- Email: [email protected]
The relation between daily travel and use of the home computer
Transportation Research Part A 36 (2002) 437–452 www.elsevier.com/locate/tra
The relation between daily travel and use of the home computer Randi Johanne Hjorthol a
a,*
Institute of Transport Economics, Box 6110, Etterstad, N-0602 Oslo, Norway
Received 13 June 2000; received in revised form 26 January 2001; accepted 28 February 2001
Abstract On the basis of the Norwegian national personal travel survey (NPTS) 1997/98 and a connected mail back survey of the use of information – and communication technology at home, the relation between mobility and use of stationary communication has been studied. On the basis of these results we cannot see any direct substitutionary effects of the use of stationary technology at people’s home on the use of mobile technology. Access to and use of information technology seems not to have a significant impact on travel activities in everyday life. Stationary communication seems to be a supplement to activities based on mobile technology. For people who work more than ‘‘normal’’ weekly working hours, stationary technology seems to give them greater flexibility in regard to where to work, but it does not necessarily reduce their travel activity. There is a tendency that people who own home computers make less work trips, but this does not affect the total number of daily trips. The spatial flexibility give a temporal flexibility, which means that work trips and other trips can be more dispersed over the day than is the situation today. The positive consequence can be a reduction in the rush-hour traffic; the negative is that it is more difficult to offer a high frequent public transport service when travel needs are more spread in time. Ownership and use of both mobile and stationary technologies are unequally distributed. Men, people with high education and income are the most frequent owners and users. Ó 2002 Elsevier Science Ltd. All rights reserved. Keywords: Home computer; Car-use; Travel pattern; Survey data; Norway
1. Introduction Both transport technology and information and communication technology (ICT) have ‘‘made the world smaller’’ at least for some groups in society. The concept time-space compression
*
Tel.: +47-22-57-38-00; fax: +47-22-57-02-90. E-mail address: [email protected] (R.J. Hjorthol).
0965-8564/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 5 - 8 5 6 4 ( 0 1 ) 0 0 0 1 2 - X
438
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
(Harvey, 1989) is descriptive of this phenomenon. In principle, these two forms of technology in combination make it possible to organise everyday life and employment in new ways in relation to time and space. In principle people can work at home, groceries can be ordered on the net, information needed can be loaded down from the net, people can be entertained in different ways and there is no need to visit banks or travel bureaus as long as they are connected to Internet. Organising everyday life in this way make technological determinists or optimists think that problems related to rush-hour traffic and environmental problems with increasing car traffic can be significantly reduced. The question is whether more use of new information – and telecommunication technology will change the organisation of everyday life, and if so, will this change have an impact on the amount of travel and car-use? Within the field of transport, the discussion of substitution of travel by electronic communication has been going on for more than 20 years. The energy crisis at the beginning of the 1970s was the start of it all (Mokhtarian, 1990). One of the first studies on telecommuting takes this as its point of departure (Nilles et al., 1976). In the debate on how to reduce environmental problems generated by road traffic, and secure a more sustainable development, great hope has been placed on stationary means of communication bringing about reduced daily travel by car (Batten, 1989; Capello and Gillespie, 1993; Engstr€ om and Johanson, 1996). In debates on planning policy, replacing travelling to work with telecommuting (e.g., working remotely with the help of a computer either from the home of the employed or from neighbourhood centres) is gaining ground. Most of the debate has been concentrated on the work trip, other types of travel activities have to a very little degree been discussed. Based on a paradigm of modernity, the belief is that new and modern technology, partly or in full, can take over for the old, e.g. that use of information and telecommunication technology to some degree will replace physical travel, the mobile technology. In principle, however, there are four ways that new technology can interact with old: 1. Substitution or replacement. New technology (ICT) replace old (transport/travel purpose) without any effects on other parts of the travel pattern or for other household members. 2. Modification. New technology is used to conduct or change planned activities. 3. Generation. New technology gives more information, new acquaintances and possibilities which induce more travel. 4. Addition. New technology comes in addition to old, and there is no specific relation between them. In addition to these four ‘‘ideal-types’’ there can also be various combinations between them. Research on the substitution of travel has been concentrated on telecommuting and often in pilot and demonstration projects (Nilles, 1991; Hamer et al., 1991; Henderson and Mokhtarian, 1996; Balepur et al., 1998). The results are not unambiguous. Although some projects have not resulted in reduced travel in total, car-use was reduced on work trips (Nilles, 1991). In others, an increase in car-use was found due to more travel outside rush-time hours, but a reduction in the total length of car travel (Balepur et al., 1998). An experiment with 30 employees at the Ministry of Transport in The Netherlands resulted in a reduction in daily trips for telecommuters (Hamer et al., 1991). In a review of eight telecommuting programs, Mokhtarian et al. (1995) claim that the
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
439
effect of telecommuting has to be analysed in relation to the total amount of daily travel, i.e. not just the journey to work. For instance, when travelling to work is eliminated, efficient travel chains (e.g., combination of different purposes) can be broken and new patterns established which are not so efficient. This might also change the travel patterns in the family or the household. What these researchers also found important was that the first telecommuters were different from employees in general. They had further to travel to work than employees on average, and for them the effect of telecommuting was greater than for people with shorter distances to work. Mokhtarian’s (1998) conclusion based on the state-of-the-art of the relation between telecommuting and travel activity is that one cannot expect any significant reduction in travel activity by substituting telecommuting for work trips. She believes that information and communication technology will result in more flexibility in relation to everyday travel. A reduction in time travelling to work may, for instance, lead to more leisure travelling or shopping trips. Reduced car-use for one member of the family can lead to increased use for another. In the long term, telecommuting and use of information and communication technology for organising everyday activities can have an impact on land-use. For example, a reduction in the number of trips to work per week can make acceptance of a long journey to work more palatable, and people may buy houses in more distant (and attractive) areas where prices are lower than in more central areas in towns and cities. The purpose of this paper is to present an analysis of the relation between use of the home computer for various purposes (the stationary communication technology) and daily travel activity in general (the mobile technology – mostly the car) and discuss what kind of interaction there is between the private ownership and use of stationary and mobile technology, represented by the home computer and the car. The questions addressed are related to the four principles outlined. In this context substitution, generation or addition are most relevant in the discussion of the results. The home computer will have different areas of application depending on the equipment. The simplest form – a computer with no Internet connection – can be used for writing, for carrying out numerical calculations, for entertainment and for processing and storing information. A CDROM player increases the value of information-seeking and entertainment. A computer connected to the Internet allows communication with the ‘‘surrounding world’’. This can be divided according to activity or purpose. Private use might be social contacts with friends, ordering goods or services, obtaining different forms of information, and so on. Use in employment/paid work can also take different forms, e.g. writing, making calculations, information-seeking, attending discussion groups, etc. The use of a car or other means of transportation can be related to activities in analogous categories, and the trip will be the analytical category in this context. The work trip is related to production, as is the business trip, which can be compared to information-seeking and communication with colleagues or discussion groups on the Internet. Shopping for groceries belongs to the private sphere and shopping trips are analogues to electronic shopping. Leisure trips are analogues to the use of the home computer for entertainment. The paper is organised in six parts. Following the Section 1 and a section on data and methodology, analyses of the ownership of both stationary and mobile technology equipments are presented, before taking a closer look at the use of the technology among different groups in Section 4. In Section 5, the results of the analysis about the relationship between the possibility for
440
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
Table 1 The distribution of sociodemographical variables for different groups (%) Variables
Group 1
Group 2
Group 3
Without PC
Have PC at home have not answered the questionnaire
Have PC at home have answered the questionnaire
Two cars in the household
Gender Men Women
45 55
49 51
55 45
53 48
Age (years) 13–17 18–24 25–34 35–44 45–54 55–66 67+
4 6 20 15 16 18 22
12 11 21 24 20 10 3
10 9 24 27 21 8 2
9 7 20 27 2 12 3
Place of living Four largest cities Smaller cities and towns The rest of the country
43 28 29
52 23 25
49 25 26
30 46
20 40
15 37
20 43
12 11 1
18 20 2
19 27 2
16 19 2
Household income (NOK 1000) <100 100–199 200–299 300–399 400–499 500+
9 18 25 18 15 14
2 4 15 18 20 41
2 4 13 19 22 41
1 2 9 18 23 47
Socioeconomic status Manual workers Non-manual workers Professionals, high level white collar Owners Students Others
33 14 8 4 5 35
35 22 14 6 13 10
32 25 18 5 11 9
36 21 16 7 10 11
Education Elementary school High school/college University Lower level Higher level Unknown
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
441
Table 1 (continued) Variables
Group 1
Group 2
Group 3
Without PC
Have PC at home Have PC at home have answered have not answered the ques- the questionnaire tionnaire
Number of cars in the household None 1 2 3+
22 56 19 3
8 50 36 6
7 53 36 5
Weekly working hours 0–20 21–36 37–40 41þ
12 17 52 20
14 16 46 24
14 14 50 22
4321
1988
2529
Total
Two cars in the household
‘‘home work’’ and the amount of travel activity and travel patterns are presented along with regression analyses of the effect of home computer ownership on travel activity. Finally, in Section 6, future development is discussed on the basis of the results.
2. Data and methodology The empirical analyses are based on two interrelated data sets; the Norwegian national personal travel survey (NPTS) from 1997/98 and a corresponding survey about the use of information and communication technology at home. The data sets are combined – the respondents are the same. The NPTS consists of a random sample of about 8800 people, 13 years of age or above. They were interviewed by telephone about their daily travel activities, household characteristics and personal information as education, income, transport resources and employment. At the end of the interview the interviewee was asked if the household owned a computer, and, if so, whether s/he would agree to fill out a questionnaire about the respondent’s use of information and telecommunication equipment at home. Of the total sample of NPTS 51% (4500) said they owned a home computer; 3400 agreed to fill out the questionnaire and, of these, 81% actually returned it. With some reduction because incomplete questionnaires, a net sample of about 2500 persons was obtained, 56% of those who owned a home computer. Table 1 shows three different groups of respondents distributed on variables that are of importance related to travel; Group 1, respondents without pc at home, Group 2, pc-owners who did not answer the questionnaire, Group 3, pc-owners who answered the questionnaire. In addition
442
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
households with two or more cars are presented for the variables gender, age, education, household income and socioeconomic status. Group 1 differs from both groups 2 and 3. Compared to pc-owners it consists of more women, elderly people, fewer from the largest cities, more with low household income, more people outside the working force and they have less cars. The differences between 2 and 3 are minor. There are some more men and people with high education level in group 3 compared to group 2, but there are no significant differences along variables like age, place of living, socioeconomic status, number of cars and weekly working hours. This means that for very important variables self-selection and sample bias of group 3 compared to group 2 seems to be a very little problem. These two sets of data make it possible to analyse the relation between stationary and mobile communication on a general level. The personal travel survey comprises of the following subjects: introduction, access to transport resources for the interviewed person and for the household, activities and travel the day before the interview (purpose, length, time-use, transport mode, when and where the trip started and ended), long trips (100 km or longer during the previous month), employment/occupational status, the journey to work, education and employment of the spouse, information about the household and the interviewee. Information about the long trips will not be used in the analyses presented in this paper. The information and telecommunication survey consists of six main subjects: type and number of computers at home, other types of information and telecommunication devices, use of computers at home, use of computers for out of home communication (both privately and related to paid work), use of other telecommunication equipment (telephone and telex) and questions about home-based paid work. Only data about ownership and use of home computers will be used. Additive indexes of the use of home computers are constructed on the basis of a number of different private and work-related tasks carried out during a week. Both surveys were conducted over an entire year, September 1997–September 1998. Information about the use of home computers is limited to the week before the respondents answered the questionnaire, the travel activity to one day about a week before (when the telephone interview was carried out). The consequence of this is that travel activities and use of the home computer do not cover the same days. The results therefore are interpreted in a general sense.
3. The ownership of computers and cars Even though cars and computers today are commonly found among most social groups in the population, there is a correlation with income (see also Table 1). A Norwegian study on the development of use of different media shows that home computers are more common among highincome households than among low-income households and more common among men than women (Vaage, 1998). A Swedish study on information technology and transportation also shows a relation between access to computers at home and income, but contrary to the Norwegian results a very high proportion among the lowest income group have access to a home computer (SIKA, 1998). The explanation is that students and young people not living at their parents home very often have a home computer but a low income. What both the Swedish and the Norwegian studies show is that elderly people very seldom have access to home computers. Surveys of both
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
443
Table 2 The probability of having an Internet connection among those who have a computer (logistic regression)a Variables
B
S.E.
Sig.
Gender (male ¼ 1, female ¼ 0) Age Household income Education, < ¼ 9 yearsb Education, < ¼ 12 yearsb Education, < ¼ 15 yearsb Manual workerc Non-manual workerc High level white collar workers, profes.c Owner of business, farmer, etc.c Student, pupilc Constant
0.3079 )0.0203 0.2187 )0.1561 )0.2616 )0.3830 )0.1667 )0.1387 )0.2022 )0.0337 )0.3194 )0.1908
0.0911 0.0041 0.0351 0.1729 0.1194 0.1244 0.1780 0.1771 0.1882 0.2494 0.2615 0.2987
0.0007 0.0000 0.0000 0.3667 0.0284 0.0021 0.3491 0.4337 0.2825 0.8924 0.2218 0.5229
N ¼ 2190; )2 log likelihood ¼ 2940,080. Reference categories: Education > ¼ 16 years. c Housewives, unemployed and others not in the work force. a
b
private consumption and personal travel show that high-income groups have better access to private cars than low-income groups (SSB, 1993; Hjorthol, 1998). As shown in Table 1 only 7–8% of pc-owners do not have a car in the household, while 22% of non-owners do not have a car. More than 40% of pc-owners have two or more cars compared to 22% of those who do not own a pc. Both the car and the home computer are technical devices that are part of the same field of consumption. They are both individual equipment more typical of men, young middle-aged people, and those with high income. These variables are also important for the quality pc-owners have on their equipment. The probability of having an Internet connection among computer owners is strongly related to both income and age (Table 2). Gender, too, has a significant impact. Men more often than women have a computer with an Internet connection. The analysis also shows that people with middle-level education less often have an Internet connection than people with a high-level university education. The reason that there is no significant difference between people with lowest and highest education is explained by age. More than 50% of the respondents on the lowest educational level are under 18 years and still living with their parents.
4. The use of computers In the postal survey questions asked about the use of the computer are divided into four different categories. (a) Private tasks 1 without need of an Internet connection (writing documents, letters, etc., numeric calculation, graphics, illustration, etc., games and other tasks).
444
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
(b) Private tasks 2 with need of an Internet connection (send/receive e-mail from friends and relatives, send/receive e-mail from others/attend discussion groups, use the Internet for information, etc., order goods/services, send/receive documents, and other tasks). (c) Work tasks 1 without need of an Internet connection (writing documents, letters, etc., numeric calculation, graphics, illustration, etc. and other tasks). (d) Work tasks 2 with need of an Internet connection (send/receive e-mail from others/attend discussion groups, use the Internet for information, etc., order goods/services, send/receive documents, and other tasks). Additive indexes are constructed on the basis of these four groups of tasks. The respondents were asked to report how many times they carried out the different tasks during the previous week. If they had not used the computer they were asked to indicate as to how long time it was since they had used it. In the analyses presented in this section only respondents with an Internet connection are included.
4.1. Private use of the home computer The use of computers without need for an Internet connection for private tasks (private tasks 1) is perhaps the kind of use least to do with, with travel. Some of it, like games, might substitute for leisure activities and travel, but there is no necessary connection between the four types of private use recorded here and travel. This is better taken as an indication of how common this kind of technology is in people’s everyday lives. It tells about the tendencies of ‘‘habituation’’ of this type of technology, and is presented for that reason. Private tasks (private tasks 2) which need an Internet connection are sending/receiving e-mail to/from friends, relatives or others, searching for information, news, etc., ordering goods or services, sending/receiving documents, and other tasks. Some of these tasks can be substituted by travel; for example, ordering goods instead of shopping, ordering tickets directly on the Internet from an airline instead of visiting a travel bureau. But ordinary mail and use of the telephone are alternatives to both travel and use of the Internet for many of these tasks. However, some of these would probably not have been done had access to the Internet not been possible. On average, people with an Internet connection send/receive mail from friends and relatives 2.4 times per week; they send/receive mail from others and attend discussion groups 1.8 times per week, they search for information and news 3.3 times per week, order goods and services 0.1 time per week, send/receive documents 1.0 and carry out other tasks 1.5 times per week. Information-seeking therefore represents 33% of these tasks and sending/receiving mail more than 40%, so the substitutionary potential of telecommunication for travel is small for these categories of private tasks. The results for both indexes, categorised as private tasks 1 and private tasks 2, are given in Table 3. Men use the computer for private purposes more than women do, and young people more than the elderly. Those with low education and not gainfully employed use the home computer more
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
445
Table 3 Average use of the home computer for private tasks 1 (the sum of writing documents, letters, etc., numerical tasks, graphics and illustrations, games and other tasks) and private tasks 2 (the sum of sending/receiving E-mail, attending discussion groups, search for information and news, ordering goods/services, sending/receiving documents and other tasks)a Variables
Private tasks 1 times per week
Number of respondents
Gender Male Female
11.0 7.0
743 513
Age (years) 13–17 18–24 25–34 35–44 45–54 55–66 67+
15.8 10.9 9.4 7.1 8.8 7.5 7.8
138 103 329 347 248 72 19
12.7
Education Compulsory school (9 years) Education up to 12 years University level Low grade High grade
a
Private tasks 2 times per week
Number of respondents
0.000 668 456
11.0 11.4 10.2 6.6 7.3 4.5 4.5
115 96 305 312 218 62 16
207
9.8
168
9.7
436
8.9
398
7.6 7.9
217 376
6.9 7.8
200 343
0.000
0.001
0.000
9.3
Sig. (F-test) 0.000
10.3 5.5
Employment Employed > ¼ 40 h per week 8.5 30–40 h per week 8.3 <30 h per week 10.5 Not employed 12.0 Total
Sig. (F-test)
0.260
0.001
0.964
267 570 180 236
8.8 8.2 8.6 8.2
242 516 158 205
1253
8.4
1121
Numbers per week for respondents with Internet.
for writing documents, correspondence, playing games, etc. (private tasks 1) than those with high education and people within the work force. Education and employment are strongly related to age, and this is the main explanation behind the frequent use among these groups. Young men are the most frequent users of the computer for these purposes, and playing games is the most decisive activity. The differences between men and women are most distinct among the youngest (13–17 years). Boys have a weekly use frequency of 16.1; girls of 6.5. In the age group 18–24 years, the frequency of use for boys is 10.4 and 5.1 for girls. Among other age groups there are no significant differences between the genders. These differences between young men and women have also been found in other studies (Telenor, 1999). Use of the home computer for private purposes bears little relation to mobility. Some of these activities can be considered analogous to trips related to various leisure activities and shopping.
446
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
Table 4 Correlation between indexes for private purposes 1 (N ¼ 1253) and private purposes 2 (N ¼ 1121) (see Table 2) – and travel activities (Pearsons r) Trips related to visits Index – use of computer for private purposes 1 Index – use of computer for private purposes 2 *
0.062 0.051
Trips related to shopping
Trips per day in total
Trips as a car driver
km as a car driver per day
0.006
)0.045
)0.009
)0.047
)0.020
)0.042
)0.042
)0.016
)0.027
0.008
Trips related to leisure
Correlation is significant at the 0.05 level (2-tailed).
The correlation 1 analysis, however, shows no significant negative relations between this type of use of the home computer and travelling, as one would have expected if stationary communication was substituted for travelling (Table 4). The activities (especially playing games) are additional to other leisure activities out of home. The correlation analysis shows that there is no relation between car-use, measured as trips as a driver and distance in kilometres per day, number of trips for different purposes, trips as a car driver and use of the computer for private purposes. There is a very small positive relation between use of the computer for private tasks 1 and number of trips related to private visits of friends and relatives. An interpretation could be that there is a slight tendency that high frequency on one activity is correlated with high activity on the other. These results show that use of stationary technology for private purposes is additional to travelling and as such not related activities. 4.2. Use of the home computer for paid work The most relevant tasks related to substitution of travel are to work at home instead of travelling to place of work. As discussed in the Introduction, there is considerable interest in ICT for a possible reduction of work trips as a means of reducing the environmental problems related to car traffic. The most significant differences in application of the home computer for work purposes are found between men and women and between those with long weekly working hours and those with part-time work or those with ‘‘normal’’ working hours (Table 5). Men’s use is about twice that of women’s use, and people who work 40 h or more per week also use it twice as much as people with shorter working hours. Men are in the majority when it comes to working long hours, so those characteristics reinforce each other. The result in Table 6 indicates that work at home for these men comes in addition to their regular work at the workplace. 1
The correlation analyses can only be taken as indications because of the differences in periods for travel activities and use of the home computer (see Section 2).
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
447
Table 5 Average use of the home computer for paid work 1 (the sum of writing documents, letters etc., numerical tasks, graphics and illustrations, and other tasks) and paid work 2 (the sum of sending/receiving E-mail, searching for information/ news, ordering goods/service, sending/receiving documents, and other tasks)a Variables
Gender Male Female Education Compulsory school (9 years) Education up to 12 years University level Lower grade High grade
Paid work 1 times per week
Number of respondents
5.4 3.3
604 364
3.5
Paid work 2 times per week
Number of respondents
5.2 2.5
508 303
97
4.5
73
4.1
339
3.6
279
4.7 5.4
189 341
3.1 5.3
161 295
0.003
258 545 165
Occupational status Manual worker White collar worker Professional Owner
3.4 4.5 5.7 7.7 4.6
a
Sig. (F-test) 0.007
0.004
Employment Employed > ¼ 40 h per week 6.9 30–40 h per week 3.6 <30 h per week 4.1
Total (occupational status)
Sig. (F-test)
0.396
0.000
0.000 6.7 3.5 2.3
217 456 138
344 319 235 58
2.8 4.7 5.1 6.4
300 258 199 43
956
4.0
800
0.033
0.330
Times per week for respondent with paid work and Internet.
Table 6 Correlation between mobile communication (number of work trips, business trips, kilometres as a car driver per day, number of car trips as a driver per day and total trips per day) and stationary communication (index for use work purposes 1 (N ¼ 968) and work purposes 2 (N ¼ 811))a
Index – use of computer for work purposes 1 Index – use of computer for work purposes 2 a
Work trips
Business trips
km as a driver per day
Trips as a driver per day
Total trips per day
0.082
0.103
0.092
0.119
0.081
0.063
0.062
0.073
0.056
0.014
Respondents in paid work with Internet. Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). *
448
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
Table 7 Linear regression of daily travel length by car in km as a driver for gainfully employed respondents with drivers licence and Internet, N ¼ 786a Variables Work purposes 1 Region (1 ¼ four biggest cities, 0 ¼ else) Gender (1 ¼ male, 0 ¼ female) Household income Number of cars Age Constant a
B
Beta
t-test
Sig.
0.344 4.803
0.079 0.050
2.239 1.414
0.025 0.168
13.949 )0.009 7.292 0.266 )7.955
0.139 )0.007 0.108 0.053
3.939 )0.197 3.014 1.493 )0.977
0.000 0.844 0.003 0.136 0.329
R2 ¼ 0:045; adjusted R2 ¼ 0:037.
As the correlation 2 analysis shows, there is no substitution of mobile communication, travelling, by use of stationary communication devices. It is rather a tendency of the opposite, especially in relation to work purposes 1. People who use the computer at home in relation to their paid work have some more and longer trips than people who do not use the computer for such tasks. However, the correlation coefficients are small, so it is difficult to give an absolute conclusion. But it seems that rather than being a substitute for mobile communication, stationary communication can be additional when the computer is used for tasks related to paid work. The correlation between the index for work purpose 2 and number of kilometres by car as a driver also shows a small positive relation. Those who use the home computer for work-related tasks tend to drive longer distances (with themselves as driver) that those who do not use it. A primary conclusion is that there is no substitution of travel by use of communication technology among those who own and use a home computer(s). To control for the effect of other variables on car-use a regression analysis of total daily travel length by car as a driver is done. The independent variables are place of living, gender, household income, age and number of cars in the household in addition to the use of the home computer for work purposes 1. For the analyses we have selected respondents with Internet and drivers licence. Table 7 shows a significant positive relation between the amount of travel (km per day as a driver) and use of the home computer for work purposes 1. This analysis reinforces the conclusions of the correlation analyses. There is a tendency that people who use the home computer frequently for work purposes also are frequent users of the car, even though the relation is not very strong. The analysis presented in Table 7 shows however, that gender and number of cars in the household are more important for car-use than use of the home computer. 5. The possibility to work at home and the amount of travel activity The analyses presented in Section 4 include only pc owners. In this section we therefore want to compare travel activity for different groups in relation to the ownership of home computers and the possibility to work at home. 2
See Footnote 1.
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
449
In the questionnaire the respondents (with home computers) were asked if they had the possibility to work at home and, if so, how many times (whole days or part of a day) they had done so during the previous week. Of the sample, 30% said they can work at home. 6% have a permanent workplace at home, which means that they work mostly at home, and 64% cannot work at home. In Table 8 the daily travel patterns of employed groups with an annual income of NOK 200.000 or higher (to get the groups more homogeneous) with different access to home computers and possibilities for working at home are compared. People without home computers represent one category. (Not all trip purposes are presented in the table, so the total number of trips is higher than the sum of those presented in the table). The results indicate that variations in travel patterns are fairly small. There is no significant difference in the total number of kilometres driven by car per day, even if there is a tendency for people who work mostly at home to travel a little shorter by car as a driver. The same group has fewer work trips, but the total number of trips per day is the same as for the other groups with a home computer who do not work permanently at home, and even more than those without a home computer. This supports the hypothesis that a reduction in one type of journey or trip will be replaced by other kinds; for instance, work trips can be replaced by trips related to leisure or shopping. In this case we see that those who work at home have more chauffeuring trips than those without home computers, which could indicate that working at home might be an adjustment to a family situation with children. To control for the effects of the differences between these groups and of other variables which have an impact on travel in addition to pc ownership regression analyses are done for work trips, total number of trips and trips related to chauffeuring. Since only pc-owners got the question about the option to work at home, it is not possible to include this aspect in the analysis. To make
Table 8 Number of trips per day for different purposes (not all purposes are shown) for employed people with different access to a home computer and possibilities for work at homea;b Different trips
Without home computers
(1) Trips related to visits per day Trips related to leisure per day Trips related to shopping per day Trips related to chauffeuring per day Work trips per day Business trips per day Number of trips per day Number of trips as car driver per day km as a car driver per day Total a b
0.37 0.46 0.71 0.25 1.03 0.15 3.26 2.07 29.2 1454
(3.4) (3) (2.3.4) (4) (3.4) (4)
With home computer, cannot work at home (2) 0.30 0.45 0.79 0.35 0.91 (4) 0.13 (3.4) 3.45 2.10 30.6 820
With home computer, can work at home (3) 0.26 0.60 0.67 0.31 0.97 (4) 0.27 3.60 2.22 28.8 444
Employed with own income NOK 200.000 or more per year. Numbers per day. (x) Significantly different from group x, P < 0:05.
With home computer, have permanent work site at home (4) 0.24 0.55 0.74 0.49 0.68 0.19 3.87 1.95 23.6 84
450
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
the analyses as comparable as possible with the analysis presented in Table 8, only respondents with a personal annual income of NOK 200.000 or more are included. Table 9 shows a linear regression analysis of the average number of work trips per day controlled for the effects of pc-ownership, residential region, gender, household income, number of cars and age. The analysis shows a small significant effect of pc-ownership on the number of work trips per day. Respondents with a pc have in average less work trips per day than people without. The analysis of the total number of trips per day in Table 10 shows that there are no significant difference in the total number of daily trips between pc-owners and non-owners. The reduction in work trips is ‘‘compensated’’ for by more other types of trips. Gender, income, number of cars and age are the important variables to explain the level of daily travel activity. Table 11 shows the results of the analysis of trips related to chauffeuring children, and we see that pc owners have more such trips than non-owners. These three regression analyses give an indication of a certain reduction of work trips for pcowners compared with non-owners. However, there is no difference in total travel activity between the groups, which means more trips for other purposes. In this case we see that the respondents with home computers have more of their travel activity related to bringing children to different activities. As we saw in the analysis presented in Table 8, this might indicate an adjustment to a family situation. Table 9 Linear regression of number of work trips per day for gainfully employed respondents with income equal or greater than NOK 200.000, N ¼ 2870a Variables
B
Beta
t-test
Sig.
PC-owners (1 ¼ yes, 0 ¼ no) Region (1 ¼ four biggest cities, 0 ¼ else) Gender (1 ¼ male, 0 ¼ female) Household income Number of cars Age Constant
)0.077 0.053
)0.035 0.025
)2.126 1.532
0.034 0.126
0.094 0.000 0.036 )0.002 0.966
0.042 0.025 0.025 )0.021
2.591 1.516 1.495 )1.290 11.020
0.010 0.130 0.135 0.197 0.000
a
R2 ¼ 0:005; adjusted R2 ¼ 0:003.
Table 10 Linear regression of the total number of trips per day for gainfully employed respondents with income equal or greater than NOK 200.000, N ¼ 2870a Variables PC-owners (1 ¼ yes, 0 ¼ no) Region (1 ¼ four biggest cities, 0 ¼ else) Gender (1 ¼ male, 0 ¼ female) Household income Number of cars Age Constant a
R2 ¼ 0:016; adjusted R2 ¼ 0:014.
B
Beta
t-test
Sig.
0.145 0.151
0.031 0.033
1.862 2.012
0.063 0.004
)0.004 0.000 0.192 )0.020 3.704
)0.001 0.033 0.062 )0.091
)0.051 1.962 3.652 )5.636 19.967
0.960 0.050 0.000 0.000 0.000
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
451
Table 11 Linear regression of number of chauffeuring trips per day for gainfully employed respondents with own income equal or greater than NOK 200.000, N ¼ 2870a Variables PC-owners (1 ¼ yes, 0 ¼ no) Region (1 ¼ four biggest cities, 0 ¼ else) Gender (1 ¼ male, 0 ¼ female) Household income Number of cars Age Constant a
B 0.072 0.0199 )0.092 0.001 0.062 )0.054 0.418
Beta
t-test
Sig.
0.040 0.028
2.430 0.698
0.015 0.485
)0.051 0.032 0.052 )0.066
)3.121 1.935 3.103 )4.061 5.925
0.002 0.053 0.002 0.000 0.000
R2 ¼ 0:013; adjusted R2 ¼ 0:012.
6. Discussion Considering the limitations related to the data sets, the results of the analyses show only to a very little extent any relation between ownership and use of a home computer and people’s daily travel patterns. On the basis of these results we cannot see any direct substitutionary effects of the use of stationary technology at people’s home on the use of mobile technology. Access to and use of information technology at home do not seem to have a significant impact on travel activities in everyday life. The analyses give some indications of adjusting work and family life, but the net effect gives no reduction in travel activity. Stationary communication seems to be a supplement to activities based on mobile technology, but it gives people more spatial and temporal options. For people who work more than normal weekly working hours, stationary technology seems to give them greater flexibility in regard to where and when to work, but it does not necessarily reduce their travel activity. Flexibility in everyday activities and travel is also what Mokhtarian (1998) suggests as an important impact of this ‘‘new’’ communication technology. The spatial flexibility will also give a temporal flexibility, which means that work trips and other trips can be more dispersed over the day than is the situation today. The positive consequence can be a reduction in the rush-hour traffic; the negative is that it is more difficult to offer a good public transport service when travel needs are more spread in time. The costs to offer a high frequency supply during all day could be high. What these analyses also reveal is a relationship between the ownership of cars and the ownership of computers, and, as such, to a high degree the same social groups who use the car and the computer. Men and high-income groups are more frequent users of the car than women and people with lower income. Men, high-income groups and people with high education more often own a home computer and have the possibility to work at home than people with low income and education. The possibility for flexibility related to the use of home computers and the new information and communication technology in general follow the existing divisions of different segments of the labour market. One crucial question is how the different segments will develop and their size. How many of the future workers of the labour stock will have the possibility to work at home, at neighbourhood centres, or to be mobile workers in general; how many want this type of flexibility and do companies want it?
452
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
If the possibility to flexibility regarding where and when to work continues to be a good for the privileged (high education/high income) in the future, it will be of great interest to know more about their preferences for living and also their daily activities in order to foresee the consequences for their travel patterns. In the long term changes in preferences related to the greater flexibility is possible, and can have impacts on urban development and land use patterns. Acknowledgements The research described in this paper was supported by the Strategic Institute Program of the Institute of Transport Economics. Thanks to Unni Lodden for her help in the preparation of the survey data and to the referees who provided many helpful comments. References Balepur, P.N., Varma, K.V., Mokhtarian, R.L., 1998. Transportation impacts of center-based telecommuting: interim findings from the neighborhood telecenters project. Transportation 25, 287–306. Batten, D., 1989. The future of transport and interface communication: debating the scope for substitution growth. In: Batten, D., Roland, T. (Eds.), Transportation for the Future. Springer, Berlin. Capello, R., Gillespie, A., 1993. Transport communications and spatial organisation: conceptual framework and future trends. In: Nijkamp, P. (Eds.), Europe on the Move. Aldershot mv, Avebury. Engstr€ om, M.G., Johanson, R., 1996. IT-utviklingens effekter p a framtida res- och transportstrukturer. (The effects of IT on transport and travel). Naturv ardverket F€ orlag, Stockholm. Hamer, R., Kroes, E., van Oostsroom, H., 1991. Teleworking in the Netherlands, an evaluation of changes in travel behaviour. Transportation 18, 365–382. Harvey, D., 1989. The Condition of Postmodernity. An Enquiry into the Origins of Cultural Change. Basil Blackwell, Oxford. Henderson, D.K., Mokhtarian, P.L., 1996. Impacts of center-based telecommuting on travel and emission: analysis of the pudget sound demonstration project. Transp. Res. D 1 (1), 29–45. Hjorthol, R., 1998. Hverdagslivets reiser. En analyse av kvinners og menns daglige reiser i Oslo (Travel activities in everyday life. An analysis of women’s and men’s daily travel activities in Oslo. Dissertation for dr. philos.). Insitutt for sosiologi og samfunnsgeografi, Universitetet i Oslo, Institute of Transport Economics, TØI Report 391/1998, Oslo. Mokhtarian, P.L., Handy, S., Salomon, I., 1995. Methodological issues in the estimation of the travel, energy and air quality impacts of telecommuting. Transp. Res. A. (4), 283–302. Mokhtarian, P.L., 1990. A typology of relationships between telecommunications and transportation. Transp. Res. A 24 (3), 231–242. Mokhtarian, P.L., 1998. A synthetic approach to estimating the impacts of telecommuting on travel. Urban Studies 35 (2), 214–241. Nilles, J.M., Carlson, F., Gray, P., Hanneman, G., 1976. The Telecommunications – Transportation Tradeoff. Wiley, New York. Nilles, J.M., 1991. Telecommunting and urban sprawl: mitigator or incsiter? Transportation 18, 411–432. SIKA, 1998. IT-utveklingen och transporterna 2. Redovisning av en kommunikationsunders€ okning 1997. (The development of IT and the transport. A study of communication from 1997) SIKA Rapport 1998:4, Stockholm. SSB, 1993. Forbrukerundersøkelsen 1989–1991 (Survey of consumer expenditure 1989–1991). Statistisk sentralbyr a. NOS C 65, Oslo, Kongsvinger. Telenor, 1999. Ungdom og internett. (Youth and Internet) Faglig nyhetsbrev fra Telenor FOU (1), January 1999, Oslo. Vaage, O.F., 1998. Norsk mediebarometer 1997. (Norwegian barometer of media, 1997) Statistisk sentralbyr a. Statistiske analyser 24, Oslo, Kongsvinger.
The relation between daily travel and use of the home computer Randi Johanne Hjorthol a
a,*
Institute of Transport Economics, Box 6110, Etterstad, N-0602 Oslo, Norway
Received 13 June 2000; received in revised form 26 January 2001; accepted 28 February 2001
Abstract On the basis of the Norwegian national personal travel survey (NPTS) 1997/98 and a connected mail back survey of the use of information – and communication technology at home, the relation between mobility and use of stationary communication has been studied. On the basis of these results we cannot see any direct substitutionary effects of the use of stationary technology at people’s home on the use of mobile technology. Access to and use of information technology seems not to have a significant impact on travel activities in everyday life. Stationary communication seems to be a supplement to activities based on mobile technology. For people who work more than ‘‘normal’’ weekly working hours, stationary technology seems to give them greater flexibility in regard to where to work, but it does not necessarily reduce their travel activity. There is a tendency that people who own home computers make less work trips, but this does not affect the total number of daily trips. The spatial flexibility give a temporal flexibility, which means that work trips and other trips can be more dispersed over the day than is the situation today. The positive consequence can be a reduction in the rush-hour traffic; the negative is that it is more difficult to offer a high frequent public transport service when travel needs are more spread in time. Ownership and use of both mobile and stationary technologies are unequally distributed. Men, people with high education and income are the most frequent owners and users. Ó 2002 Elsevier Science Ltd. All rights reserved. Keywords: Home computer; Car-use; Travel pattern; Survey data; Norway
1. Introduction Both transport technology and information and communication technology (ICT) have ‘‘made the world smaller’’ at least for some groups in society. The concept time-space compression
*
Tel.: +47-22-57-38-00; fax: +47-22-57-02-90. E-mail address: [email protected] (R.J. Hjorthol).
0965-8564/02/$ - see front matter Ó 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 9 6 5 - 8 5 6 4 ( 0 1 ) 0 0 0 1 2 - X
438
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
(Harvey, 1989) is descriptive of this phenomenon. In principle, these two forms of technology in combination make it possible to organise everyday life and employment in new ways in relation to time and space. In principle people can work at home, groceries can be ordered on the net, information needed can be loaded down from the net, people can be entertained in different ways and there is no need to visit banks or travel bureaus as long as they are connected to Internet. Organising everyday life in this way make technological determinists or optimists think that problems related to rush-hour traffic and environmental problems with increasing car traffic can be significantly reduced. The question is whether more use of new information – and telecommunication technology will change the organisation of everyday life, and if so, will this change have an impact on the amount of travel and car-use? Within the field of transport, the discussion of substitution of travel by electronic communication has been going on for more than 20 years. The energy crisis at the beginning of the 1970s was the start of it all (Mokhtarian, 1990). One of the first studies on telecommuting takes this as its point of departure (Nilles et al., 1976). In the debate on how to reduce environmental problems generated by road traffic, and secure a more sustainable development, great hope has been placed on stationary means of communication bringing about reduced daily travel by car (Batten, 1989; Capello and Gillespie, 1993; Engstr€ om and Johanson, 1996). In debates on planning policy, replacing travelling to work with telecommuting (e.g., working remotely with the help of a computer either from the home of the employed or from neighbourhood centres) is gaining ground. Most of the debate has been concentrated on the work trip, other types of travel activities have to a very little degree been discussed. Based on a paradigm of modernity, the belief is that new and modern technology, partly or in full, can take over for the old, e.g. that use of information and telecommunication technology to some degree will replace physical travel, the mobile technology. In principle, however, there are four ways that new technology can interact with old: 1. Substitution or replacement. New technology (ICT) replace old (transport/travel purpose) without any effects on other parts of the travel pattern or for other household members. 2. Modification. New technology is used to conduct or change planned activities. 3. Generation. New technology gives more information, new acquaintances and possibilities which induce more travel. 4. Addition. New technology comes in addition to old, and there is no specific relation between them. In addition to these four ‘‘ideal-types’’ there can also be various combinations between them. Research on the substitution of travel has been concentrated on telecommuting and often in pilot and demonstration projects (Nilles, 1991; Hamer et al., 1991; Henderson and Mokhtarian, 1996; Balepur et al., 1998). The results are not unambiguous. Although some projects have not resulted in reduced travel in total, car-use was reduced on work trips (Nilles, 1991). In others, an increase in car-use was found due to more travel outside rush-time hours, but a reduction in the total length of car travel (Balepur et al., 1998). An experiment with 30 employees at the Ministry of Transport in The Netherlands resulted in a reduction in daily trips for telecommuters (Hamer et al., 1991). In a review of eight telecommuting programs, Mokhtarian et al. (1995) claim that the
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
439
effect of telecommuting has to be analysed in relation to the total amount of daily travel, i.e. not just the journey to work. For instance, when travelling to work is eliminated, efficient travel chains (e.g., combination of different purposes) can be broken and new patterns established which are not so efficient. This might also change the travel patterns in the family or the household. What these researchers also found important was that the first telecommuters were different from employees in general. They had further to travel to work than employees on average, and for them the effect of telecommuting was greater than for people with shorter distances to work. Mokhtarian’s (1998) conclusion based on the state-of-the-art of the relation between telecommuting and travel activity is that one cannot expect any significant reduction in travel activity by substituting telecommuting for work trips. She believes that information and communication technology will result in more flexibility in relation to everyday travel. A reduction in time travelling to work may, for instance, lead to more leisure travelling or shopping trips. Reduced car-use for one member of the family can lead to increased use for another. In the long term, telecommuting and use of information and communication technology for organising everyday activities can have an impact on land-use. For example, a reduction in the number of trips to work per week can make acceptance of a long journey to work more palatable, and people may buy houses in more distant (and attractive) areas where prices are lower than in more central areas in towns and cities. The purpose of this paper is to present an analysis of the relation between use of the home computer for various purposes (the stationary communication technology) and daily travel activity in general (the mobile technology – mostly the car) and discuss what kind of interaction there is between the private ownership and use of stationary and mobile technology, represented by the home computer and the car. The questions addressed are related to the four principles outlined. In this context substitution, generation or addition are most relevant in the discussion of the results. The home computer will have different areas of application depending on the equipment. The simplest form – a computer with no Internet connection – can be used for writing, for carrying out numerical calculations, for entertainment and for processing and storing information. A CDROM player increases the value of information-seeking and entertainment. A computer connected to the Internet allows communication with the ‘‘surrounding world’’. This can be divided according to activity or purpose. Private use might be social contacts with friends, ordering goods or services, obtaining different forms of information, and so on. Use in employment/paid work can also take different forms, e.g. writing, making calculations, information-seeking, attending discussion groups, etc. The use of a car or other means of transportation can be related to activities in analogous categories, and the trip will be the analytical category in this context. The work trip is related to production, as is the business trip, which can be compared to information-seeking and communication with colleagues or discussion groups on the Internet. Shopping for groceries belongs to the private sphere and shopping trips are analogues to electronic shopping. Leisure trips are analogues to the use of the home computer for entertainment. The paper is organised in six parts. Following the Section 1 and a section on data and methodology, analyses of the ownership of both stationary and mobile technology equipments are presented, before taking a closer look at the use of the technology among different groups in Section 4. In Section 5, the results of the analysis about the relationship between the possibility for
440
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
Table 1 The distribution of sociodemographical variables for different groups (%) Variables
Group 1
Group 2
Group 3
Without PC
Have PC at home have not answered the questionnaire
Have PC at home have answered the questionnaire
Two cars in the household
Gender Men Women
45 55
49 51
55 45
53 48
Age (years) 13–17 18–24 25–34 35–44 45–54 55–66 67+
4 6 20 15 16 18 22
12 11 21 24 20 10 3
10 9 24 27 21 8 2
9 7 20 27 2 12 3
Place of living Four largest cities Smaller cities and towns The rest of the country
43 28 29
52 23 25
49 25 26
30 46
20 40
15 37
20 43
12 11 1
18 20 2
19 27 2
16 19 2
Household income (NOK 1000) <100 100–199 200–299 300–399 400–499 500+
9 18 25 18 15 14
2 4 15 18 20 41
2 4 13 19 22 41
1 2 9 18 23 47
Socioeconomic status Manual workers Non-manual workers Professionals, high level white collar Owners Students Others
33 14 8 4 5 35
35 22 14 6 13 10
32 25 18 5 11 9
36 21 16 7 10 11
Education Elementary school High school/college University Lower level Higher level Unknown
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
441
Table 1 (continued) Variables
Group 1
Group 2
Group 3
Without PC
Have PC at home Have PC at home have answered have not answered the ques- the questionnaire tionnaire
Number of cars in the household None 1 2 3+
22 56 19 3
8 50 36 6
7 53 36 5
Weekly working hours 0–20 21–36 37–40 41þ
12 17 52 20
14 16 46 24
14 14 50 22
4321
1988
2529
Total
Two cars in the household
‘‘home work’’ and the amount of travel activity and travel patterns are presented along with regression analyses of the effect of home computer ownership on travel activity. Finally, in Section 6, future development is discussed on the basis of the results.
2. Data and methodology The empirical analyses are based on two interrelated data sets; the Norwegian national personal travel survey (NPTS) from 1997/98 and a corresponding survey about the use of information and communication technology at home. The data sets are combined – the respondents are the same. The NPTS consists of a random sample of about 8800 people, 13 years of age or above. They were interviewed by telephone about their daily travel activities, household characteristics and personal information as education, income, transport resources and employment. At the end of the interview the interviewee was asked if the household owned a computer, and, if so, whether s/he would agree to fill out a questionnaire about the respondent’s use of information and telecommunication equipment at home. Of the total sample of NPTS 51% (4500) said they owned a home computer; 3400 agreed to fill out the questionnaire and, of these, 81% actually returned it. With some reduction because incomplete questionnaires, a net sample of about 2500 persons was obtained, 56% of those who owned a home computer. Table 1 shows three different groups of respondents distributed on variables that are of importance related to travel; Group 1, respondents without pc at home, Group 2, pc-owners who did not answer the questionnaire, Group 3, pc-owners who answered the questionnaire. In addition
442
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
households with two or more cars are presented for the variables gender, age, education, household income and socioeconomic status. Group 1 differs from both groups 2 and 3. Compared to pc-owners it consists of more women, elderly people, fewer from the largest cities, more with low household income, more people outside the working force and they have less cars. The differences between 2 and 3 are minor. There are some more men and people with high education level in group 3 compared to group 2, but there are no significant differences along variables like age, place of living, socioeconomic status, number of cars and weekly working hours. This means that for very important variables self-selection and sample bias of group 3 compared to group 2 seems to be a very little problem. These two sets of data make it possible to analyse the relation between stationary and mobile communication on a general level. The personal travel survey comprises of the following subjects: introduction, access to transport resources for the interviewed person and for the household, activities and travel the day before the interview (purpose, length, time-use, transport mode, when and where the trip started and ended), long trips (100 km or longer during the previous month), employment/occupational status, the journey to work, education and employment of the spouse, information about the household and the interviewee. Information about the long trips will not be used in the analyses presented in this paper. The information and telecommunication survey consists of six main subjects: type and number of computers at home, other types of information and telecommunication devices, use of computers at home, use of computers for out of home communication (both privately and related to paid work), use of other telecommunication equipment (telephone and telex) and questions about home-based paid work. Only data about ownership and use of home computers will be used. Additive indexes of the use of home computers are constructed on the basis of a number of different private and work-related tasks carried out during a week. Both surveys were conducted over an entire year, September 1997–September 1998. Information about the use of home computers is limited to the week before the respondents answered the questionnaire, the travel activity to one day about a week before (when the telephone interview was carried out). The consequence of this is that travel activities and use of the home computer do not cover the same days. The results therefore are interpreted in a general sense.
3. The ownership of computers and cars Even though cars and computers today are commonly found among most social groups in the population, there is a correlation with income (see also Table 1). A Norwegian study on the development of use of different media shows that home computers are more common among highincome households than among low-income households and more common among men than women (Vaage, 1998). A Swedish study on information technology and transportation also shows a relation between access to computers at home and income, but contrary to the Norwegian results a very high proportion among the lowest income group have access to a home computer (SIKA, 1998). The explanation is that students and young people not living at their parents home very often have a home computer but a low income. What both the Swedish and the Norwegian studies show is that elderly people very seldom have access to home computers. Surveys of both
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
443
Table 2 The probability of having an Internet connection among those who have a computer (logistic regression)a Variables
B
S.E.
Sig.
Gender (male ¼ 1, female ¼ 0) Age Household income Education, < ¼ 9 yearsb Education, < ¼ 12 yearsb Education, < ¼ 15 yearsb Manual workerc Non-manual workerc High level white collar workers, profes.c Owner of business, farmer, etc.c Student, pupilc Constant
0.3079 )0.0203 0.2187 )0.1561 )0.2616 )0.3830 )0.1667 )0.1387 )0.2022 )0.0337 )0.3194 )0.1908
0.0911 0.0041 0.0351 0.1729 0.1194 0.1244 0.1780 0.1771 0.1882 0.2494 0.2615 0.2987
0.0007 0.0000 0.0000 0.3667 0.0284 0.0021 0.3491 0.4337 0.2825 0.8924 0.2218 0.5229
N ¼ 2190; )2 log likelihood ¼ 2940,080. Reference categories: Education > ¼ 16 years. c Housewives, unemployed and others not in the work force. a
b
private consumption and personal travel show that high-income groups have better access to private cars than low-income groups (SSB, 1993; Hjorthol, 1998). As shown in Table 1 only 7–8% of pc-owners do not have a car in the household, while 22% of non-owners do not have a car. More than 40% of pc-owners have two or more cars compared to 22% of those who do not own a pc. Both the car and the home computer are technical devices that are part of the same field of consumption. They are both individual equipment more typical of men, young middle-aged people, and those with high income. These variables are also important for the quality pc-owners have on their equipment. The probability of having an Internet connection among computer owners is strongly related to both income and age (Table 2). Gender, too, has a significant impact. Men more often than women have a computer with an Internet connection. The analysis also shows that people with middle-level education less often have an Internet connection than people with a high-level university education. The reason that there is no significant difference between people with lowest and highest education is explained by age. More than 50% of the respondents on the lowest educational level are under 18 years and still living with their parents.
4. The use of computers In the postal survey questions asked about the use of the computer are divided into four different categories. (a) Private tasks 1 without need of an Internet connection (writing documents, letters, etc., numeric calculation, graphics, illustration, etc., games and other tasks).
444
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
(b) Private tasks 2 with need of an Internet connection (send/receive e-mail from friends and relatives, send/receive e-mail from others/attend discussion groups, use the Internet for information, etc., order goods/services, send/receive documents, and other tasks). (c) Work tasks 1 without need of an Internet connection (writing documents, letters, etc., numeric calculation, graphics, illustration, etc. and other tasks). (d) Work tasks 2 with need of an Internet connection (send/receive e-mail from others/attend discussion groups, use the Internet for information, etc., order goods/services, send/receive documents, and other tasks). Additive indexes are constructed on the basis of these four groups of tasks. The respondents were asked to report how many times they carried out the different tasks during the previous week. If they had not used the computer they were asked to indicate as to how long time it was since they had used it. In the analyses presented in this section only respondents with an Internet connection are included.
4.1. Private use of the home computer The use of computers without need for an Internet connection for private tasks (private tasks 1) is perhaps the kind of use least to do with, with travel. Some of it, like games, might substitute for leisure activities and travel, but there is no necessary connection between the four types of private use recorded here and travel. This is better taken as an indication of how common this kind of technology is in people’s everyday lives. It tells about the tendencies of ‘‘habituation’’ of this type of technology, and is presented for that reason. Private tasks (private tasks 2) which need an Internet connection are sending/receiving e-mail to/from friends, relatives or others, searching for information, news, etc., ordering goods or services, sending/receiving documents, and other tasks. Some of these tasks can be substituted by travel; for example, ordering goods instead of shopping, ordering tickets directly on the Internet from an airline instead of visiting a travel bureau. But ordinary mail and use of the telephone are alternatives to both travel and use of the Internet for many of these tasks. However, some of these would probably not have been done had access to the Internet not been possible. On average, people with an Internet connection send/receive mail from friends and relatives 2.4 times per week; they send/receive mail from others and attend discussion groups 1.8 times per week, they search for information and news 3.3 times per week, order goods and services 0.1 time per week, send/receive documents 1.0 and carry out other tasks 1.5 times per week. Information-seeking therefore represents 33% of these tasks and sending/receiving mail more than 40%, so the substitutionary potential of telecommunication for travel is small for these categories of private tasks. The results for both indexes, categorised as private tasks 1 and private tasks 2, are given in Table 3. Men use the computer for private purposes more than women do, and young people more than the elderly. Those with low education and not gainfully employed use the home computer more
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
445
Table 3 Average use of the home computer for private tasks 1 (the sum of writing documents, letters, etc., numerical tasks, graphics and illustrations, games and other tasks) and private tasks 2 (the sum of sending/receiving E-mail, attending discussion groups, search for information and news, ordering goods/services, sending/receiving documents and other tasks)a Variables
Private tasks 1 times per week
Number of respondents
Gender Male Female
11.0 7.0
743 513
Age (years) 13–17 18–24 25–34 35–44 45–54 55–66 67+
15.8 10.9 9.4 7.1 8.8 7.5 7.8
138 103 329 347 248 72 19
12.7
Education Compulsory school (9 years) Education up to 12 years University level Low grade High grade
a
Private tasks 2 times per week
Number of respondents
0.000 668 456
11.0 11.4 10.2 6.6 7.3 4.5 4.5
115 96 305 312 218 62 16
207
9.8
168
9.7
436
8.9
398
7.6 7.9
217 376
6.9 7.8
200 343
0.000
0.001
0.000
9.3
Sig. (F-test) 0.000
10.3 5.5
Employment Employed > ¼ 40 h per week 8.5 30–40 h per week 8.3 <30 h per week 10.5 Not employed 12.0 Total
Sig. (F-test)
0.260
0.001
0.964
267 570 180 236
8.8 8.2 8.6 8.2
242 516 158 205
1253
8.4
1121
Numbers per week for respondents with Internet.
for writing documents, correspondence, playing games, etc. (private tasks 1) than those with high education and people within the work force. Education and employment are strongly related to age, and this is the main explanation behind the frequent use among these groups. Young men are the most frequent users of the computer for these purposes, and playing games is the most decisive activity. The differences between men and women are most distinct among the youngest (13–17 years). Boys have a weekly use frequency of 16.1; girls of 6.5. In the age group 18–24 years, the frequency of use for boys is 10.4 and 5.1 for girls. Among other age groups there are no significant differences between the genders. These differences between young men and women have also been found in other studies (Telenor, 1999). Use of the home computer for private purposes bears little relation to mobility. Some of these activities can be considered analogous to trips related to various leisure activities and shopping.
446
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
Table 4 Correlation between indexes for private purposes 1 (N ¼ 1253) and private purposes 2 (N ¼ 1121) (see Table 2) – and travel activities (Pearsons r) Trips related to visits Index – use of computer for private purposes 1 Index – use of computer for private purposes 2 *
0.062 0.051
Trips related to shopping
Trips per day in total
Trips as a car driver
km as a car driver per day
0.006
)0.045
)0.009
)0.047
)0.020
)0.042
)0.042
)0.016
)0.027
0.008
Trips related to leisure
Correlation is significant at the 0.05 level (2-tailed).
The correlation 1 analysis, however, shows no significant negative relations between this type of use of the home computer and travelling, as one would have expected if stationary communication was substituted for travelling (Table 4). The activities (especially playing games) are additional to other leisure activities out of home. The correlation analysis shows that there is no relation between car-use, measured as trips as a driver and distance in kilometres per day, number of trips for different purposes, trips as a car driver and use of the computer for private purposes. There is a very small positive relation between use of the computer for private tasks 1 and number of trips related to private visits of friends and relatives. An interpretation could be that there is a slight tendency that high frequency on one activity is correlated with high activity on the other. These results show that use of stationary technology for private purposes is additional to travelling and as such not related activities. 4.2. Use of the home computer for paid work The most relevant tasks related to substitution of travel are to work at home instead of travelling to place of work. As discussed in the Introduction, there is considerable interest in ICT for a possible reduction of work trips as a means of reducing the environmental problems related to car traffic. The most significant differences in application of the home computer for work purposes are found between men and women and between those with long weekly working hours and those with part-time work or those with ‘‘normal’’ working hours (Table 5). Men’s use is about twice that of women’s use, and people who work 40 h or more per week also use it twice as much as people with shorter working hours. Men are in the majority when it comes to working long hours, so those characteristics reinforce each other. The result in Table 6 indicates that work at home for these men comes in addition to their regular work at the workplace. 1
The correlation analyses can only be taken as indications because of the differences in periods for travel activities and use of the home computer (see Section 2).
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
447
Table 5 Average use of the home computer for paid work 1 (the sum of writing documents, letters etc., numerical tasks, graphics and illustrations, and other tasks) and paid work 2 (the sum of sending/receiving E-mail, searching for information/ news, ordering goods/service, sending/receiving documents, and other tasks)a Variables
Gender Male Female Education Compulsory school (9 years) Education up to 12 years University level Lower grade High grade
Paid work 1 times per week
Number of respondents
5.4 3.3
604 364
3.5
Paid work 2 times per week
Number of respondents
5.2 2.5
508 303
97
4.5
73
4.1
339
3.6
279
4.7 5.4
189 341
3.1 5.3
161 295
0.003
258 545 165
Occupational status Manual worker White collar worker Professional Owner
3.4 4.5 5.7 7.7 4.6
a
Sig. (F-test) 0.007
0.004
Employment Employed > ¼ 40 h per week 6.9 30–40 h per week 3.6 <30 h per week 4.1
Total (occupational status)
Sig. (F-test)
0.396
0.000
0.000 6.7 3.5 2.3
217 456 138
344 319 235 58
2.8 4.7 5.1 6.4
300 258 199 43
956
4.0
800
0.033
0.330
Times per week for respondent with paid work and Internet.
Table 6 Correlation between mobile communication (number of work trips, business trips, kilometres as a car driver per day, number of car trips as a driver per day and total trips per day) and stationary communication (index for use work purposes 1 (N ¼ 968) and work purposes 2 (N ¼ 811))a
Index – use of computer for work purposes 1 Index – use of computer for work purposes 2 a
Work trips
Business trips
km as a driver per day
Trips as a driver per day
Total trips per day
0.082
0.103
0.092
0.119
0.081
0.063
0.062
0.073
0.056
0.014
Respondents in paid work with Internet. Correlation is significant at the 0.05 level (2-tailed). ** Correlation is significant at the 0.01 level (2-tailed). *
448
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
Table 7 Linear regression of daily travel length by car in km as a driver for gainfully employed respondents with drivers licence and Internet, N ¼ 786a Variables Work purposes 1 Region (1 ¼ four biggest cities, 0 ¼ else) Gender (1 ¼ male, 0 ¼ female) Household income Number of cars Age Constant a
B
Beta
t-test
Sig.
0.344 4.803
0.079 0.050
2.239 1.414
0.025 0.168
13.949 )0.009 7.292 0.266 )7.955
0.139 )0.007 0.108 0.053
3.939 )0.197 3.014 1.493 )0.977
0.000 0.844 0.003 0.136 0.329
R2 ¼ 0:045; adjusted R2 ¼ 0:037.
As the correlation 2 analysis shows, there is no substitution of mobile communication, travelling, by use of stationary communication devices. It is rather a tendency of the opposite, especially in relation to work purposes 1. People who use the computer at home in relation to their paid work have some more and longer trips than people who do not use the computer for such tasks. However, the correlation coefficients are small, so it is difficult to give an absolute conclusion. But it seems that rather than being a substitute for mobile communication, stationary communication can be additional when the computer is used for tasks related to paid work. The correlation between the index for work purpose 2 and number of kilometres by car as a driver also shows a small positive relation. Those who use the home computer for work-related tasks tend to drive longer distances (with themselves as driver) that those who do not use it. A primary conclusion is that there is no substitution of travel by use of communication technology among those who own and use a home computer(s). To control for the effect of other variables on car-use a regression analysis of total daily travel length by car as a driver is done. The independent variables are place of living, gender, household income, age and number of cars in the household in addition to the use of the home computer for work purposes 1. For the analyses we have selected respondents with Internet and drivers licence. Table 7 shows a significant positive relation between the amount of travel (km per day as a driver) and use of the home computer for work purposes 1. This analysis reinforces the conclusions of the correlation analyses. There is a tendency that people who use the home computer frequently for work purposes also are frequent users of the car, even though the relation is not very strong. The analysis presented in Table 7 shows however, that gender and number of cars in the household are more important for car-use than use of the home computer. 5. The possibility to work at home and the amount of travel activity The analyses presented in Section 4 include only pc owners. In this section we therefore want to compare travel activity for different groups in relation to the ownership of home computers and the possibility to work at home. 2
See Footnote 1.
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
449
In the questionnaire the respondents (with home computers) were asked if they had the possibility to work at home and, if so, how many times (whole days or part of a day) they had done so during the previous week. Of the sample, 30% said they can work at home. 6% have a permanent workplace at home, which means that they work mostly at home, and 64% cannot work at home. In Table 8 the daily travel patterns of employed groups with an annual income of NOK 200.000 or higher (to get the groups more homogeneous) with different access to home computers and possibilities for working at home are compared. People without home computers represent one category. (Not all trip purposes are presented in the table, so the total number of trips is higher than the sum of those presented in the table). The results indicate that variations in travel patterns are fairly small. There is no significant difference in the total number of kilometres driven by car per day, even if there is a tendency for people who work mostly at home to travel a little shorter by car as a driver. The same group has fewer work trips, but the total number of trips per day is the same as for the other groups with a home computer who do not work permanently at home, and even more than those without a home computer. This supports the hypothesis that a reduction in one type of journey or trip will be replaced by other kinds; for instance, work trips can be replaced by trips related to leisure or shopping. In this case we see that those who work at home have more chauffeuring trips than those without home computers, which could indicate that working at home might be an adjustment to a family situation with children. To control for the effects of the differences between these groups and of other variables which have an impact on travel in addition to pc ownership regression analyses are done for work trips, total number of trips and trips related to chauffeuring. Since only pc-owners got the question about the option to work at home, it is not possible to include this aspect in the analysis. To make
Table 8 Number of trips per day for different purposes (not all purposes are shown) for employed people with different access to a home computer and possibilities for work at homea;b Different trips
Without home computers
(1) Trips related to visits per day Trips related to leisure per day Trips related to shopping per day Trips related to chauffeuring per day Work trips per day Business trips per day Number of trips per day Number of trips as car driver per day km as a car driver per day Total a b
0.37 0.46 0.71 0.25 1.03 0.15 3.26 2.07 29.2 1454
(3.4) (3) (2.3.4) (4) (3.4) (4)
With home computer, cannot work at home (2) 0.30 0.45 0.79 0.35 0.91 (4) 0.13 (3.4) 3.45 2.10 30.6 820
With home computer, can work at home (3) 0.26 0.60 0.67 0.31 0.97 (4) 0.27 3.60 2.22 28.8 444
Employed with own income NOK 200.000 or more per year. Numbers per day. (x) Significantly different from group x, P < 0:05.
With home computer, have permanent work site at home (4) 0.24 0.55 0.74 0.49 0.68 0.19 3.87 1.95 23.6 84
450
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
the analyses as comparable as possible with the analysis presented in Table 8, only respondents with a personal annual income of NOK 200.000 or more are included. Table 9 shows a linear regression analysis of the average number of work trips per day controlled for the effects of pc-ownership, residential region, gender, household income, number of cars and age. The analysis shows a small significant effect of pc-ownership on the number of work trips per day. Respondents with a pc have in average less work trips per day than people without. The analysis of the total number of trips per day in Table 10 shows that there are no significant difference in the total number of daily trips between pc-owners and non-owners. The reduction in work trips is ‘‘compensated’’ for by more other types of trips. Gender, income, number of cars and age are the important variables to explain the level of daily travel activity. Table 11 shows the results of the analysis of trips related to chauffeuring children, and we see that pc owners have more such trips than non-owners. These three regression analyses give an indication of a certain reduction of work trips for pcowners compared with non-owners. However, there is no difference in total travel activity between the groups, which means more trips for other purposes. In this case we see that the respondents with home computers have more of their travel activity related to bringing children to different activities. As we saw in the analysis presented in Table 8, this might indicate an adjustment to a family situation. Table 9 Linear regression of number of work trips per day for gainfully employed respondents with income equal or greater than NOK 200.000, N ¼ 2870a Variables
B
Beta
t-test
Sig.
PC-owners (1 ¼ yes, 0 ¼ no) Region (1 ¼ four biggest cities, 0 ¼ else) Gender (1 ¼ male, 0 ¼ female) Household income Number of cars Age Constant
)0.077 0.053
)0.035 0.025
)2.126 1.532
0.034 0.126
0.094 0.000 0.036 )0.002 0.966
0.042 0.025 0.025 )0.021
2.591 1.516 1.495 )1.290 11.020
0.010 0.130 0.135 0.197 0.000
a
R2 ¼ 0:005; adjusted R2 ¼ 0:003.
Table 10 Linear regression of the total number of trips per day for gainfully employed respondents with income equal or greater than NOK 200.000, N ¼ 2870a Variables PC-owners (1 ¼ yes, 0 ¼ no) Region (1 ¼ four biggest cities, 0 ¼ else) Gender (1 ¼ male, 0 ¼ female) Household income Number of cars Age Constant a
R2 ¼ 0:016; adjusted R2 ¼ 0:014.
B
Beta
t-test
Sig.
0.145 0.151
0.031 0.033
1.862 2.012
0.063 0.004
)0.004 0.000 0.192 )0.020 3.704
)0.001 0.033 0.062 )0.091
)0.051 1.962 3.652 )5.636 19.967
0.960 0.050 0.000 0.000 0.000
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
451
Table 11 Linear regression of number of chauffeuring trips per day for gainfully employed respondents with own income equal or greater than NOK 200.000, N ¼ 2870a Variables PC-owners (1 ¼ yes, 0 ¼ no) Region (1 ¼ four biggest cities, 0 ¼ else) Gender (1 ¼ male, 0 ¼ female) Household income Number of cars Age Constant a
B 0.072 0.0199 )0.092 0.001 0.062 )0.054 0.418
Beta
t-test
Sig.
0.040 0.028
2.430 0.698
0.015 0.485
)0.051 0.032 0.052 )0.066
)3.121 1.935 3.103 )4.061 5.925
0.002 0.053 0.002 0.000 0.000
R2 ¼ 0:013; adjusted R2 ¼ 0:012.
6. Discussion Considering the limitations related to the data sets, the results of the analyses show only to a very little extent any relation between ownership and use of a home computer and people’s daily travel patterns. On the basis of these results we cannot see any direct substitutionary effects of the use of stationary technology at people’s home on the use of mobile technology. Access to and use of information technology at home do not seem to have a significant impact on travel activities in everyday life. The analyses give some indications of adjusting work and family life, but the net effect gives no reduction in travel activity. Stationary communication seems to be a supplement to activities based on mobile technology, but it gives people more spatial and temporal options. For people who work more than normal weekly working hours, stationary technology seems to give them greater flexibility in regard to where and when to work, but it does not necessarily reduce their travel activity. Flexibility in everyday activities and travel is also what Mokhtarian (1998) suggests as an important impact of this ‘‘new’’ communication technology. The spatial flexibility will also give a temporal flexibility, which means that work trips and other trips can be more dispersed over the day than is the situation today. The positive consequence can be a reduction in the rush-hour traffic; the negative is that it is more difficult to offer a good public transport service when travel needs are more spread in time. The costs to offer a high frequency supply during all day could be high. What these analyses also reveal is a relationship between the ownership of cars and the ownership of computers, and, as such, to a high degree the same social groups who use the car and the computer. Men and high-income groups are more frequent users of the car than women and people with lower income. Men, high-income groups and people with high education more often own a home computer and have the possibility to work at home than people with low income and education. The possibility for flexibility related to the use of home computers and the new information and communication technology in general follow the existing divisions of different segments of the labour market. One crucial question is how the different segments will develop and their size. How many of the future workers of the labour stock will have the possibility to work at home, at neighbourhood centres, or to be mobile workers in general; how many want this type of flexibility and do companies want it?
452
R.J. Hjorthol / Transportation Research Part A 36 (2002) 437–452
If the possibility to flexibility regarding where and when to work continues to be a good for the privileged (high education/high income) in the future, it will be of great interest to know more about their preferences for living and also their daily activities in order to foresee the consequences for their travel patterns. In the long term changes in preferences related to the greater flexibility is possible, and can have impacts on urban development and land use patterns. Acknowledgements The research described in this paper was supported by the Strategic Institute Program of the Institute of Transport Economics. Thanks to Unni Lodden for her help in the preparation of the survey data and to the referees who provided many helpful comments. References Balepur, P.N., Varma, K.V., Mokhtarian, R.L., 1998. Transportation impacts of center-based telecommuting: interim findings from the neighborhood telecenters project. Transportation 25, 287–306. Batten, D., 1989. The future of transport and interface communication: debating the scope for substitution growth. In: Batten, D., Roland, T. (Eds.), Transportation for the Future. Springer, Berlin. Capello, R., Gillespie, A., 1993. Transport communications and spatial organisation: conceptual framework and future trends. In: Nijkamp, P. (Eds.), Europe on the Move. Aldershot mv, Avebury. Engstr€ om, M.G., Johanson, R., 1996. IT-utviklingens effekter p a framtida res- och transportstrukturer. (The effects of IT on transport and travel). Naturv ardverket F€ orlag, Stockholm. Hamer, R., Kroes, E., van Oostsroom, H., 1991. Teleworking in the Netherlands, an evaluation of changes in travel behaviour. Transportation 18, 365–382. Harvey, D., 1989. The Condition of Postmodernity. An Enquiry into the Origins of Cultural Change. Basil Blackwell, Oxford. Henderson, D.K., Mokhtarian, P.L., 1996. Impacts of center-based telecommuting on travel and emission: analysis of the pudget sound demonstration project. Transp. Res. D 1 (1), 29–45. Hjorthol, R., 1998. Hverdagslivets reiser. En analyse av kvinners og menns daglige reiser i Oslo (Travel activities in everyday life. An analysis of women’s and men’s daily travel activities in Oslo. Dissertation for dr. philos.). Insitutt for sosiologi og samfunnsgeografi, Universitetet i Oslo, Institute of Transport Economics, TØI Report 391/1998, Oslo. Mokhtarian, P.L., Handy, S., Salomon, I., 1995. Methodological issues in the estimation of the travel, energy and air quality impacts of telecommuting. Transp. Res. A. (4), 283–302. Mokhtarian, P.L., 1990. A typology of relationships between telecommunications and transportation. Transp. Res. A 24 (3), 231–242. Mokhtarian, P.L., 1998. A synthetic approach to estimating the impacts of telecommuting on travel. Urban Studies 35 (2), 214–241. Nilles, J.M., Carlson, F., Gray, P., Hanneman, G., 1976. The Telecommunications – Transportation Tradeoff. Wiley, New York. Nilles, J.M., 1991. Telecommunting and urban sprawl: mitigator or incsiter? Transportation 18, 411–432. SIKA, 1998. IT-utveklingen och transporterna 2. Redovisning av en kommunikationsunders€ okning 1997. (The development of IT and the transport. A study of communication from 1997) SIKA Rapport 1998:4, Stockholm. SSB, 1993. Forbrukerundersøkelsen 1989–1991 (Survey of consumer expenditure 1989–1991). Statistisk sentralbyr a. NOS C 65, Oslo, Kongsvinger. Telenor, 1999. Ungdom og internett. (Youth and Internet) Faglig nyhetsbrev fra Telenor FOU (1), January 1999, Oslo. Vaage, O.F., 1998. Norsk mediebarometer 1997. (Norwegian barometer of media, 1997) Statistisk sentralbyr a. Statistiske analyser 24, Oslo, Kongsvinger.