wages of the different workers are collected using a primary survey by interviewing the staffs of the buses.

The wage related information during pre-covid situation was primarily collected from the staffs but that data is validated from the operators too. The revenue collection and expenditure related data from the operators were collected from their offices. These offices for various areas are situated at the starting point of the routes/common parking lots of the buses. The surveyor initially collected the data from several operators regarding personal information (name and contact number) revenue collection of the buses, maintenance cost, cost of fuel etc. In this survey, the wage related information was also asked. But the wages provided by the operators were approximate value of total wage of all the staffs and it was also stated that the wages vary depending on the skill and experience of the staffs. The surveyor further took permission from the operators to have some interactions with the staffs so that the exact information regarding the wages could be collected. The wage related information was collected from the staff in the specific transport offices only so that somewhat sincere response can be gathered from the staffs. For this primary survey 200 buses were selected randomly and the staffs of the selected buses were interviewed to collect the information regarding their wages. The surveyor first initiated a friendly talk with the staffs and state the purpose of data collection and after that, the information regarding the wages were asked. The data was collected using random sampling method. The survey data were collected on three different time periods, pre-covid (January 2019) situation, unlock with 50% seating capacity (November 2020) and unlock with 100% seat capacity plus some restricted amount of standee (January 2021) situation.

4.1.2. Users’ perception data

Users’ perception data was collected through a questionnaire survey. It is not possible to interview the whole population of the study area for the questionnaire survey. Therefore, it is necessary to calculate the sample size for the survey. The minimum size of the sample is calculated as per Eq. (1) [24]:

$$n\ge N{\left[1+\frac{N-1}{P\left(1-P\right)}{\left(\frac{d}{z_{\alpha /2}}\right)}^2\right]}^{-1}$$ (1)

where.

n is the minimum sample size to be considered.

N is the population of the city.

P is the quality characteristics which are to be measured. For neutral cases or where no previous experience exists then the value of P is taken as 0.5 [24].

d is the margin of error which is taken as 5%

$z_{\alpha /2}=1.96$ for 95% confidence interval

As per the 2011 census data, the population of Guwahati is 957,352. The minimum sample size is determined on the basis of Eq. (1) and found to be 385. Considering the minimum sample size, a total number of 650 and 470 valid responses were collected for the questionnaire survey before and after COVID-19 situation respectively.

The purpose of the questionnaire survey was to gather information about the users’ perception of the bus service. Therefore, the current users of the bus service and the users having a previous city bus experience were interviewed for the survey. The survey was conducted on various locations of the city. The locations include bus stops and other places like markets, parks, offices etc. which are accessible by bus service. The questionnaire survey was conducted using face to face interview method and by online method (through google forms).

The respondents for the fact-to-face survey were selected based on voluntary response sampling. For face-to-face interview, some users in each bus stops/on board were approached and stated the purpose of the survey. Then their willingness to participate in the survey were asked. Around 40% of the users stated their willingness to took part in the survey. In this way a total number of 750 survey sheets were distributed and collected after sometime. From this response sheet, only 300 responses (incomplete responses, uniform types of responses were not considered) were found to be valid. The response rate for the face-to-face survey was found to be 40%.

The online survey was conducted using google form. For the online survey, different offices/educational institutes along the bus routes were visited and the form was shared with some of the higher authorities. For online survey, different offices, schools/universities were selected based on random sampling method. The purpose of the survey was stated and along with that it is requested to share with their colleagues/friends. The respondents for the survey were selected based on snowball sampling. A total number of 57 offices/educational institutes were visited and survey form was shared. In this way a total number of 843 survey responses were collected among which only 350 responses were found to be valid for further analysis.

After the relaxation of the lockdown norms another survey was conducted and a total number of 470 valid responses (170 responses through face-to-face interview and 300 responses through online survey) were collected. The face-to-face interview was conducted on board as well as different bus stops. A total number of 550 survey sheets were distributed and collected. After verification it was found that only 170 responses were valid (with a response rate of 31%). For the online survey, the same respondents were contacted who took part in the online survey in the previous instance of survey before covid-19. A total number of 300 valid responses were collected with a response rate of 35%.

The attributes of the questionnaire survey were finalized by conducting a preliminary survey. The preliminary survey was conducted in pre-covid situation. The preliminary survey could be considered as a normal conversation between the two passengers rather than a traditional survey. The preliminary survey was conducted on board. Therefore, the survey respondents were the city bus users. A total number of 102 respondents were interviewed for the survey. Among the 70 respondents, 35 respondents were found to be school/college going students (less than 25 years of age), 54 respondents were the normal working-class people and 11 respondents were found to belong in a higher age group (may be more than or near about 60). Out of 102 respondents, a total number of 40 respondents were female passengers. Among these 40 female users, 15 were completing the trip for their school or universities, 25 respondents were the normal office goers. Each respondent was requested in the preliminary survey to state the attributes effecting their level of satisfaction. Based on their stated responses the attributes for the final survey were fixed. The questionnaire survey consists of two parts. In the first part the respondents were requested to answer various questions related to the socioeconomic and demographic variables. In the second part, the respondents were requested to rate various statements corresponding to different attribute of the bus service based on their perception. The respondents were requested to rate the statements in 9-point scale where 1 indicates that they are completely disagreeing with the statement and 9 indicates that they are completely agreeing with the statement. The questionnaire survey was conducted in two time periods, just before the spread of COVID-19 (December 2019) and after the complete relaxation of COVID-19 precautionary measures (September 2021).

4.2. Tools and techniques used

4.2.1. Paired sample t-test

We perform paired t-test to determine whether the mean value of user's perception attribute before COVID 19 is statistically difference from the mean value of user's perception attribute after COVID 19. The test was carried out under the null hypothesis that there is no difference in mean value of all the attribute before and after COVID 19. A 95% confidence interval was considered during the analysis. Paired t-test is a widely used statistical test; hence it is not discussed in detail. For detail, the reader may refer to Ref. [25].

4.2.2. Factor analysis

Factor analysis is used to analyze the user's perception data. Factor analysis is a two-step procedure, i.e., Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA). EFA is used to know the number of unobservable summary variables or factors which are needed to explain the correlations between variables [26] and the link between the latent factors and observed variables [26]. Therefore, the objective of the EFA is to reduce the number of variables (data reduction) and to identify the relationship between observed variables and the latent factors. EFA is used when the researcher does not have any idea about the number of underlying factors and their structural relationship with the observed. EFA is used to determine the numbers of latent factors which affect the observed variables. CFA was used to verify and confirm the structural relationship between the latent variables and the observed variables [13, 16] through the different goodness of fit statistics values [8, 13]. The researcher must have specific knowledge of the total number of factors and the relationship between observed variables and the latent factor prior to the CFA model [13, 16]. Therefore, before conducting CFA, EFA is to be conducted. Two types of outcomes are estimated from the CFA analysis: factor loadings and factor scores [7, 10]. EFA and CFA was conducted using STATA 14 software.

5. Results

5.1. Revenue collection and expenditure of the buses

Average revenue collection/day/bus was estimated from the revenue collection data collected from regional transport offices and from several private transport operators. It could be seen from Table 4 that the average wages of the bus staffs (driver, ticket collector and helper) reduced considerably after COVID-19. In Guwahati city, the wages of the bus staffs are not fixed. The wages depend on the total number of runs they made in each day. As the number of passengers have considerably reduced after the spreading of the COVID-19, therefore buses made very few runs between different stations. This leads to the reduction in their wages. Average revenue collection/day/bus have reduced tremendously post COVID-19 outbreak compared to pre COVID-19 period. Moreover, it is also found that there is a slight improvement in the revenue collection after the 2nd unlock phase compared to the 1st unlock. The improvement of the revenue collection in the second unlock phase indicate that the users of the buses increase slightly compared to the 1st unlock phase. But compared to pre covid situation, that improvement is very marginal. In this regard it is very much essential to understand the users concern about the city bus service so that necessary action could be taken to improve the user base.

Table 4.

Revenue collection and expenditure of the buses.

Items	Before COVID-19 Lockdown (Rs.)	1st Unlock with 50% Seat Capacity (Rs.)	2nd Unlock With 100% Seat Capacity Plus Restricted amount of Standee (Rs.)
Average revenue collection/day/bus	6122	3205	3805
Average wages of driver/day/bus	780	339	476
Average wages of ticket collector/day/bus	592	273	353
Average wages of helper/day/bus	391	215	245
Fuel price and maintenance cost/day/bus	2237	1807	2035
Average additional expenditure in lieu of Covid-19	0	50	50
Profit/bus/day	2122	521	646
Revenue loss/bus/day	–	1601	1476

5.2. Preliminary data analysis of questionnaire data

Reliability of the data set is checked after collecting the data using Cronbach's alpha value. Cronbach's alpha value indicates the internal consistency of the data as a group. A Cronbach's alpha value of more than 0.7 indicates a reasonable consistency of the data. For both the pre covid and post covid data set the Cronbach's alpha value were found to be more than 0.9 which indicate that the data sets are highly consistent to proceed with further analysis. The socioeconomic characteristics of the respondents are shown in Table 5 .

Table 5.

Socioeconomic characteristics of the respondents.

Socioeconomic characteristics of the respondents		Percentage of the total surveyed data (%) pre-covid	Percentage of the total surveyed data (%) post-covid
Gender	Male	47.8	54
	Female	52.2	46
Monthly income in Rupees	Less than 10,000 (Group 1)	37.2	36.2
	10,000–15,000 (Group 2)	34.2	31.3
	15,000–20,000 (Group 3)	12.8	16.6
	20,000–25,000 (Group 4)	6.1	9.4
	More than 25,000 (Group 5)	9.4	6.5
Age (years)	Less than 25 (Group 1)	20.6	21
	25–40 (Group 2)	55.5	54
	40–55 (Group 3)	22.2	23
	Over 55 (Group 4)	1.7	2

In Pre-covid situation, the samples included 311 male respondents and 339 female respondents. In post covid situation, a total number of 254 male respondents and 216 female respondents took part in the survey. In both the time period, maximum of the city bus users belongs to the income group 1 and 2. The qualitative attributes of the questionnaire survey are given in Table 6 .

Table 6.

Service quality attributes.

Qualitative attributes	Assigned code to the perception data set	Average values of user perception ratings before COVID-19	Average values of user perception ratings after COVID-19
Buses are well maintained	a1	4.5	3.8
Seats are generally clean	a2	4.5	3.5
Seats are comfortable and sufficient leg room is provided	a3	5.2	5.1
Doors and windows are in proper working condition	a4	4.6	4.4
Vehicles are odor free and clean	a5	4.2	3.0
While standing sufficient distance is maintained from the fellow passengers	a6	3.5	3.2
Seats are available while travelling in the buses	a7	3.5	5.6
Staffs are well behaved	a8	4.5	5.5
Some reserved seats are provided for women and older people	a9	4.2	5.7
Very few breakdowns of the vehicles are experienced during journey	a10	5.5	5.6
Adequate travel speed is maintained	a11	4.2	4.4
Buses are punctual	a12	4.6	3.5
When you are travelling with the bus you know exactly when you reached your destination	a13	4.0	3.5
You know how much you have to wait for the bus in a particular bus stop	a14	4.5	3.7
Bus service is Consistent irrespective of the situation	a15	5.3	4.5
Bus fares are justifiable	a16	6.0	6.2
Bus fares are fixed for all the buses	a17	6.5	6.7
Bus service could be easily accessed from your location	a18	4.5	4.7
You feel safe while travelling through bus with respect to theft and other mishaps	a19	4.5	4.7
Buses run safely maintaining all the traffic rules	a20	6.5	6.4
Bus stops are safe and tidy	a21	4.6	4.5
Adequate time is provided for boarding and alighting	a22	4.7	5.5
Overall satisfaction		5.2	4.3

The statement a1 is rated below after the COVID-19. It may be due to the fact that bus service was closed for nearly 144 days in Guwahati. Within that period no revenue was generated by the buses. Therefore, during that period very little or no maintenance works were carried out which disrupts the swift operation of the buses. Statement a2 and a5 are also rated considerably below after the COVID-19. Though after the COVID-19 lockdown some cleaning work was initiated by the buses after each run and that are continued till now. But covid 19 has inflict some sense of hygiene and cleanliness in the mind of the users. Therefore, the cleanliness and hygiene of the buses become an important parameter for the users. After the covid 19, though the buses may look more clean and hygiene compare to the pre covid situation but it was not able to meet up with the users level of expectation.

Statement a3, a4, a10, a11, a16, a17, a18, a19, a20, and a21 are rated similar before and after COVID-19 lockdown as there were no changes in those particular attributes of the buses before and after COVID-19. Statement a6 is also rated very closely before and after COVID-19. Before COVID-19 during the peak hours' buses became overly crowded and this fact led to low rating of this particular attribute. After the relaxation of COVID-19 norms though very little passengers are allowed after the seats are full but passengers felt uncomfortable in the buses due to the proximity of other passengers. For this reason, this attribute is rated very low after the COVID-19. Statement a7 is rated better after the COVID-19, because after the relaxation of COVID-19 norms very less people are allowed after the seats are full and many users opted for other mode of transport. These two factors lead to more availability of the seats. Moreover, as the user base of the buses reduces considerably, therefore the attitudes of the staffs of the buses changed considerably after the reopening of the bus service which leads to a higher rating of the statement a8 after the COVID-19. In Guwahati, as per the Govt. norms some of the seats were reserved for the older people and female passengers. But before the COVID-19, due to overly crowding of the buses in some of the routes, this norm was not strictly followed. After the COVID-19, as the buses remain empty in most of the time, the norm is now strictly followed. Moreover, some new buses are introduced by the Govt. only for the female users and older people. That's why the statement a9 is rated higher after the COVID-19. Statements a12, a13, a14, a15 is rated below than the pre covid situation. This is because, as the users base of the buses reduced considerably post covid, therefore, buses tend to wait in a particular stop for a considerable period of time just to catch some more passengers. This leads to increase the journey time and therefore will reduce the consistency of the buses. The punctuality of buses (a12) is defined on the basis of users' perception regarding arrival of buses in each stop. All the users may not have the information regarding the official bus schedule, but they have some perception regarding the arrival and departure of buses from each station. They perceive this punctuality from their experience of day-to-day travel and by comparing the travel time with the travel time required by para transit or other mode of transport.

Gender wise distribution of the users’ ratings on overall satisfaction is indicated in Table 7 . 26.8% of the male respondents in pre covid situation and 34.8% male respondents in post covid situation were found to provide a rating of 4 or below on overall satisfaction. For the female respondents, these values were found to be 52.2% in pre covid situation and 58.4% in post covid situation. 27.9% of male respondents in pre covid situation and 41.7% male respondents in post covid situation were found to be provided a neutral rating. On the other hand, in pre covid situation 26.6% of the female respondents and 27.7% female respondents in post covid situation provided a neutral rating value. 45.3% of the male respondents in pre covid situation and 23.5% of the male respondents in post covid situation were found to be satisfied with the bus service (provided a rating of 6 or more). But for the female respondents, these values were found to be 21.3% and 13.8% respectively for pre and post covid situation. Percentage of female respondents were found to be more before the neutral ratings while the percentage of male respondents were more after the neutral rating. This indicates that the female respondents are less satisfied with respect to the male respondents.

Table 7.

Gender wise distribution of the users’ ratings on overall satisfaction.

Time Period	Sex	Distribution of users' ratings on the overall satisfaction of the bus service (%)
		1	2	3	4	5	6	7	8	9
Pre covid	Male	1.2	0.0	3.5	22.1	27.9	27.9	5.8	11.6	0.0
	Female	4.3	3.2	13.8	30.9	26.6	13.8	3.2	4.3	0.0
Post covid	Male	2.6	10.4	6.1	15.7	41.7	15.7	5.2	2.0	0.6
	Female	3.1	1.5	20.0	33.8	27.7	7.7	4.6	1.0	0.5

Age wise distribution of the users’ ratings on overall satisfaction of the bus service is indicated in Table 8 . In pre covid situation, 39% respondents of the age group 1 provided a rating of 4 or below, 31.7% of respondents provided a neutral rating while 29.3% of the respondents provided a rating of 6 or more. 39.4% of the respondents provided a rating of 4 or below, 24.7% provided a neutral rating and 35.9% of the respondents provided a rating of 6 or more among the users belonging to age group 2. For the respondents belonging to age group 3, these percentage values were found to be 57.1%, 7.1% and 35.8%. For the respondents belonging to age group 4 these percentage values were found to be 30%, 20% and 50%. Therefore, it can be said that with increasing age the dissatisfaction of the users increases.

Table 8.

Age wise distribution of the users’ ratings on overall satisfaction.

Time period	Age Groups	Distribution of users' ratings on the overall satisfaction of the bus service (%)
		1	2	3	4	5	6	7	8	9
Pre covid	1	0.0	3.7	8.5	26.8	31.7	20.7	2.4	6.1	0.0
	2	2.9	2.0	8.6	25.9	24.7	21.0	6.2	8.6	0.0
	3	7.1	0.0	14.3	35.7	7.1	14.3	7.1	14.3	0.0
	4	0.0	0.0	0.0	30.0	20.0	40.0	10.0	0.0	0.0
Post covid	1	4.2	6.9	6.9	26.4	33.3	15.3	1.4	4.2	1.4
	2	2.7	9.6	8.2	20.5	39.7	13.7	5.5	0.0	0.0
	3	0.0	4.0	4.0	12.0	44.0	24.0	12.0	0.0	0.0
	4	10.3	8.5	20.5	30.3	21.7	8.7	0.0	0.0	0.0

For the respondents of age group 1 in post covid situation, 44.4% provided a rating of 4 or below, 33.3% provided a neutral rating and 22.3% provided a rating of 6 or more. For the respondents of age group 2 these values were 41%, 39.7% and 19.3% respectively. For the respondents belonging to age group 3, these values were found to be 20%, 44% and 36%. For the respondents belonging to age group 4, these values were 69.6%, 21.7% and 8.7%.

Income wise distribution of the users’ ratings on overall satisfaction of the bus service is indicated in Table 9 . The percentage of respondents provided ratings of 4 or less are 43.3%, 20.9%, 47.8%, 45.5% and 82.3% respectively for income group 1, 2, 3, 4 and 5 in pre covid situation. In post covid situation these are found to be 49%, 27%, 54%, 62.5% and 70.9% respectively for income group 1, 2, 3, 4 and 5. It indicates that users in post covid situation poorly perceived the bus service in comparison to pre covid situation.

Table 9.

Income wise distribution of the users’ ratings on overall satisfaction.

Time period	Income groups	Distribution of users' ratings on the overall satisfaction of the bus service (%)
		1	2	3	4	5	6	7	8	9
Pre covid	1	0.0	3.0	10.4	29.9	28.4	20.9	3.0	4.5	0.0
	2	0.0	1.6	3.2	16.1	24.2	32.3	9.7	12.9	0.0
	3	0.0	0.0	13.0	34.8	34.8	8.7	8.7	0	0.0
	4	10.5	7.7	9.1	18.2	45.5	9.1	0.0	0.0	0.0
	5	12.6	9.1	13.5	47.1	11.8	5.9	0.0	0	0.0
Post covid	1	3.4	6.9	7.8	30.9	28.8	10.3	3.4	6.0	2.5
	2	0.0	8.1	5.4	13.5	40.5	10.8	8.7	9.5	3.5
	3	2.6	10.3	2.6	38.5	25.6	20.5	0.0	0.0	0.0
	4	3.7	14.8	3.7	40.3	22.2	7.4	6.5	1.5	0.0
	5	4.2	8.3	14.6	43.8	20.8	8.3	0.0	0.0	0.0

Correlations among the different variables are checked using Variance Inflation Factor (VIF). VIF represents the amount of variance of a regressor explained by the remaining regressors present in the regression model due to correlation among them [27]. Different cut-off values for the VIF have been suggested by different researcher depending on the nature of the study. In this study, the cut-off value of the VIF is considered as 2 [28] i.e. the variable with a VIF value greater than 2 is highly correlated with the remaining variables. For pre covid situation it is found that all the variables included in the questionnaire have a VIF value more than 2. Therefore, it could be said that all the variables are highly correlated. For post covid data, it is found that except bus fare (a16) all the variables are highly correlated. It indicates that most of the variables are highly correlated and unsuitable for regression. For this reason, factor analysis is performed to represent the highly correlated variables with a smaller number of variables.

5.3. Paired sample t-test

The mean and standard deviation value of each attribute are shown in Table 10 . From Table 10, it can be seen that the t value and p value of every attribute is very small (p < 0.05), hence the null hypothesis was rejected. Thus, the difference in mean value of user's perception before COVID 19 is statistically significance from the mean value of user's perception attribute after COVID 19.

Table 10.

Paired sample t-test statistics.

Attributes	Average values of user perception ratings before COVID-19	Standard Deviation	Average values of user perception ratings after COVID-19	Standard Deviation	t statistics	degree of freedom	std error of difference	p value
a1	4.5	0.52	3.8	0.91	16.27	1118	0.043	0.0001
a2	4.5	1.01	3.5	1.02	16.28	1118	0.061	0.0001
a3	5.2	0.77	5.1	0.9	1.997	1118	0.05	0.0461
a4	4.6	0.79	4.4	0.92	3.9	1118	0.051	0.0001
a5	4.2	0.81	3	1.1	21.0261	1118	0.057	0.0001
a6	3.5	0.73	3.2	0.62	7.2225	1118	0.042	0.0001
a7	3.5	0.62	5.6	0.54	59.00079	1118	0.036	0.0001
a8	4.5	0.86	5.5	0.84	19.39	1118	0.052	0.0001
a9	4.2	0.53	5.7	0.83	36.846	1118	0.041	0.0001
a10	5.5	0.76	5.6	0.46	2.536	1118	0.039	0.013
a11	4.2	0.62	4.4	0.84	4.584	1118	0.044	0.0001
a12	4.6	0.73	3.5	0.26	31.262	1118	0.035	0.0001
a13	4	0.78	3.5	0.73	10.874	1118	0.046	0.0001
a14	4.5	0.56	3.7	0.85	18.97	1118	0.042	0.0001
a15	5.3	0.6	4.5	0.57	22.486	1118	0.036	0.0001
a16	6	0.73	6.2	0.82	4.295	1118	0.047	0.0001
a17	6.5	0.62	6.7	1.09	3.889	1118	0.51	0.0001
a18	4.5	0.87	4.7	0.53	4.425	1118	0.045	0.0001
a19	4.5	0.69	4.7	0.77	4.558	1118	0.044	0.0001
a20	6.5	0.57	6.4	0.84	2.372	1118	0.042	0.0178
a21	4.6	0.63	4.5	0.72	2.468	1118	0.041	0.0137
a22	4.7	0.57	5.5	0.76	20.128	1118	0.04	0.0001

5.4. Factor analysis

For Pre covid situation, all the variable are analyzed using factor analysis and for post covid situation all the variable except a16 are used for factor analysis. Factor analysis is performed to find the factors which affect users’ perception. For pre covid situation, four latent factors with eigenvalues greater than one [29] were extracted from the 23 correlated attributes which explained 75.7% of the total variance. For post covid situation, five latent factors with eigenvalues greater than one [29] were extracted from the 22 correlated attributes on the city bus service which explained 77.5% of the total variance. Before conducting EFA sample adequacy have been checked by Kaiser-Meyer-Olkin (KMO) tests. The KMO estimates the ratio of the squared correlation between variables to the squared partial correlation between variables [30]. When the KMO is near 0, it is difficult to extract a factor and when the KMO is near 1, a factor or factors can probably be extracted. For an acceptable sample adequacy, the KMO value should be more than 0.5 [31]. The tests statistics values are found to be more than 0.9 for both the cases. For both the cases, EFA was conducted separately. EFA also indicated the possible relationship between the factors and observed variables. Based on the results obtained from the EFA, CFA was conducted. The EFA aims at reducing (data reduction) the number of variables and defining the relations between observed and latent factors. EFA is conducted using IBM SPSS 20 software. In SPSS, different extraction methods for the EFA analysis are there-principal components, principal axis factoring, unweighted least squares, maximum likelihood etc. Principal components is the default extraction method in SPSS and it extracts uncorrelated linear combinations of the variables. As the data set was highly correlated, therefore principal components was used as extraction method for the EFA to extract uncorrelated factor. After extracting the factors, factor rotation is necessary to better fit the data. The most commonly used method is varimax. Varimax is an orthogonal rotation method that tends to produce factor loading that are either very high or very low, making it easier to match each item with a single factor. For this reason, varimax rotation technique was used.

The factor structure for the CFA analysis is shown in Fig. 3 for pre covid situation and in Fig. 4 for post covid situation. In both the figures, the square boxes represent different observable variables, the oval shapes represent latent factors. The values written beside the single arrowed lines represent standardized factor loadings. Square of these factor loading represent correlation coefficient between the latent factor and observed variables. The values shown beside the small round shapes are standardized error variance which represent unexplained portion of the variance. Four factors namely comfort, fairness, reliability and safety have been extracted from the attributes in Pre-covid situation. Five factors extracted from the attributes in post covid situation are vehicle condition and hygiene, comfort, Reliability & convenience, fairness, and safety.

Fig. 3.

CFA analysis of the perception data for pre COVID-19 situation.

Fig. 4.

CFA analysis of the perception data for post COVID-19 situation.

Four factors namely comfort, fairness, reliability and safety have been extracted from the attributes in Pre-covid situation. Ten attributes (a1 to a10) are grouped under factor comfort as users relate these attributes with a comfortable journey. For example, attribute a1 relates to the maintenance of the vehicle. If the vehicle is well maintained, it would be visually appealing as well as comfortable to ride. Attribute a2 and a4 are related to the cleanliness of the vehicle and seats. But these two attributes are also grouped under factor comfort as users perceive that cleanliness and odor free journey will make their ride more comfortable. Good behavior of the staffs inflicts confident among the passengers and hence make their ride more comfortable. For this reason, this attribute is also grouped under factor comfort. Doors and windows in proper working condition is very necessary during the adverse weather conditions, winter seasons etc. Therefore, condition of the windows and doors (a4) are necessary for a comfortable ride of the users. Attributes a11, a12, a15, a16 and a18 are grouped under factor reliability & convenience. Adequate travel speed (a11) and punctuality of the buses (a12) are very much necessary for the convenient of the users. Moreover, maintaining punctuality makes the buses more reliable. Therefore, these two attributes grouped under factor reliability and convenience. Similarly, Consistency of the service (a15), accessibility of the service (a18) also makes the ride more reliable. One thing could be noticed that, the attribute a16 (bus fares are justifiable) also fall under the factor reliability and convenience. It may be due to the fact that a reasonable fare will make the bus service more affordable for the users and hence it will make the service more convenient for the users. Attributes a13, a14 and a17 fall under factor fairness. Attribute a17 (bus fares are fixed for all the buses) indicates the consistency of the bus fare. A good performance of this attribute indicate that the bus fares are fixed for different routes irrespective of the buses and any time period. Therefore, this factor is grouped under fairness. Attribute a13 (When you are travelling with the bus you know exactly when you reached your destination) indicates that the passengers are well informed about their journey time i.e., they know exactly when they will reach their destination if the ride the bus. Therefore, this attribute is also grouped under factor fairness. Similarly, attribute a14 (You know how much you have to wait for the bus in a particular bus stop) is also grouped under factor fairness. Attribute a19, a20, a21, and a21 are grouped under factor safety as they relate to the safety of the passengers and the bus ride.

The factor structure for the post Covid situation is found to be different from the pre covid situation as the users may perceive the attributes differently. Five factors extracted from the attributes in post covid situation are vehicle condition and hygiene, comfort, Reliability & convenience, fairness, and safety. Attributes a1, a2, a4, and a5 grouped under factor vehicle condition and hygiene. Attribute a2 and a5 relate to the cleanliness of the vehicle and seats while attributes a1 and a4 relate to the condition of the vehicle. In pre covid situation, all these attributes were grouped under factor comfort but for the post covid situation users view these attributes separately than from the pre covid era as the importance of cleanliness and hygiene have greatly increased in post covid situation. Attributes a3, a6, a7, a8, and a9 grouped under factor comfort as all these attributes relate to the comfort of the users. The attributes under factor safety, and fairness remain same in both the pre and post covid situation. It is to be noted that, attribute a16 (Bus fares are justifiable) is not included in the factor structure of post covid situation as this attribute is seen as a separate entity in post covid situation.

Comparative fit index (CFI), chi-square to degrees of freedom ratio (χ²/d.f.), standardized root mean squared residual (SRMR), root mean squared error of approximation (RMSEA), and coefficient of determination (CD) values were used to check the goodness of fit for the proposed CFA model. The CFI, χ2/d.f., SRMR, RMSEA, and CD values for the model representing pre-covid situation are found to be 0.911, 2.23, 0.069, 0.067, and 0.998 and for the post covid situation these values are found to be 0.921, 2.66, 0.058, 0.071 and 0.987. All these mentioned values are found to be within their limit [32,33]. Therefore, it can be said that the proposed CFA models fits the data well. The internal consistency of the variables within the factors (i.e. reliability of the proposed factor structures) were checked using factor loading values, composite reliability (CR) values, and average variance extracted (AVE) values. Factor loadings, composite reliability, and average variance extracted values should be more than 0.5, 0.7, and 0.5 respectively for a reasonable internal consistency of the variables [31]. From Fig. 3, Fig. 4 it is found that all the factor loading values are more than 0.5. The composite reliability and average variance extracted values are shown in Table 7. It is found that composite reliability and average variance extracted values are within their acceptable limits. Therefore, it can be said that the reliabilities of the proposed CFA models are acceptable.

Discriminant validity of the factor is checked by comparing maximum shared variance (MSV) and average shared variance (ASV) with the AVE values. The MSV and AVE values are reported in Table 11 . It is found that MSV and ASV values are lesser than AVE values which indicates that there does not exist any problems regarding discriminant validity of the factors.

Table 11.

Values of the composite reliability and average variance extracted of the latent factors.

CFA model for Pre covid data set					CFA model for post covid data set
Latent factors	CR	AVE	MSV	ASV	Latent factors	CR	AVE	MSV	ASV
Safety	0.86	0.60	0.21	0.18	Safety	0.89	0.67	0.41	0.31
Comfort	0.92	0.54	0.26	0.23	Comfort	0.85	0.53	0.41	0.35
Reliability & Convenience	0.81	0.52	0.26	0.18	Reliability & Convenience	0.87	0.58	0.38	0.32
Fairness	0.77	0.53	0.23	0.16	Fairness	0.79	0.55	0.28	0.22
					Vehicle condition & Hygiene	0.94	0.80	0.39	0.32

5.5. Relative importance of the factors

The relative importance of the factors on overall satisfaction is estimated using path analysis. Path analysis and hierarchical regression (HR) are the two most commonly used methods to analyze if the variable of interest explained a statistically significant amount of variance in a model. HR is commonly used when a model's variance is explained by predictor variables that correlate with each other. As there are negligible intercorrelation among the factors, path analysis rather than hierarchical regression analysis was used in this study. Moreover, in the mentioned section, path analysis is conducted to estimate the relative weight of all the factors which is indicative of their relative importance over the overall satisfaction of the users. HR in this case may not serve the purpose of this study. Therefore, it was not used in this section. In the path analysis, factors scores have been used as independent variables and the overall satisfaction as dependent variable. The path analysis for pre covid (a) and post covid (b) is shown in Fig. 5 respectively. The values shown beside the single arrowed lines are the coefficient values indicating relative importance of the factors on overall satisfaction. The values shown beside the small round shapes are the standardized error variance which represent unexplained portion of the variance. From Fig. 5(a) and (b) it could be said that the unexplained portion of the variance in pre-covid situation is 0.11 and in post covid situation is 0.15. Therefore, 89% of variance in the pre covid situation and 85% variance in post covid situation can be explained by the factors.

Fig. 5.

Relative importance of the factors.

For pre covid situation Reliability & Convenience is found to be most important factor followed by safety, comfort, and Fairness. But for the post covid situation, the most important factor is found to be Vehicle Condition & Hygiene. It indicates that the users after the covid 19 breakdown are more concerned with the vehicle condition and hygiene than the other factors. In the path analysis of the post covid analysis the bus fares (a16) is also included. The input variable for the bus fares are the data obtained from the questionnaire survey. In pre covid situation the bus fare (a16) was found to have strong correlation with other variables. For this reason, bus fare was included in the factor analysis of pre covid data. Factor analysis of the correlated data indicates that the bus fare fall under the factor reliability and convenience. For post covid data, it is found that except bus fare (a16) all the variables are highly correlated. It indicates that most of the variables are highly correlated and unsuitable for regression. For this reason, factor analysis is performed on all the 21 variables except a16. In post covid situation users perceived a16 as a separate entity. This may be due to the fact that, during the covid situation many people have lost their jobs and many people were forced to do the same job with lesser payment. And after that time period also, due to the increase in petrol prices in India there were frequent changes in the bus fare. For this reason bus fare is included in the path analysis in post covid situation. The relative importance of the bus fares is found to be low which indicate that the users are willing to pay more for a better service which can also be observed by the responses of users provided in Table 4. As vehicle condition and hygiene is found to be very important for the users in post covid situation, therefore, this factor should be improved to increase the users base of the city buses.

6. Discussion

In India, a major portion of the users are dependent on public transit. Therefore, a modal shift of the users from public transit to private vehicles will create many problems like pollution, congestion etc. In this regard some studies [4,17,18] have highlighted the impact of COVID-19 on the ridership of the public transport in India. Research findings showed that the public is in favor of personal vehicle compared to public vehicle during the early stages of COVID-19 [4]. Poor vehicle condition & hygiene could be a reason of this user's choice as evident from this study. From the previous literatures [18,19] it is also found that the fear of pandemic has greatly impacted people travel decision-decision and it will affect the ridership of the public transport even in the absence of restrictive measures. Dandapat et al. [4] found that an expected delayed recovery of the ridership of public transport post-COVID era. This is also evident from the study area where a massive decline of 59.09% of the ridership observes in the absence of restrictive measures. Similarly, Meena [17] reported a drop of shared mobility by 35% compared to the normal condition due to the fear of contact with unknown people. But till now no comparative studies of users' perceptions in pre and post covid situations are available as far as our best knowledge which limit the complete policy making process for the public transport users. A comparative study is necessary to highlight the changing behavior and priorities of users after the covid situation so that necessary action could be taken. Therefore, this study will open a new dimension in the area of public transport policy making and will be helpful for taking necessary action to recover the declining public transport sector in developing economic.

7. Conclusion and recommendations

COVID-19 has greatly affected the travel demand of the people which marked a huge dent on the public transit sector. Public transit authorities faced a huge financial loss which is also evident from this study. A cost analysis of the city buses indicate that the city bus operators faced a financial loss due to the COVID19 outbreak and it affects greatly to the livelihood of many workers associated with the city bus service. It has been found that the average profit of the buses reduced from 2122 Rs/day/buses to 521 Rs./day/buses in first unlock phase and 646 Rs./day/bus in second unlock phase. The average wage of the drivers reduced from 780 Rs./day to 339 Rs./day in first unlock phase and 476 Rs./day in second unlock phase. This loss in wage is not only limited to the drivers but extended to other workers associated with the buses. Loss in the average profit margin of the buses could be attributed to local travel restrictions imposed by the authorities and negative perception of the city bus users toward the bus service. The ridership of city buses has declined to 9% (in post covid situation) from 22% (pre covid situation). Therefore, in this study, the users’ perceptions in both pre covid and post covid situation were analyzed. It has been found that the users perception varies significantly in both the situation. Ridership decline is found to be a major issue in the post covid situation. Before the COVID-19 outbreak, users perceive vehicle condition and hygiene as an integral part of their comfort but in the wake of corona outbreak users tend to give more importance to the vehicle condition and hygiene than the other factors. Bus service in the city seem to have performed very poorly in this respect which may lead to shifting of users towards other mode of transport. Safety and reliability & convenience is also found to be important for the users. Under performing factors can be improved by improving the attributes associated with them. Some of the recommendations to improve the bus service in the city are as follows:

• •The most important factor in post covid situation is found to be vehicle condition and hygiene. This particular factor can be improved by making the buses clean and visually attractive for the passengers. Cleaning should be carried out after each run. Some incentive should be provided to the staffs of the buses for cleaning and maintaining the buses. Grievance box should be there in every buses so that users can put up their grievances. Staffs as well as passengers should be informed regarding various public transport healthy practices through different awareness program.
• •Safety of the buses is also found to be important. Safety measures need to be increased in the bus stops, and within the buses. The drivers of the city buses should be instructed to avoid rash driving. Moreover, utmost care and vigilance should be there during the boarding and alighting process. Driver needs to start the vehicle only after successful boarding and alighting process.
• •The bus service needs to follow a fixed arrival and departure time. This can be achieved by minimizing the unnecessary delays between the bus stops. The servicing of the buses should be done within the stipulated time to avoid any unnecessary breakdowns during the journey. The bus service should be consistent and easily accessible. To make the bus service more accessible some new routes can be introduced. Otherwise, the existing bus routes can be made more accessible by regulating the para transit services in the city. For example, in Guwahati city E-Rickshaw services are unregulated. This service can be made regulated so that they can act as a feeder to the public transport.
• •Number of buses should be increased to prevent the overcrowding during busy hours. This can be achieved by introducing additional buses during peak hours. The staffs may be provided adequate training in respect of their behavioral aspect towards passengers. Seating arrangements may be redesigned to make it more comfortable. From the preliminary data analysis, it is also found that the female users and people with relatively higher ages are not so satisfied with the service. Therefore, some additional buses could be introduced in the city keeping in mind about the female users and people with higher ages.
• •It is recommended to publish the bus fare and schedule of buses on different routes through posters, banners etc. in the buses and bus stops. Another way of sharing the information is total digitalization of the buses through development and implementation tools such as automated fare-collection systems, secure wayfinding and protocol app within terminal etc.

8. Future scope

This study is based on users' perception data. User perception of the service is not same for all the users. It differs between individuals and different market segments based on socioeconomic variables. Therefore, it would be advantageous to study in detail about the effect of demographic and socioeconomic characteristics of the passengers on their perception of the service to make the service more appealing across different groups. These effect of the users' socio-economic characteristics on perception was not discussed in detail in the article. Users' perception data can be collected in different ways (for example: face-to-face interview method, online data collection, data collection through mobile app etc.) and there may be variation of the data collected by different methods which will need further analysis and discussion. Moreover, the condition of the city buses will not be same throughout an entire day. For example, in peak hours the buses may be very crowdy but during the odd hours the buses may remain mostly empty. This type of condition will lead to temporal variation in users’ satisfaction which is also not discussed in this article. In this regard, it is felt that certain aspects of the city bus service deserve further studies. Some of them are mentioned below:

• •This temporal variability in the user's satisfaction can be studied in future to make the transit service more effective.
• •Effect of users' socio economic and demographic characteristics on users' satisfaction may be studied in future.
• •Effect of different data collection technique on the quality of collected data may be studied in future for better understanding about the different data collection technique.

Funding

No external funding is received in this study.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.