The Effectiveness of Telecommuting as a Transportation Control Measure

Full text

THE EFFECTIVENESS OF
TELECOMMUTING AS A
TRANSPORTATION CONTROL MEASURE
by
Srikanth Sampath,
Somitra Saxena, and
Patricia L. Mokhtarian
Institute of Transportation Studies
and
Department of Civil Engineering
University of California
Davis, CA 95616
(916) 752-7062
Fax: (916) 752-6572
August 1991
Institute of Transportation Studies
Research Report Number
UCD-ITS-RR-91-10
Proceedings of the ASCE Urban Transportation Division
National Conference on Transportation Planning and Air Quality,
Santa Barbara, CA, July 28-31, 1991 ( 1992), 347-362

THE EFFECTIVENESS OF TELECOMMUTING
AS A TRANSPORTATION CONTROL MEASURE
Srikanth Sampath1
Somitra Saxena1
Patricia L. Mokhtarian2
ABSTRACT
This paper examines the potential of telecommuting as a strategy for managing travel
demand. In particular, the paper focuses on the travel and air quality implications of
telecommuting. A study of travel impacts has been carried out using data obtained from the State
of California Telecommuting Pilot Project. This paper presents preliminary findings from the first
known empirical study of the emission impacts of telecommuting.
Previously-reported travel-related findings include significant reductions in work trips,
peak-period travel and distance travelled due to telecommuting, while no increase was found in
non-work trips. New emission-related findings include substantial reductions in the number of cold
starts (60% fewer), and emissions of organic gases (64% lower), carbon monoxide (63% lower),
and oxides of nitrogen (73% lower) on telecommuting days. These reductions are nearly
proportional to the decrease in distance travelled by auto (76%). Work is ongoing to refine and
extend the analysis of emissions impacts.
---------------------
1 Post Graduate Research Assistant, Department of Civil Engineering and Institute of
Transportation Studies, University of California, Davis, CA 95616.
2 Assistant Professor, Department of Civil Engineering and Institute of Transportation Studies,
University of California, Davis, CA 95616.
INTRODUCTION
Transportation Control Measures (TCMs), strategies for reducing travel or improving the
flow of traffic, form an important component of plans designed to improve air quality in
non-attainment areas. One TCM that is still fairly novel is telecommuting, defined as "the partial or
total substitution of telecommunications, with or without assistance of computers, for the twice
daily commute to/from work" [Nilles, 1988]. While early treatments of telecommuting envisioned
it to involve computer-based information employees, working from home, full time, it is now
widely acknowledged that none of those characteristics are essential [Mokhtarian, 1991a]. In a
number of telecommuting programs, the average participation rate is 1-2 days a week. A large
variety of jobs deal with information to the extent that part- time telecommuting at those levels is
1

feasible. And while telecommuting today is predominantly home-based, several telecommuting
center demonstrations are currently taking place throughout the country.
This paper presents preliminary findings from the first known empirical study of the
emission impacts of telecommuting. The organization of the paper is as follows: the next section
describes several public policy documents in which telecommuting plays a role as a TCM. The
following section reviews some of the commonly-raised hypotheses about the transportation-related
impacts of telecommuting. Next, travel-related factors affecting the air quality impacts of telecom-
muting are described. The succeeding three sections present the empirical research on emission
impacts of telecommuting, including a description of the study setting and data collection, travel
findings, and emissions findings. The closing section discusses conclusions and directions for
future research.
TELECOMMUTING IN PUBLIC POLICY
There is growing interest among planners, researchers and policy makers in telecommuting
as a strategy for reducing travel demand. No one suggests that telecommuting alone will provide
the solution to congestion and pollution problems, but it does have appeal as one among many
measures for addressing these problems. This appeal can be traced to several factors: First,
telecommuting can be implemented now, as it does not require any lengthy planning, design and
construction lead times. Second, it is relatively inexpensive to implement, and third, it expands
personal choices rather than restricting them, by offering more flexibility in workstyle and lifestyle.
Finally, it addresses a variety of public- and private-sector concerns. For businesses, it offers the
potential for improved productivity, recruitment and retention, savings in space costs, and other
benefits. For public policy makers, telecommuting can contribute to issues dealing with the
American family, employment for people with disabilities, rural economic development, global
competitiveness, health care, and community involvement -- in addition to transportation, energy
and air quality.
On the basis of this multi-dimensional appeal, telecommuting has found its way into a
number of public policy statements -- especially as a transportation control measure, but also
addressing other policy concerns. For example:
• The 1989 Air Quality Management Plan for the South Coast (California) Air Basin sets the
goal of reducing work trips by 30% in the year 2010, due to the combined impacts of
telecommuting and alternative work schedules (SCAQMD and SCAG, 1989).
• Regulation XV of the South Coast Air Quality Management District (SCAQMD, 1990)
includes telecommuting on a menu of strategies large employers must use to decrease
peak-period vehicle trips.
• California Governor George Deukmejian issued Executive Order D-82-89 on October 30,
1989, which directed state agencies to include telecommuting in their emergency response
to the Loma Prieta (Northern California) earthquake.
2

• The statement of National Transportation Policy (USDOT, 1990) has a short but positive
section on telecommuting, and President Bush has publicly endorsed the concept.
• Legislation supportive of telecommuting has recently been enacted by the States of
California (Chapter 1389, 1990), Washington (Chapter 202, 1991), Florida (Chapter
90-291, 1990) and Virginia (HJRs 77 [1990] and 339 [1991]). In particular, the State of
Washington statute requires trip reduction plans to be prepared and provides a "20% bonus"
for work-at-home and alternative work schedule strategies. That is, a vehicle trip reduced
by one of those measures counts as 1.2 trips reduced.
POTENTIAL TRAVEL RELATED IMPACTS OF TELECOMMUTING
Telecommuting leads to a certain amount of change in the lifestyle and hence travel
behavior of the telecommuter. Furthermore, it may also lead to changes in the travel behavior of
other household members [Garrison and Deakin, 1988]. These potential impacts are not always
positive from a transportation policy maker's point of view. A comprehensive assessment of the
possible impacts of telecommuting is therefore important to an evaluation of the effectiveness of
telecommuting as a travel demand reduction strategy.
Many different hypotheses have been formulated regarding the impacts of telecommuting
on household travel [Jovanis, 1983; Salomon, 1986]. They can be classified into short-term,
medium-term and long-term hypotheses.
First, consider hypotheses regarding short-term travel behavior. One of the more obvious
results of telecommuting should be the reduction of commute trips. A direct consequence of this
will likely be the reduction of peak hour trips, as most often work trips are made during peak hours.
However, due to factors such as a psychological need for mobility, the availability of a vehicle to
other household members, or the need to engage in work-related activities directly because of tele-
commuting (e.g., trips to the office supply store or post office), non-commute trips may increase.
The eliminated need to travel to work may increase flexibility in activity scheduling. Given
the flexibility to do so, the timing of trips may change. Trips may be shifted to off-peak periods to
avoid congestion delays, and/or to different days of the week.
The flexibility and lower frequency of work trips brought about by telecommuting may
have negative impacts on mode choice. Carpools and vanpools might dissolve if telecommuters
drop out, and transit operators may lose revenue. On the positive side, trips made closer to home
may shift to non-motorized modes such as bicycle and walk. Another consequence might be
changes in destination choice. Work trips may now be made to a local center rather than to a more
distant office building; non-work trips may be made closer to home rather than closer to work. This
may have negative impacts on the local street networks (although it should have positive economic
impacts on local businesses).
3

Elimination of the work trip may break up efficiently-linked activity patterns, creating
several one-stop trips instead of one multi-stop trip. Finally, telecommuting might cause
reallocation of activities among household members, resulting in changed travel patterns.
In the medium term, telecommuting might lead to changes in vehicle ownership. The
ability to telecommute may eliminate the need for a car altogether, or more likely the need for a
second car. In the long term, telecommuting reduces the need to reside close to the work site.
Hence, a long-term impact of telecommuting might be a shift to housing in more desirable and/or
more affordable outlying locations. This may or may not lead to increases in travel. Job location
may change as well. Once the ability to telecommute has been established, the worker may change
jobs, moving to a more distant employer. Or, telecommuting may make it feasible to move a
corporate facility without either relocating or losing some employees.
FACTORS AFFECTING POTENTIAL AIR QUALITY IMPACTS OF
TELECOMMUTING
The air quality impacts of telecommuting can either be direct or indirect. Direct impacts are
those resulting from changes in transportation due to telecommuting. Indirect impacts include the
net air quality effects of non-transportation energy consumed while telecommuting (see CEC, 1983
and JALA Associates, 1990 for hypothetical and empirical evaluations of direct and indirect energy
impacts of telecommuting. These findings obviously have implications for air quality). This paper
only addresses the direct air quality impacts of telecommuting.
The transportation-derived air quality impacts of telecommuting may or may not be as
favorable as the underlying transportation impacts themselves. A number of factors affect the direct
air quality impacts of telecommuting. These include: distance travelled by auto, number of cold
starts, number of hot starts, speed, type of vehicle, and ambient temperature [Horowitz, 1982].
Each of these is discussed below.
The distance travelled by auto is important since, other things being equal, the lower the
distance travelled, the lower the emissions. Thus, if telecommuting results in a reduction of
distance travelled through the elimination of commute trips, positive impacts on emissions will
result. Conversely, if increased non-work trips or a move to a more distant location leads to a net
increase in auto distance travelled, the impact on emissions will be undesirable.
The number of cold starts is important because cold starts have higher emission rates than
warmed-up vehicles. (A vehicle's engine is cold if it has been turned off for more than one hour for
vehicles with a catalytic converter, and four hours for vehicles without a catalytic converter). If
telecommuting were to generate additional trips or more unlinked trips, then the number of cold
starts could increase, resulting in higher emissions. The number of hot starts or the number of
stops in a trip is also a factor to be considered, as the emissions during a hot start are higher than
during the stabilized phase.
4

In general, there is a U-shaped relationship between speed and emissions [California Air
Resources Board, 1990]. That is, emissions decline as speed increases, up to about 80-96
kilometers per hour (50-60 miles per hour), then increase with higher speeds. Further, trips with
more accelerations and decelerations result in higher emissions than those with constant speed. If
telecommuting promotes off-peak travel, with fewer accelerations and decelerations at higher (but
not too high) average speeds, it can be beneficial for air quality.
Cold start emissions are sensitive to the surrounding air temperature. In general, the lower
the ambient temperature, the higher the emissions. If telecommuting reduces trips in early morning
and late evening hours and induces trips to be made later in the daytime, it may have a significant
positive effect on air quality.
Emissions are also somewhat dependent on the type of vehicle used in trip-making. For
example, the presence or absence of a catalytic converter affects the emissions from a vehicle.
Emissions are different between diesel engines and gasoline-powered engines. Diesel engines tend
to have lower hydrocarbon (HC) and carbon monoxide (CO) emissions and considerably higher
particulate emissions than gasoline engines. Telecommuting may prompt a reassignment of
vehicles within households. For example, the telecommuter may use a less fuel-efficient vehicle
for the shorter trips being made on telecommuting days, leaving the more efficient vehicle to be
used by the spouse for commuting. If the two vehicles belong to different classes, changes in
emissions may result. Thus, all the vehicles in the household should be considered to fully analyze
the transportation-related impacts of telecommuting on air quality.
SETTING FOR THE EMPIRICAL ANALYSIS
Many of the hypotheses discussed above have been tested in the context of the two-year
State of California Telecommuting Pilot Project (see JALA Associates, 1990 for an overall
evaluation of the project). The sample for this study included more than 200 state workers as
telecommuters and control group members.
Three-(consecutive)-day travel diaries were completed by all state participants and their
driving-age household members. These diaries were completed in two waves (i.e, at two points in
time). In the first wave, which spanned from January to June 1988 (due to the gradual phasing-in of
telecommuting), all employees commuted to work conventionally. In the second wave, which
covered April-June 1989, telecommuting had been in effect for about a year. For the
telecommuters in the sample, the wave 2 diaries were specified to contain at least one telecom-
muting day.
The empirical findings reported below pertain to the 73 "stayer" telecommuters for whom
before and after trip diaries were available. Additional transportation findings, including an
analysis of 65 control group employees, 54 telecommuter household members and 36 control group
household members are reported in Kitamura, et al., (1991) and Pendyala, et al. (1991).
5

Two types of data files were created with the travel diaries. The first one contains personal
and household information and the other contains trip information. The person file contains
information such as the participant status (telecommuter, control group member, telecommuter
household member, or control group household member), age, gender, home and work locations,
locations frequently visited, transit lines used and household car ownership.
The trip files contain the trip characteristics for every trip reported by the respondents. The
information for each trip includes the origin and destination, beginning and ending trip times,
purpose, approximate trip length as reported by the respondent, mode used, beginning and ending
odometer readings if a car were used, the number of passengers and the percentage of time spent on
the freeway for each trip. The complete trip files contain 2706 first wave trips and 2235 second
wave trips. For the 73 telecommuters, the files contain 874 trips in wave 1 and 680 trips in wave 2.
The person and trip files formed the basis for analyzing the travel impacts of telecom-
muting, reported in the following section. To analyze emission impacts, however, the information
was needed in a different form. Emissions are vehicle-based, not person-based. While the
person-trip files indicated which vehicle was being used for each trip, it would not be possible to
determine with certainty from them whether a certain trip involved a hot start or a cold start for the
vehicle used.
Thus, it was necessary to create vehicle movement profiles, itemizing the trips made by each
household vehicle in the sample throughout the three-day diary period. The "vehicle profile"
contains the following information for every trip made by every household vehicle: trip duration;
the trip length; the time parked; origin and destination coordinates; the average speed; vehicle
make, model and year; vehicle class (light-duty auto, light-duty truck, medium-duty truck, heavy-
duty truck, and motorcycle); participant status (for the driver of the vehicle); the trip purpose; and
percentage of freeway use. The vehicle profiles contain 2061 trips in wave 1 and 1726 trips in
wave 2 for the full sample of 219 stayers. For the sample of 73 telecommuters, the vehicle profiles
contain 722 wave 1 trips and 549 wave 2 trips.
TRAVEL IMPACTS
Table 1 summarizes some travel-related indicators before and after telecommuting began.
All the statistics in the second wave (after telecommuting) are further divided into days on which
the employee telecommuted and days on which the employee commuted conventionally. Any
characteristic in the second wave that is marked with an asterisk is significantly different (at a 5%
level of significance) from the first wave. It is observed that on average, telecommuters eliminated
two trips on telecommuting days (the trips to and from work) while showing no significant change
on non-telecommuting days. They made almost no work trips on telecommuting days. Also,
telecommuters made significantly fewer car trips on telecommuting days. Thus, the hypothesis that
the telecommuter may make additional trips due to "cabin fever" or work-related requirements is
not supported by the data. While household-level impacts are not discussed in detail here, the
evidence shows that household members did not increase their travel, either [Pendyala, et al.,
1991].
6

TABLE 1
TRAVEL IMPACTS OF TELECOMMUTING
BEFORE AFTER
(telecom)
AFTER
(commute)
Number of trips in sample
874
184
496
Total trips
3.99 1.94* 4.00
Work trips
1.02 0.09* 1.11
Car trips
3.25 1.77* 3.25
AM Peak
0.89 0.24* 0.82
PM Peak
0.99 0.46* 1.16
Total km (miles),
all modes
85.9
(53.7) 21.1*
(13.2*)
89.8
(56.1)
AVERAGE PER PERSON PER DAY FOR SAMPLE OF 73 TELECOMMUTERS
* SIGNIFICANTLY DIFFERENT FROM "BEFORE" WITH 95% CONFIDENCE
Source : Pendyala, et al., 1991.
Large reductions in peak period travel are observed on telecommuting days, more so for the
AM peak but significant for both peaks. There is no significant change in peak-period travel on
non-telecommuting days. There is also more than a 75% reduction in total distance travelled (by all
modes) on telecommuting days, while there is no significant increase on non-telecommuting days.
The reduction in total distance travelled, along with the reductions in car trips and peak-period trips,
suggest that telecommuting has promise as a strategy for reducing congestion and improving air
quality. The air quality implications of these positive transportation findings are examined below.
7

EMISSION IMPACTS
Table 2 summarizes some emissions-related indicators before and after telecommuting.
The first four rows of figures present travel factors relevant to emissions; number of cold starts,
number of hot starts, average speed, and auto kilometers travelled. The last three rows of numbers
present actual average emissions, taking into account the effects of those travel indicators. For the
four travel factors, the statistics in the second wave which are significantly different from those in
the first wave are marked with an asterisk.
As would be expected from the reduction in trips shown in Table 1, a significant reduction
in the number of cold starts is apparent, from more than 2 per day before telecommuting, to fewer
than 1 on telecommuting days. No significant change is found on non-telecommuting days. The
number of hot starts also decreased on both telecommuting and non-telecommuting days, though
the change is statistically insignificant (at the 5% level).
The reduction in average speed on telecommuting days is important, and counter to the
hypothesis that travel would shift to off-peak times and uncongested facilities where speeds would
be higher. The observed decrease is due to the fact that trips on telecommuting days are more likely
to be shorter, local trips, involving a much lower proportion of freeway usage [Pendyala, et al.,
1991]. The auto distance travelled declined by 76% on telecommuting days and reduced
marginally on non-telecommuting days also.
The reduction in the number of cold starts and reduction in distance travelled by auto will
have a beneficial impact on air quality, while the reduction in average speed will work in the
opposite direction. The net impact on emissions is discussed below.
In this preliminary analysis, the emissions calculations were performed using the AQAT-3
program of the California Air Resources Board (Randall and Diamond, 1989). AQAT-3 is a
microcomputer software package containing simplified versions of programs commonly used for
air quality analysis in California. This analysis employed the EMFAC7D program of the package,
using the program defaults for temperature (24o C / 75o F) and fleet age mix. User-specified inputs
included the year in which the emissions are to occur, percentage vehicle miles traveled (VMT) by
vehicle class, average speed, and the percentage of VMT in cold start and hot start modes.
The results for total organic gases (TOG), carbon monoxide (CO), and oxides of nitrogen
(NOx) are as shown in the rest of Table 2. On telecommuting days the reduction in emissions of
TOG and CO are 64% and 63% respectively. There is a 73% reduction in NOx emissions on
telecommuting days. Even on non-telecommuting days there is a modest decrease in the emissions
of those three classes of pollutants.
8

TABLE 2
EMISSION IMPACTS OF TELECOMMUTING
BEFORE AFTER
(telecom)
AFTER
(commute)
Number of trips in sample
722
163
386
Cold starts
2.26 0.92* 2.23
Hot starts
0.99 0.74 0.81
Average speed,
kmph (mph)
47.0
(29.4) 37.1*
(23.2)* 47.8
(29.9)
Auto VKmT
(VMT)
79.5
(49.7) 19.2*
(12.0)* 71.0
(44.4)
TOG (gms) **
45.2 16.1 41.7
CO (gms) **
467.7 175.0 433.3
NOx (gms) **
49.7 13.4 44.8
AVERAGE PER PERSON PER DAY FOR SAMPLE OF 73 TELECOMMUTERS
AQAT - 3 (EMFAC7D, 24o C / 75o F , DEFAULT FLEET AGE MIX)
* SIGNIFICANTLY DIFFERENT FROM "BEFORE" WITH 95% CONFIDENCE.
** NO STATISTICAL TEST PERFORMED ON THESE INDICATORS.
9

It is worth pointing out that emission rates, in grams per unit distance, are actually higher on
telecommuting days than non-telecommuting days. For example, the rate for CO is 5.9 gm/km (9.4
gm/mi) before telecommuting, and 9.1 gm/km (14.6 gm/mi) on telecommuting days, a 55%
increase. Rates for TOG and NOx are 32% and 11% higher, respectively, on telecommuting days.
The rates are higher for two reasons: first, the average speeds are lower on telecommuting
days, as mentioned earlier. Second, even though the number of cold starts and hot starts are lower
on telecommuting days, the proportion of distance traveled in cold start and hot start modes is
higher (since total distance traveled by auto is so much lower). In the EMFAC model, the
emissions rate calculations (especially for CO and TOG) are a function of the proportion of distance
in cold start and hot start modes. Nevertheless, even though emissions rates are higher on
telecommuting days, multiplying the higher factor of grams per unit distance by the far lower
distance traveled results in total emissions that are still greatly reduced due to telecommuting.
It is further of interest to compare the reductions in emissions against the reductions in
vehicle distance traveled. A priori, no specific relationships between the two can be assumed: due
to the mitigating influences already discussed (number of cold and hot starts, speed, and so on), the
emissions reductions may be higher or lower than the reductions in distance.
The ratio of emissions reductions (for a given class of pollutant) to distance reduction might
be taken as an index of efficiency for a particular TCM oriented toward reducing travel. This ratio
can theoretically take on any non-negative value, but in practice is likely to fall between zero and
one. For the data presented here, the index of efficiency is 0.85 for TOG (64.4% reduction in
emissions compared to 75.9% reduction in distance traveled), 0.83 for CO, and 0.96 for NOx. The
index for NOx is close to unity because that class of pollutants is less affected by cold starts and by
changes in speed, within the range of the trips in this sample. In all cases, however, the index of
efficiency is quite high, indicating that emissions benefits nearly proportional to the reductions in
distance traveled are being achieved.
CONCLUSIONS AND DIRECTIONS FOR FUTURE RESEARCH
The results of the empirical analysis reported in this paper present strong evidence that
telecommuting is a viable transportation control measure. It appears to have beneficial
transportation and air quality impacts, at least in the specific context studied here. In brief, the
findings are:
• Telecommuters in the Pilot Project reported making virtually no work trips on telecom-
muting days.
• Counter to hypothesis, no increase in non-commute trips was observed for telecommuters
(or their household members). Thus, on average two fewer trips were made by the
telecommuter on telecommuting days.
10

• Distance traveled by auto declined by 76% (60.3 km, or 37.7 miles per telecommuter on
average) on telecommuting days.
• On telecommuting days, telecommuters showed a 60% reduction in the number of cold
starts per person per day, while no significant change was observed on the commuting days.
The number of hot starts per person per day decreased by 25%.
• As a consequence, significant reductions in emissions were observed on telecommuting
days: 64% for TOG, 63% for CO, and 73% for NOx.
• Lower average speeds and higher proportions of travel in cold and hot start modes, had the
combined effect that emissions rates (grams per unit distance) were higher on telecom-
muting days. However, total distance traveled was so much lower that the total emissions
still declined substantially.
• The ratios of emissions reductions to reduction in distance traveled were quite high: 0.85 for
TOG, 0.83 for CO, and 0.96 for NOx. This indicates that telecommuting (at least for this
study) is a relatively efficient TCM.
Several near-term extensions of this study are currently underway. Additional insight will
be gained through analyzing not only trips made by telecommuters, but also trips made by their
household members and by the control group households. A more refined and detailed air quality
analysis will be performed by using EMFAC7E instead of the simplified EMFAC7PC version used
here. By using EMFAC7E, actual fleet age mix values can be used, instead of the default fleet age
mix used in this analysis.
While the travel and preliminary air quality findings presented here are encouraging, it must
be remembered that they represent only one application of telecommuting, and deal only with short-
term, day-to-day travel behavior. In the long term, several questions remain to be answered
regarding the large-scale transportation and air quality impacts of telecommuting.
First, the impacts of telecommuting on residential location must be monitored. Preliminary
evidence suggests that telecommuting will motivate at least some people to move significantly
further away from the work place [Mokhtarian, 1991b]. The important question is whether these
moves will be the exception (whose negative impacts will be outweighed by the travel savings for
the many who do not move) or the rule.
Second, the findings presented here apply to the home-based form of telecommuting. A
number of people expect to see substantial future growth in the center-based forms of
telecommuting, which provide certain advantages over the home in many cases. The air quality
benefits of telecommuting are likely to be lower for telecommuting centers, because a vehicle trip
(although a much shorter one than a conventional commute) may still be made from home to the
center. However, that trip may be combined with other trips (such as to a child care center) in such
a way as to have little impact on emissions. And even if air quality benefits are lower for
11

telecommuting centers, they may still be worth achieving. Thus, this form of telecommuting
deserves additional study.
Finally, the most critical question regarding the transportation/air quality impacts of
telecommuting is whether enough people will do it, often enough, to matter. It is vital, therefore, to
develop causal/ behavioral models of the adoption of telecommuting. Included in that effort should
be identification of barriers to adoption and the likely importance of those barriers in the future.
ACKNOWLEDGEMENTS
This research was funded by the State of California Department of Transportation, the
University of California Transportation Center, and the University of California Energy Research
Group. Significant contributions to the project as a whole, including earlier work reviewed in this
paper, have been made by Prof. Ryuichi Kitamura, Prof. David S. Bunch, Ram M. Pendyala, Kons-
tadinos G. Goulias, and Huichun Zhao.
REFERENCES
California Air Resources Board (1990) Derivation of the EMFAC7E Emission and Correction
Factors for On-road Motor Vehicles (Exec. Sum.), Mobile Source Division, Sacramento, CA, July.
California Energy Commission (CEC) (1983) Telecommunications and Energy: The Energy
Conservation Implications for California of Telecommunications Substitutes for Transportation,
prepared by JALA Associates, Document No. P400-83-042, December.
Garrison, W.L. and E. Deakin (1988) Travel, work, and telecommunications: A view of the
electronics revolution and its potential impacts. Transportation Research A, 22A(4), 239-245.
Horowitz, J.L. (1982) Air Quality Analysis for Urban Transportation Planning, Cambridge, MA:
MIT Press.
JALA Associates, Inc. (1990) California Telecommuting Pilot Project Final Report. Stock No.
7540-930-1400-0. State of California Department of General Services. North Highlands, CA
95660, June.
Jovanis, P. (1983) Telecommunications and alternative work schedules: Options for managing
transit travel demand. Urban Affairs Quarterly, Vol. 19, No. 2 (December), pp. 167-189.
Kitamura, R., P.L. Mokhtarian, R.M. Pendyala, and K.G. Goulias (1991) An evaluation of
telecommuting as a trip reduction measure. Institute of Transportation Studies, University of
California, Davis, CA.
Mokhtarian, P.L. (1991a) Defining telecommuting. Forthcoming Transportation Research Record.
12

Mokhtarian, P.L. (1991b) Telecommuting and travel: State of the practice, state of the art.
Forthcoming in Transportation, Volume 18, Number 4.
Nilles, J. M. (1988) Traffic reduction by telecommuting: A status review and selected bibliography.
Transportation Research A, 22A(4), 301-317.
Pendyala, R.M., K.G. Goulias, and R. Kitamura (1991) Impact of telecommuting on spatial and
temporal patterns of household travel. Forthcoming in Transportation, Volume 18, Number 4.
Randall, P.C. and A. Diamond (1989) Air Quality Analysis Tools (AQAT-3), State of California Air
Resources Board, Stationary Source Division, Sacramento, CA.
Salomon, I. (1986) Telecommunications and travel relationships: A review. Transportation
Research 20A, 223-238.
South Coast Air Quality Management District and Southern California Association of Governments
(1989), 1989 Air Quality Management Plan, El Monte, CA, March.
South Coast Air Quality Management District (1990) Regulation XV: Trip Reduction/Indirect
Source. El Monte, CA, adopted December 11, 1987; amended May 17, 1990.
U. S. Department of Transportation (1990) Moving America: New Directions, New Opportunities,
February, p.17.
13