Socio-economic characteristics
The socio-economic characteristics of the sample are presented in Table 1. Remarkably, the majority of ride-hailing users are women (64.86%), young people who are 18–25 years old (69.15%), high school degree holders (52.32%), and singles (74.01%). The most common income is 1–3 M IDR (33.67%) and the most common occupation is student (46.10%). It is important to note that the income data collected represents individual incomes. In the case of students, many of them may have lower reported incomes due to their dependency on their families for financial support. It has been widely recognized already in other geographical contexts that young, well-educated, and high-income groups are more likely to be ride-hailing users (e.g., Alemi, Circella et al. (2018); Tirachini (2019); Tirachini and del Río (2019); Brail (2020); Kong, Moody et al. (2020); (Irawan, Bastarianto et al. 2021)). For vehicle ownership, most of the respondents (97.51%) own one or more motorcycles, in line with the major role of motorcycles in Indonesian transportation (Yogyakarta 2021). Most respondents live in the Sleman Region (55.71%), which also has the largest population of the three study areas (Statistics 2021).
Table 1
Socio-economic characteristics
Descriptions
|
n
|
%
|
|
Descriptions
|
n
|
%
|
Socio-Demography
|
|
|
|
Household Characteristics
|
|
|
Gender
|
|
|
|
Home Status
|
|
|
|
a) Man
|
285
|
35.4
|
|
|
a) Own by myself /family
|
720
|
89.44
|
|
b) Woman
|
520
|
64.6
|
|
|
b) Rent
|
85
|
10.56
|
Education
|
|
|
|
Type of House
|
|
|
|
a) Elementary or under
|
3
|
0.37
|
|
|
a) Landed house
|
659
|
81.86
|
|
b) Senior school certificate
|
2
|
0.25
|
|
|
b) Apartment / Boarding / others
|
146
|
18.14
|
|
c) Highschool certificate
|
418
|
51.93
|
|
Total Family Member
|
|
|
|
d) Bachelor equivalent
|
362
|
44.97
|
|
|
a) < 2 members
|
119
|
14.78
|
|
e) Master’s or higher
|
19
|
2.36
|
|
|
b) 2–4 members
|
385
|
47.83
|
|
f) Doctor’s degree
|
1
|
0.12
|
|
|
c) 4–6 members
|
249
|
30.93
|
Age
|
|
|
|
|
d) ≥ 6 members
|
52
|
6.46
|
|
a) 18–25
|
557
|
69.19
|
|
Driving License
|
|
|
|
b) 25–35
|
197
|
24.47
|
|
|
a) Have
|
147
|
18.26
|
|
c) More than 35
|
51
|
6.34
|
|
|
b) No Have
|
658
|
81.74
|
Marital Status
|
|
|
|
Motorcycle License
|
|
|
|
a) Single or not married
|
596
|
74.04
|
|
|
a) Have
|
618
|
76.77
|
|
b) Married
|
209
|
25.96
|
|
|
b) No Have
|
187
|
23.23
|
Income
|
|
|
|
Car in the Household
|
|
|
|
a) No monthly income
|
249
|
30.93
|
|
|
a) No have
|
544
|
67.58
|
|
b) Less than 1 M IDR
|
197
|
24.47
|
|
|
b) 1 car
|
224
|
27.83
|
|
c) 1–3 M IDR
|
277
|
34.41
|
|
|
c) 2 cars
|
27
|
3.35
|
|
d) > 3 M IDR
|
82
|
10.19
|
|
|
d) > 2 cars
|
10
|
1.24
|
Employment Status
|
|
|
|
Motorcycle in the Household
|
|
|
|
a) Student
|
368
|
45.71
|
|
|
a) No have
|
25
|
3.11
|
|
b) Full time workers
|
156
|
19.38
|
|
|
b) 1 motorcycle
|
357
|
44.35
|
|
c) Full time self-employee
|
106
|
13.17
|
|
|
c) 2 motorcycles
|
197
|
24.47
|
|
d) Par timer
|
60
|
7.45
|
|
|
d) > 2 motorcycles
|
226
|
28.07
|
|
e) No job
|
115
|
14.29
|
|
Bicycles in the Household
|
|
|
Living Area
|
|
|
|
|
a) No have
|
25
|
3.11
|
|
a) Bantul region
|
204
|
25.34
|
|
|
b) 1 bicycle
|
357
|
44.35
|
|
b) Sleman region
|
451
|
56.02
|
|
|
c) 2 bicycles
|
197
|
24.47
|
|
c) Yogyakarta city
|
150
|
18.63
|
|
|
d) > 2 bicycles
|
226
|
28.07
|
Exploratory factor analysis of food delivery attitudes
Table 2 presents a summary of the survey questions related to the reasons for using ride-hailing services. Respondents rated these reasons on a 5-point Likert scale, ranging from "very disagree" (1) to "very agree" (5), to indicate the extent to which they believe each reason applies. The table also includes the mean and standard deviation values for these ratings.
Table 2
Reasons for ordering food delivery via ride-hailing app
No.
|
I ordered food delivery because..
|
Mean
|
Std. Dev
|
1
|
The delivery cost is cheap
|
3.68
|
0.885
|
2
|
I get a discount when using food delivery
|
4.28
|
0.71
|
3
|
The delivery time is still reasonable
|
3.97
|
0.585
|
4
|
It’s difficult to reach the real restaurant
|
3.77
|
0.861
|
5
|
I can avoid a trip to the restaurant
|
4
|
0.791
|
6
|
It’s simple and easy to use it
|
4.23
|
0.639
|
7
|
The easy of payment
|
4.18
|
0.671
|
8
|
I got the best service even in the bad weather
|
4.03
|
0.774
|
9
|
They have a lot of choices of restaurant
|
4.17
|
0.703
|
10
|
I can easily compare the price and type of food
|
4.21
|
0.682
|
11
|
I use food delivery to avoid crowd during COVID-19
|
4.24
|
0.726
|
12
|
During the COVID-19, food delivery is the best choice to replace my eating-out.
|
4.09
|
0.828
|
After trying various rotations, we eliminated the variables number 4 with the lowest factor loading and conducted another rotation. Eventually, we identified the two most significant factors that represent the same constructs, namely "Convenience" and "Safety and Health Concern". Subsequently, we performed a reliability test, which revealed that the Cronbach's alpha values for each factor were 0.849 and 0.782, respectively. These values suggest that both factors are reliable and can be used for further analysis. The factor loading scores obtained from EFA are reported in Table 3.
Table 3
Exploratory factor analysis of attitudes toward food delivery services
No.
|
Indicators
|
Factor
|
Convenience
|
Safety and Health Concern
|
1
|
The delivery cost is cheap
|
0.444
|
|
2
|
I get a discount when using food delivery
|
0.613
|
|
3
|
The delivery time is still reasonable
|
0.525
|
|
5
|
I can avoid a trip to the restaurant
|
|
-0.483
|
6
|
It’s simple and easy to use it
|
0.743
|
|
7
|
The easy of payment
|
0.781
|
|
8
|
I got the best service even in the bad weather
|
0.546
|
|
9
|
They have a lot of choices of restaurant
|
0.725
|
|
10
|
I can easily compare the price and type of food
|
0.707
|
|
11
|
I use food delivery to avoid crowd during COVID-19
|
|
-0.951
|
12
|
During the COVID-19, food delivery is the best choice to replace my eating-out.
|
|
-0.569
|
Behavioral changes in food delivery and eating out.
The responses about behavioral changes for food delivery and eating out are presented in Table 4. The results indicate that the frequency of food delivery decreased somewhat during the pandemic. The share of people ordering food more than once a month has decreased, whereas the share of people ordering food only once a month or never has increased. However, the decrease is gradual. A much bigger change is seen for eating out. Whereas 25% used to visit a restaurant at least several times a month (and 25% even multiple times per week), the majority (55%) does not visit restaurants during COVID.
Table 4
The frequencies of food delivery and eating out before and during COVID
No.
|
Frequency
|
Food Delivery
|
Eating Out
|
Before COVID
|
During COVID
|
Before COVID
|
During COVID
|
|
n
|
%
|
n
|
%
|
n
|
%
|
n
|
%
|
|
a.
|
Never
|
82
|
9
|
137
|
15
|
79
|
9
|
491
|
55
|
|
b.
|
Once a month
|
183
|
21
|
250
|
28
|
89
|
10
|
172
|
19
|
|
c.
|
2–3 times a month
|
289
|
33
|
233
|
26
|
222
|
25
|
79
|
9
|
|
d.
|
Once a week
|
125
|
14
|
80
|
9
|
169
|
19
|
80
|
9
|
|
e.
|
2–3 times a week
|
187
|
21
|
149
|
17
|
217
|
25
|
54
|
6
|
|
f.
|
Daily
|
19
|
2
|
36
|
4
|
109
|
12
|
9
|
1
|
|
The cross-tabulations in Fig. 1 and Fig. 2 confirm that food delivery services before and during COVID do not change much, indicated by the high percentage of observations on the diagonal line-box from “never-never” to “daily-daily” with the total percentage not changing being 52. In contrast, most respondents switch to never visiting restaurants, as indicated by the high frequencies on the bottom line with a total of 55%. However, some respondents who visited restaurants frequently before COVID retain a relatively high frequency.
Based on survey responses, the total monthly frequency of restaurant visits and home deliveries has been calculated. To this end, we convert every frequency category to a number per month, as indicated in Table 5. Based on this total aggregation, eating out significantly drops by 78% and food delivery only slightly decreases by 3% during the pandemic.
We also display the comparison of both food delivery and eating out based on age, gender and employment status (see Fig. 3 and Fig. 4). The figures suggest that across the board, most people reduce their frequency of food delivery, but the share of daily users also increases slightly. Older people seem to reduce their frequency of food delivery more than younger people. With respect to eating out, we see an even stronger difference between ages, with older people reducing their frequency of eating out much more. We also see that men and women show a similar trend in the reduction of food delivery, although women remain more frequent users. For eating out, we see that women reduce their frequency of this activity more than men. With respect to daily occupation, workers (e.g., employees, part-timers) and the unemployed more strongly decrease food delivery during COVID. However, students were more likely to increase “daily” frequency of food delivery during COVID, although some decreased use. Compared to students, workers and unemployed respondents decreased the frequency of eating out during COVID.
Table 5
Number of food deliveries and eating out activities
No
|
Frequency
|
Conversion (freq/month)
|
Food Delivery (orders/month)
|
Eating Out (visits/month)
|
Before COVID
|
During COVID
|
Before COVID
|
During COVID
|
1
|
Never
|
0
|
0
|
0
|
0
|
0
|
2
|
Once a month
|
1
|
183
|
250
|
89
|
172
|
3
|
2–3 times a month
|
2,5
|
722.5
|
582.5
|
555
|
197.5
|
4
|
Once a week
|
4
|
500
|
320
|
676
|
320
|
5
|
2–3 times a week
|
10
|
1870
|
1490
|
2170
|
540
|
6
|
Daily
|
30
|
570
|
1080
|
3270
|
270
|
|
TOTAL
|
3,845.5
|
3,722.5
|
6,760
|
1,499.5
|
|
PERCENTAGE
|
57%
|
55%
|
100%
|
22%
|
|
AVERAGE NUMBER OF
VISITS (person per month)
|
4.35
|
4.21
|
7.64
|
1.69
|
Model estimation results
The results of model estimation of ordered logistic random effect are presented in Table 6, only displaying significant values. We also estimated the value of Variance Inflation Factors (VIF) to assess the multicollinearity among independent variables that present in the Appendix 2. Based on that VIF statistics, we did not find the multicollinearity issues, since the values are less than 5. Furthermore, estimation results suggest that ride-hailing food delivery and eating out have a complementary relationship. A higher frequency of eating out is associated with a higher frequency of food delivery and vice versa. We also find that people who have a higher frequency of work, education/school, grocery shopping, and leisure activities also more frequently eat out. Apparently, these activities are often combined with eating out. Food delivery is more frequent for people who have middle frequency for non-grocery shopping (such as shopping for long-term or special needs).
Based on the socio-economic variables, men order food delivery services less frequently. In addition, lower-income households have a lower frequency of ordering food delivery.
Interestingly, people who have at least one car or a motorcycle at home or have a drivers’ license have a higher frequency of both food delivery and eating out. Car ownership, motorcycle ownership and having a drivers’ license are possible indicators of a more out-of-home lifestyle where both eating out and ordering food substitute for preparing food at home. Finally, with respect to the duration of internet use, we find that shorter internet use is associated with a lower frequency of food delivery and eating out. Longer internet usage is possibly related to the use of online services, such as food delivery, but also with a need to go out for diner.
We also found that if people choose to use food delivery for convenience reasons, their frequency of using the service is higher. If they choose it for safety and health reasons, their frequency tends to be lower.
It is noted that while being a student does not have a significant effect, it does have a strongly positive effect on the frequency of eating out in a model where activity frequencies are left out (see Appendix 1). This suggests that the out-of-home activities that go together with eating out are typical for students.
Table 6
Model estimation of ordered logistic random effect panel model
No
|
Variables
|
Food Delivery
|
Eating Out
|
Food Delivery and Out-of-Home activity:
|
Coef.
|
Robust Std. Error
|
Coef.
|
Robust Std. Error
|
1
|
Food Delivery
|
|
|
|
|
|
a) > 4 times a month or daily
|
|
|
1.02***
|
0.25
|
|
b) 2–4 times a month
|
|
|
0.64**
|
0.193
|
|
c) < once a month or never
|
|
|
(ref. category)
|
-
|
2
|
Eating Out
|
|
|
|
|
|
a) > 4 times a month or daily
|
0.95**
|
0.295
|
|
|
|
b) 2–4 times a month
|
0.84***
|
0.223
|
|
|
|
c) < once a month or never
|
(ref. category)
|
-
|
|
|
3
|
Education
|
|
|
|
|
|
a) > 4 times a month or daily
|
-0.35
|
0.253
|
0.92***
|
0.213
|
|
b) 2–4 times a month
|
-0.05
|
0.281
|
0.05
|
0.284
|
|
c) < once a month or never
|
(ref. category)
|
-
|
(ref. category)
|
-
|
4
|
Work
|
|
|
|
|
|
a) > 4 times a month or daily
|
0.18
|
0.241
|
0.63**
|
0.212
|
|
b) 2–4 times a month
|
-0.43
|
0.385
|
0.31
|
0.373
|
|
c) < once a month or never
|
(ref. category)
|
-
|
(ref. category)
|
-
|
5
|
Grocery Shopping
|
|
|
|
|
|
a) > 4 times a month or daily
|
0.56
|
0.289
|
1.28***
|
0.277
|
|
b) 2–4 times a month
|
0.02
|
0.202
|
0.62**
|
0.213
|
|
c) < once a month or never
|
(ref. category)
|
-
|
(ref. category)
|
-
|
6
|
Non-Grocery Shopping
|
|
|
|
|
|
a) > 4 times a month or daily
|
-0.66
|
0.563
|
0.32
|
0.461
|
|
b) 2–4 times a month
|
0.70**
|
0.235
|
0.02
|
0.194
|
|
c) < once a month or never
|
(ref. category)
|
-
|
(ref. category)
|
-
|
7
|
Leisure Activity
|
|
|
|
|
|
a) > 4 times a month or daily
|
-0.13
|
0.388
|
2.97***
|
0.354
|
|
b) 2–4 times a month
|
0.07
|
0.213
|
2.11***
|
0.21
|
|
c) < once a month or never
|
(ref. category)
|
-
|
(ref. category)
|
-
|
Socio Demographics:
|
|
|
|
|
1
|
Gender
|
|
|
|
|
|
a) Man
|
-0.55*
|
0.291
|
0.22
|
0.216
|
|
b) Woman
|
(ref. category)
|
-
|
(ref. category)
|
-
|
2
|
Education
|
|
|
|
|
|
a) Highschool or under
|
0.39
|
0.845
|
0.44
|
0.583
|
|
b) Bachelor degree
|
0.61
|
0.81
|
0.61
|
0.555
|
|
c) Master degree or higher
|
(ref. category)
|
-
|
(ref. category)
|
-
|
3
|
Age
|
|
|
|
|
|
a) 18–25 years old
|
1.05
|
0.67
|
0.72
|
0.489
|
|
b) 25–35 years old
|
0.6
|
0.532
|
0.45
|
0.4
|
|
c) More than 35 years old
|
(ref. category)
|
-
|
(ref. category)
|
-
|
4
|
Marital Status
|
|
|
|
|
|
a) Not Married
|
0.39
|
0.42
|
0.54
|
0.304
|
|
b) Married
|
(ref. category)
|
-
|
(ref. category)
|
-
|
5
|
Income per-month
|
|
|
|
|
|
a) No monthly income
|
-1.5**
|
0.528
|
-0.01
|
0.396
|
|
b) < 1 M IDR
|
-1.28*
|
0.53
|
0.36
|
0.385
|
|
c) 1–3 M IDR
|
-0.61
|
0.428
|
0.37
|
0.312
|
|
d) > 3 M IDR
|
(ref. category)
|
-
|
(ref. category)
|
-
|
6
|
Employment status
|
|
|
|
|
|
a) Students
|
-0.1
|
0.413
|
0.47
|
0.289
|
|
b) Workers
|
-0.49
|
0.396
|
-0.51
|
0.283
|
|
c) No Job
|
(ref. category)
|
-
|
(ref. category)
|
-
|
7
|
Living Area
|
|
|
|
|
|
a) Yogyakarta City
|
0.55
|
0.3
|
0.12
|
0.248
|
|
b) Bantul Region
|
-0.34
|
0.287
|
0.27
|
0.214
|
|
c) Sleman Region
|
(ref. category)
|
-
|
(ref. category)
|
-
|
Household Characteristics:
|
|
|
|
|
1
|
Home Status
|
|
|
|
|
|
a) Own by myself or my family
|
-0.24
|
0.438
|
0.19
|
0.384
|
|
b) Rent
|
(ref. category)
|
-
|
(ref. category)
|
-
|
2
|
Type of House
|
|
|
|
|
|
a) Landed house
|
-0.64
|
0.358
|
-0.38
|
0.307
|
|
b) Apartment/boarding/others
|
(ref. category)
|
-
|
(ref. category)
|
-
|
3
|
Total Family Member
|
|
|
|
|
|
a) < 2 members
|
-0.37
|
0.576
|
-0.62
|
0.417
|
|
b) 2–4 members
|
-0.01
|
0.517
|
-0.72
|
0.381
|
|
c) 4–6 members
|
-0.19
|
0.532
|
-0.7
|
0.396
|
|
d) > 6 members
|
(ref. category)
|
-
|
(ref. category)
|
-
|
4
|
Driving License
|
|
|
|
|
|
a) Yes
|
0.89**
|
0.341
|
0.07
|
0.272
|
|
b) No
|
(ref. category)
|
-
|
(ref. category)
|
-
|
5
|
Motorcycle License
|
|
|
|
|
|
a) Yes
|
-0.07
|
0.306
|
0.06
|
0.225
|
|
b) No
|
(ref. category)
|
-
|
(ref. category)
|
-
|
6
|
Car in the Household
|
|
|
|
|
|
a) 0 car
|
-1.42*
|
0.654
|
-0.8
|
0.535
|
|
b) 1 car
|
-0.32
|
0.348
|
-0.51
|
0.278
|
|
c) 2 cars
|
-0.53
|
0.329
|
0.56
|
0.269
|
|
d) > 2 cars
|
(ref. category)
|
-
|
(ref. category)
|
-
|
7
|
Motorcycle in the Household
|
|
|
|
|
|
a) 0 motorcycle
|
0.87
|
0.713
|
0.09
|
0.839
|
|
b) 1 motorcycle
|
0.91**
|
0.717
|
-0.25
|
0.834
|
|
c) 2 motorcycles
|
0.16
|
0.921
|
-0.06
|
0.98
|
|
d) > 2 motorcycles
|
(ref. category)
|
-
|
(ref. category)
|
-
|
8
|
Bicycle in the Household
|
|
|
|
|
|
a) 0 bicycle
|
0.29
|
0.406
|
-0.12
|
0.366
|
|
b) 1 bicycle
|
-0.01
|
0.449
|
-0.26
|
0.391
|
|
c) 2 bicycles
|
0.48
|
0.525
|
-0.03
|
0.407
|
|
d) > 2 bicycles
|
(ref. category)
|
-
|
(ref. category)
|
-
|
ICT literacy:
|
|
|
|
|
|
Internet Duration
|
|
|
|
|
|
a) 0–2 hours / day (0 or 1)
|
-0.19
|
0.553
|
-1.14*
|
0.472
|
|
b) 2–4 hours / day (0 or 1)
|
-0.53*
|
0.276
|
-0.57*
|
0.222
|
|
c) > 4 hours / day (base = 0)
|
(ref. category)
|
-
|
(ref. category)
|
-
|
Timeframe
|
|
|
|
|
1
|
Time Before COVID (0 or 1)
|
0.46
|
0.4
|
2.4***
|
0.408
|
2
|
Time During COVID (base = 0)
|
(ref. category)
|
-
|
(ref. category)
|
-
|
Attitudes:
|
|
|
|
|
1
|
Convenience
|
0.68**
|
0.218
|
|
|
2
|
Safety and health concern
|
-0.43*
|
0.21
|
|
|
Variable Interaction
|
|
|
|
|
1
|
Women * During COVID
|
0.1
|
0.252
|
-0.81**
|
0.292
|
2
|
Having 1 motorcycle * During COVID
|
-0.67**
|
0.25
|
-0.15
|
0.295
|
3
|
Age 18–25 years old * During COVID
|
0.76*
|
0.321
|
0.85*
|
0.378
|
Threshold:
|
|
|
|
|
|
\({\kappa }_{1}\)
|
-1.26
|
1.538
|
3.29
|
1.296
|
|
\({\kappa }_{2}\)
|
2.81
|
1.543
|
6.74
|
1.322
|
Random Effect
|
|
|
|
|
|
\({\sigma }_{u}^{2}\)
|
6.69
|
1.069
|
2.74
|
0.61
|
Goodness of fit:
|
|
|
|
|
|
Number of observations
|
1,610
|
1,610
|
|
Number of groups
|
805
|
805
|
|
Log Likelihood (model)
|
-1,407
|
-1,117
|
|
df
|
52
|
50
|
|
AIC
|
2917.499
|
2333.059
|
|
BIC
|
3197.467
|
2602.258
|
Significant Level: *** p < 0.001; ** p < 0.01; * p < 0.05.
|
Regarding the effect of COVID, we find—in line with the earlier reported aggregate numbers— that people slightly (but not significantly) reduce the frequency of food delivery during COVID and strongly reduce the frequency of eating out. However, these effects occur differently for different groups, as indicated by the estimated interaction effects. Younger people (,25) reduce the frequency of home delivery less during COVID, while people with one motorcycle reduce food delivery more. Young people also reduce eating out less than older people, but women reduce the frequency of eating out more. Altogether, this suggests that older people (i.e. older than 25 years are more sensitive to the health risks of COVID in their food consumption activities, and that women are more sensitive than men.
Based on the result of sigma panel-level standard deviation from both models, we found that food delivery has the higher amount of variation (6.69) compared with eating out model (2.74). This indicates that food delivery has more variability in the frequency and users’ characteristics, including their activities behavior compared with eating out activity.