How do activities conducted while commuting influence mode
choice? Using revealed preference models to inform public
transportation advantage and autonomous vehicle scenarios
Aliaksandr Malokin, Giovanni Circella, Patricia L. Mokhtarian, 2019, in Transportation Research Part A
doi:10.1016/j.tra.2018.12.015
Location |
Northern California |
Population |
Other (specify) |
Sample size |
2229 |
Factor analysis type |
exploratory factor analysis, Oblique rotation |
Stepwise regression |
no |
Removal of insignificant variables |
yes |
Reviewed by |
SH |
Abstract
From early studies of time allocation onward, it has been acknowledged that the “productive” nature of travel could affect its utility. Currently, at the margin an individual may choose transit over a shorter automobile trip, if thereby she is able to use the travel time more productively. On the other hand, recent advancements toward partly/fully automated vehicles are poised to revolutionize the perception and utilization of travel time in cars, and are further blurring the role of travel as a crisp transition between location-based activities. To quantify these effects, we created and administered a survey to measure travel multitasking attitudes and behaviors, together with general attitudes, mode-specific perceptions, and standard socioeconomic traits (N = 2229 Northern California commuters). In this paper, we present a revealed preference mode choice model that accounts for the impact of multitasking attitudes and behavior on the utility of various alternatives. We find that the propensity to engage in productive activities on the commute, operationalized as using a laptop/tablet, significantly influences utility and accounts for a small but non-trivial portion of the current mode shares. For example, the model estimates that commuter rail, transit, and car/vanpool shares would respectively be 0.11, 0.23, and 1.18 percentage points lower, and the drive-alone share 1.49 percentage points higher, if the option to use a laptop or tablet while commuting were not available. Conversely, in a hypothetical autonomous vehicles scenario, where the car would allow a high level of engagement in productive activities, the drive-alone share would increase by 1.48 percentage points. The results empirically demonstrate the potential of a multitasking propensity to reduce the disutility of travel time. Further, the methodology can be generalized to account for other properties of autonomous vehicles, among other applications.
Factors
Models
Dependent variable |
propensity to use a laptop, netbook, or tablet computer |
Model type |
binary logit |
Sample size |
265.0 |
R2 |
0.9051 |
Adjusted R2 |
|
Pseudo R2
|
nan |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
-183.684 |
Log-likelihood at constants |
-16.426 |
Log-likelihood at convergence |
-16.426 |
Variable |
Coefficient |
p-value |
Constant |
-4.47
|
0.0 |
Dependent variable |
propensity to use a laptop, netbook, or tablet computer |
Model type |
binary logit |
Sample size |
197.0 |
R2 |
0.3326 |
Adjusted R2 |
|
Pseudo R2
|
nan |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
-136.55 |
Log-likelihood at constants |
-136.426 |
Log-likelihood at convergence |
-84.128 |
Variable |
Coefficient |
p-value |
Constant |
0.313
|
0.705 |
Has to/would like to work on commute |
1.148
|
0.0 |
Would like to take same route |
-0.543
|
0.008 |
Female |
-1.36
|
0.002 |
Age |
-0.049
|
0.001 |
Hourly Waged (=1 if ‘yes’,=0 otherwise) |
-3.276
|
0.01 |
Travel distance, mi |
0.026
|
0.0 |
Dependent variable |
propensity to use a laptop, netbook, or tablet computer |
Model type |
binary logit |
Sample size |
811.0 |
R2 |
0.509 |
Adjusted R2 |
|
Pseudo R2
|
nan |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
-562.142 |
Log-likelihood at constants |
-292.922 |
Log-likelihood at convergence |
-272.025 |
Variable |
Coefficient |
p-value |
Constant |
-2.268
|
0.0 |
Pro-technology |
0.549
|
0.0 |
Polychronicity |
0.241
|
0.045 |
Has to/would like to work on commute |
0.368
|
0.001 |
Dependent variable |
propensity to use a laptop, netbook, or tablet computer |
Model type |
binary logit |
Sample size |
389.0 |
R2 |
0.5449 |
Adjusted R2 |
|
Pseudo R2
|
nan |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
-269.634 |
Log-likelihood at constants |
-186.341 |
Log-likelihood at convergence |
-113.711 |
Variable |
Coefficient |
p-value |
Constant |
-4.408
|
0.0 |
Travel is wasted time |
0.564
|
0.001 |
Has to/would like to work on commute |
1.262
|
0.0 |
Has to/would like to do recreation on commute |
0.685
|
0.002 |
Has to/would like to multitask at work |
-0.456
|
0.021 |
Has to/would like to be available to people |
0.486
|
0.009 |
Would like to take same route |
-0.383
|
0.042 |
Annual household per capita income, $000 |
-0.021
|
0.001 |
Travel distance, mi |
0.029
|
0.0 |
Dependent variable |
propensity to use a laptop, netbook, or tablet computer |
Model type |
binary logit |
Sample size |
1001.0 |
R2 |
0.799 |
Adjusted R2 |
|
Pseudo R2
|
nan |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
-693.84 |
Log-likelihood at constants |
-158.328 |
Log-likelihood at convergence |
-132.445 |
Variable |
Coefficient |
p-value |
Constant |
-2.178
|
0.0 |
Multitasking is normative |
0.401
|
0.03 |
Time spent working |
-0.372
|
0.045 |
Has to/would like to work on commute |
0.77
|
0.0 |
Has to do recreation on commute |
0.946
|
0.0 |
would like to do recreation on commute |
-0.389
|
0.091 |
Vehicle age |
-0.102
|
0.013 |
Dependent variable |
Mode: Driving alone (base), Biking, Commuter rail, Transit, Shared ride |
Model type |
multinomial logit |
Sample size |
2229.0 |
R2 |
0.5756 |
Adjusted R2 |
|
Pseudo R2
|
nan |
AIC |
nan |
BIC |
nan |
Log-likelihood at zero |
-2655.817 |
Log-likelihood at constants |
-1555.064 |
Log-likelihood at convergence |
-1127.247 |
Biking |
Variable |
Coefficient |
p-value |
Constant |
-5.327
|
0.0 |
In-vehicle travel time, min |
-0.163
|
0.006 |
out-of-vehicle travel time, min |
-0.048
|
0.0 |
one-way commute cost, ln($) |
-1.175
|
0.0 |
pro-active modes |
2.088
|
0.0 |
Mode convenience |
0.455
|
0.0 |
Mode benefit/cost |
0.368
|
0.0 |
Mode comfort |
0.405
|
0.0 |
Mode multitaskability |
0.098
|
0.023 |
Propensity to use a laptop/tablet/netbook |
1.24
|
0.0 |
Commuter rail |
Variable |
Coefficient |
p-value |
Constant |
-2.959
|
0.0 |
In-vehicle travel time, min |
-0.016
|
0.004 |
out-of-vehicle travel time, min |
-0.048
|
0.0 |
one-way commute cost, ln($) |
-1.175
|
0.0 |
Pro-transit |
0.954
|
0.001 |
Mode convenience |
0.455
|
0.0 |
Mode benefit/cost |
0.368
|
0.0 |
Mode comfort |
0.405
|
0.0 |
Mode multitaskability |
0.098
|
0.023 |
Propensity to use a laptop/tablet/netbook |
1.24
|
0.0 |
Driving alone |
Variable |
Coefficient |
p-value |
In-vehicle travel time, min |
-0.016
|
0.004 |
out-of-vehicle travel time, min |
-0.048
|
0.0 |
one-way commute cost, ln($) |
-1.175
|
0.0 |
Mode convenience |
0.455
|
0.0 |
Mode benefit/cost |
0.368
|
0.0 |
Mode comfort |
0.405
|
0.0 |
Mode multitaskability |
0.098
|
0.023 |
Propensity to use a laptop/tablet/netbook |
1.24
|
0.0 |
Shared ride |
Variable |
Coefficient |
p-value |
Constant |
-2.752
|
0.0 |
Female |
0.393
|
0.009 |
Limitation on walking |
0.166
|
0.003 |
In-vehicle travel time, min |
-0.016
|
0.004 |
out-of-vehicle travel time, min |
-0.048
|
0.0 |
one-way commute cost, ln($) |
-1.175
|
0.0 |
Pro-transit |
0.201
|
0.014 |
Polychronicity |
0.191
|
0.006 |
Mode convenience |
0.455
|
0.0 |
Mode benefit/cost |
0.368
|
0.0 |
Mode comfort |
0.405
|
0.0 |
Mode multitaskability |
0.098
|
0.023 |
Propensity to use a laptop/tablet/netbook |
1.24
|
0.0 |
Transit |
Variable |
Coefficient |
p-value |
Constant |
0.785
|
0.343 |
Driver's license |
-1.89
|
0.022 |
Race: white |
0.532
|
0.014 |
In-vehicle travel time, min |
-0.016
|
0.004 |
out-of-vehicle travel time, min |
-0.048
|
0.0 |
out-of-vehicle travel time, min |
-0.048
|
0.0 |
one-way commute cost, ln($) |
-1.175
|
0.0 |
one-way commute cost, ln($) |
-1.175
|
0.0 |
Pro-transit |
0.825
|
0.0 |
Mode convenience |
0.455
|
0.0 |
Mode convenience |
0.455
|
0.0 |
Mode benefit/cost |
0.368
|
0.0 |
Mode benefit/cost |
0.368
|
0.0 |
Mode comfort |
0.405
|
0.0 |
Mode comfort |
0.405
|
0.0 |
Mode multitaskability |
0.098
|
0.023 |
Mode multitaskability |
0.098
|
0.023 |
Propensity to use a laptop/tablet/netbook |
1.24
|
0.0 |
Propensity to use a laptop/tablet/netbook |
1.24
|
0.0 |