Cecilia Jakobsson Bergstad, Amelie Gamble, Olle Hagman, Lars E. Olsson, Merritt Polk & Tommy Gärling

Transcription

1 Göteborg Psychological Reports University of Gothenburg SWEDEN Cecilia Jakobsson Bergstad, Amelie Gamble, Olle Hagman, Lars E. Olsson, Merritt Polk & Tommy Gärling Affective-Symbolic and Instrumental-Independence Psychological Motives Mediating Effects of Socio- Demographic Variables on Daily Car Use. Number 3 Volume

2 ISSN ISRN GU/PSYK/REP--2009/3--SE

3 Affective-Symbolic and Instrumental-Independence Psychological Motives Mediating Effects of Socio-Demographic Variables on Daily Car Use Cecilia Jakobsson Bergstad, Amelie Gamble, Olle Hagman, Lars E. Olsson, Merritt Polk, and Tommy Gärling Jakobsson Bergstad, C., Gamble, A., Hagman, O., Olsson, L. E., Polk, M., and Gärling, T. Affective-symbolic and instrumentalindependence psychological motives mediating effects of sociodemographic variables on daily car use. Göteborg Psychological Reports, 2009, 39, No. 3. This study investigates the extent to which affective-symbolic and instrumental-independence psychological motives mediate effects of socio-demographic variables on daily car use in Sweden. Available national travel diary data for 16,944 car users collected in 2005 and 2006 and questionnaire data from a mail survey to 1,134 car users collected in 2007 were used to assess the relationships daily car use as driver or passenger have to sex, age, household type (single or cohabiting with or without children), education, employment, income, residential area (urban, semi-rural or rural), access to cars and annual driving distance. On the basis of the questionnaire data reliable measures of affective-symbolic and instrumental-independence motives were constructed. The results show that households with children use the car more than households with no children, that men make more car trips as drivers than women who use the car as passenger more than men, and that households living in rural areas use the car more than households living in semirural areas who use the car more than households living in urban areas. An affective-symbolic motive partially mediates the relationship between the number of weekly car trips and sex, the instrumental-independence motive partially mediates the relationships between weekly car use and percent car use as driver and several of the socio-demographic variables (living in urban vs. rural residential area for both measures; sex and living in urban vs. semi-rural residential area for percent car use as driver). Of several other sociodemographic variables (age, employment, and income) affecting car use, only the relationship of the number of cars to percent car use as driver was (partially) mediated by the instrumental-independence motive. Key words: Car use, socio-demographic variables, psychological motives In the industrialised world the private car is the primary mode of daily travel for a majority of people. This was not the case half a century ago. The forces behind increasing car use are many. At a societal level, increasing mobility is associated with technological development, economic growth, and positive social change. At the individual level, these changes interact with psychological motives resulting in strong individual preferences for cars and car use. Lars E. Olsson is affiliated with Karlstad University, Sweden.

4 No.3:39, 2 There is a general consensus that car use needs to be reduced in the face of negative impacts on the human environment (Gärling & Steg, 2007), the societal and human costs of traffic accidents (Groeger & Rothengatter, 1998), and the growing health problems caused by car-related sedentary life styles (Dora & Phillips, 2000). Still, the car is difficult to replacebecause of its many beneficial effects for people, such as comfort and door-to-door flexibility. For many people driving is also in itself a positive and enjoyable activity (Gatersleben, 2007; Mokhtarian, 2005). Car use depends on access to a car and increases with the number of cars in a household (Gärling & Loukopoulos, 2007). Car use is also related to several socio-demographic variables. It is higher among households in rural areas than among households in urban areas (Stradling, 2007), higher among households with children than households without children (Polk, 2004), and increases with income and employment (Dargay & Hanly, 2007). Men furthermore drive more than women (Polk, 1998, 2004). Three complementary approaches explaining travel are discernible in transport research. In one approach pioneered by geographers (e.g., Hanson & Schwab, 1995), the focus is on objective descriptions of activity/travel patterns. Structural determinants of these activity/travel patterns are examined, including socio-demographic variables. A second approach emanates from disaggregate travel-choice modeling (Ben-Akiva & Lerman, 1985; McFadden, 2001). Its goal is to understand how activity/travel patterns are determined by the travel choices made by people in different socio-demographic strata. Consequently, it is believed that the effects of socio-demographic variables need an explanation. A third approach (Hagman, 2003) goes further in its attempts at identifying explanations for travel. Its goal is to find psychological motives for the travel choices people make. In line with the last approach, the aim of the present research is to investigate whether psychological motives account for the effects of socio-demographic variables on daily car use. It is assumed, consistent with the second approach, that choices of car use mediate the effects of the socio-demographic variables on car use. These effects should then also be mediated by psychological motives which determine choices of car use. Previous research has documented two types of psychological motives for car use. One motive is referred to as instrumental and is related to travel time, costs, physical effort, and flexibility. Another motive is referred to as affective-symbolic, focusing on the affective and symbolic outcomes of driving, such as enjoyment, feelings of status, stress, autonomy, and safety (Anable & Gatersleben, 2005; Jakobsson, 2007; Steg, Vlek, & Slotegraaf, 2001). While some previous studies have succeeded in empirically distinguishing between these motives (e.g. Steg, 2005), other studies suggest that the affective and instrumental motives for car use are not possible to distinguish (e.g. Mann & Abraham, 2006). In a review of previous research, Gatersleben (2007) also posited a third motive referred to as feelings of independence found to be related to the positive experiences of using the car. In accordance with previous research on psychological motives and car use (e.g. Steg, 2005), affective-symbolic motives are expected to be more influential on car use among frequent car drivers than instrumental and independence motives. The latter two motives may play a more important role for car use in rural areas and for multi-person households with children but a weaker role for members of younger, urban, or single households. Despite such expectations inferred from previous research, there is no extant theory from which explicit hypotheses can be derived, except the general hypothesis that psychological motives, due to their effects on choices of car use, mediate the relationship between car use and sociodemographic variables. The specific aim of the present study is to determine whether the psychological motives for car use mediate the relationships between car use and such socio-demographic variables that are known to affect car use. Both measures of car use and motives for car use are obtained from questions in a mail survey conducted in Sweden where car access is high

5 No.3:39, 3 (Swedish Institute for Transport and Communications Analysis, 2007). In the Swedish population 82% men and 70% women have access to one or more cars. A car is used at least once a week by 90% of the households, and for 64% of the total distance travelled the car is the primary travel mode. The measures of car use obtained from questions in the mail survey are validated by analyses of national travel diary data. Method National Travel Survey In order to determine the relationship between socio-demographic variables and car use in the Swedish population, data from the national travel surveys conducted in 2005 and 2006 (Swedish Institute for Transport and Communications Analysis, 2007) were analyzed. The surveys consist of filled-in travel diaries for a single day. Randomly sampled respondents between the ages of 6 and 84 years old received a letter describing the survey with instructions for how to fill in the travel diary on a specified day, which was randomly selected throughout the year. The respondents were phoned the day after and asked questions about their travel. Respondents between 6 and 17 years of age were contacted via their parents. The sample consisted of 41,225 respondents representing a response rate of 67.6%. An analysis of dropouts revealed no sampling biases. In the present analyses 16,944 respondents with a driver s license and a car in the household are included. Questionnaire Survey Between October and November 2007 a questionnaire was mailed to a sample of 3,000 Swedish residents consisting of approximately 1,000 randomly selected from each of metropolitan areas with more than 200,000 residents, semi-rural areas with between 20,000 and 200,000 residents, and rural areas with less than 20,000 residents. A total of 1,332 usable questionnaires were obtained (representing a response rate of 44.4%). Respondents were selected for the analyses if they stated in the questionnaire that they have a driving license and access to a car. This resulted in a sample of 1,134 respondents. A description of the sample is given in Table 1. The questionnaire was sent by regular mail including a free-of-charge return envelope. It took about 15 minutes to answer. The respondents were offered the possibility to obtain two lottery tickets in compensation for their participation. A combined thank-you and reminder card was sent after approximately one week giving respondents the option to answer the questionnaire via the web. About 5% of the usable questionnaires were obtained in this way. Several modules were included in the questionnaire covering also issues that will be reported elsewhere (Jakobsson Bergstad et al., 2009a, 2009b). The modules providing the data for the present analyses consisted of (1) answers to 32 statements related to the, affectivesymbolic, instrumental, and independence psychological motives for car use (adapted from Steg, 2005, see Table 2) obtained on seven-point rating scales ranging from totally disagree (0) to agree completely (6), (2) questions about the frequency of performance of daily activities 1 and whether a car was used as a driver or passenger to travel to these activities 2, and (3) questions about socio-demographic factors including sex, age, cohabiting status, presence of children, education, employment, income, number of cars, and annual driving distance. 1 These included the most frequently reported out-of-home activities in the Swedish national travel surveys including work or study; purchases of non-durables; other purchases; participating in sports, exercise or outdoor activities; participating in out-of-home hobbies, religious, course or club activities; visiting relatives and friends; visiting restaurants, cafés or entertainment/cultural events; picking up or leaving children at school or day care centre, and; participating in children s leisure activities. 2 For each activity respondents reported how many times during the previous week (never, one, two, three, four, five, or more than five times) they had travelled by car (as driver or passenger) or used any other travel mode with the purpose of performing the activity.

6 No.3:39, 4 Table 1. Sample descriptives (total sample size = 1134). Variable Sample size Descriptive Sex (% women) Age (years) (M/SD) /12.4 Household type (%) Single households with no children Single households with children Cohabiting households no children Cohabiting households with children Other University education (%) Full-time employment (%) Percent employment (M/SD) /42.8 Annual household gross income, in 000 SEK* (%) < >600 Residential area (%) Rural (< inhabitants) Semi-rural ( inhabitants) Urban (> inhabitants) Number of cars in household (M/SD) /0.7 Estimated annual driving distance in km (M/SD) ,865/16,026 1 SEK is approximately equal to 0.10 Euro Results Travel Diary Data Table 2 shows that households with children make more car trips and travel by car longer distances than households with no children. Table 3 indicates that men make more car trips and drive longer distances than women, whereas women use the car as passenger more than men. As may be seen in Table 4, car use is larger in rural than in semi-rural residential areas, and larger in semi-rural than in urban residential areas. Table 2. Mean number of daily car trips and mean daily car use in km as driver and passenger related to household type. (Respondents with a driver s license and with one or more cars in the household). Single without children N= 2527 Single with children n= 637 Cohabiting without children n= 7614 Cohabiting with children n= 5693 # daily car trips Km car use as driver Km car use as passenger

7 No.3:39, 5 Table 3. Mean number of daily car trips and mean daily car use in km as driver and passenger related to sex. (Respondents with a driver s license and with one or more cars in the household). Men N = 9066 Women n= 7878 # daily car trips Km car use as driver Km car use as passenger Table 4. Mean number of daily car trips and mean daily car use in km as driver and passenger related to residential area. (Respondents with a driver s license and with one or more cars in the household). Urban n = 2019 Semi-rural n = 4680 Rural n = 589 # daily car trips Km car use as driver Km car use as passenger Questionnaire Data Table 5 displays means, standard deviations, skewness, and intercorrelations of the ratings of the motives. The conditions for performing a principal component analysis (PCA) were met in that some correlations were larger than.30, the Kaiser-Meyer-Oklin measure exceeded.60, and the Bartlett s test of sphericity was statistically significant (p <.001) (Hair Anderson, Tatham, & Black, 2006). Two components were extracted in the PCA and submitted to an oblique (oblimin) rotation. The two components accounted for 41.5% of the total variance. After removing 7 items with component loadings less than.30 or communalities less than.50, an additional PCA extracted two components explaining 46.8 % of the variance. The component loadings after rotation are given in Table 6. As may be seen, the first component loads on affective-symbolic motives, whereas the second component loads on both instrumental and independence motives for car use. A reliable scale of affective-symbolic motives were constructed by averaging across the ratings with high loadings (.54, boldface in the table) only on the first component (Cronbach s α =.91), and a reliable scale of instrumental-independence motives were constructed by averaging across the ratings with high loadings (.51, boldface in the table) only on the second component (Cronbach s α =.87). In order to determine the effects of the socio-demographic variables on the measures of the affective-symbolic and the instrumental-independence motives, two OLS multiple linear regression analyses were performed (see Table 7). The independent variables were sex, age (36-54 years vs years and 55+ years vs years), having children or not, cohabiting or not, education (university degree vs. no university degree), employment (percent of fulltime), income (coded 1-6), residential area (rural vs. urban and semirural vs. urban) and number of cars. In each analysis the motive not used as a dependent variable was entered as an independent variable in order to control for its influence on the other independent variables. Table 4 shows that several of the independent variables had a significant effect on both the symbolic-affective and instrumental-independence motives, accounting for 22% and 19% of the variance, respectively. A set of OLS multiple linear regression analyses was performed to determine the extent to which the affective-symbolic and instrumental-independence motives mediate the effects of the socio-demographic variables on daily car use. Three different measures of daily car use were used as dependent variables: (1) Number of car trips to the activities during previous week; (2) Percent car use as driver to the activities during previous week, and; (3) Percent car

8 Table 5. Means (M), standard deviations (SD), skewness (Skew) and inter-correlations of ratings of motives for car use (n = 1134) Statement M SD Skew Express myself through my car My car suits me The car brings me wherever I want Driving saves a lot of time Driving is sporty and adventurous Driving is my hobby My car has a nice road-holding The car gives me power in traffic Driving makes my life more easy I can visit friends, family Feelings of freedom Provides privacy Protection against bad weather Driving is comfortable I enjoy driving a nice, good-looking car The car is always available I am free to stop everywhere I am safe in my car I can distinguish myself from others I get a kick out of driving I am not dependent on others I can go out I am a bit in love with my car Driving is enjoyable I can chose my own route The car gives me prestige Its carrying capacity (luggage, purchases) Enables recreational trips and holidays I can pick up and see off others Driving is relaxing I love driving fast I love the drone of my engine and muffler *

9 Table 5 (continued) Statement I am free to stop everywhere I am safe in my car I can distinguish myself from others I get a kick out of driving I am not dependent on others I can go out I am a bit in love with my car Driving is enjoyable I can chose my own route The car gives me prestige Its carrying capacity (luggage, purchases) Enables recreational trips and holidays I can pick up and see off others Driving is relaxing I love driving fast I love the drone of my engine and muffler Note. Statements are shown in the order they were presented in the questionnaire.

10 8 Table 6. Obliquely (oblimin) rotated principal-component loadings of ratings of affectivesymbolic and instrumental-independence psychological motives for car use. Component 1 Affective-symbolic Component 2 Instrumental-independence I get a kick out of driving Driving is sporty and adventurous The car gives me prestige Driving is my hobby Express myself through my car I love the drone of my engine and muffler I can distinguish myself from others The car gives me power in traffic I enjoy driving a nice, good-looking car I love driving fast I am a bit in love with my car Driving is enjoyable Provides privacy Driving is relaxing I can visit family and friends I can go out I am free to stop everywhere I can chose my own route Driving makes my life more easy I am not dependent on others The car brings me wherever I want Driving is comfortable Feelings of freedom the car gives me Driving saves a lot of time Enables recreational trips and holidays Table 7. OLS multiple linear regression analyses with the dependent variables affective-symbolic motive and instrumental-independence motive. Affective-symbolic motives Instrumental-independence motives n β t p n β t p Affective-symbolic motive <.001 Instrumental-independence motives < Sex (man 1, woman -1) < Age (36 54 years 1 vs years -1) Age (55 + years 1 vs years -1) < Cohabitant (1 vs. single -1) marital status Children (1 vs. no children -1) University (1 vs. lower -1) education Employment (%) Income (1-6) Urban (-1) vs semi-rural (1) residential area Urban (-1) vs rural (1) residential area <.001 Number of cars Model R 2 adj =.22, F(12, 995) = 24.89, p<.001 R 2 adj =.19, F(12, 995) = 21.24, p<.001 use as passenger to the activities during previous week. Means and standard deviations of the car use and motive measures are given in Table 8 for each household type, in Table 9 for men and women, and in Table 10 for each residential area.

11 9 Table 8. Mean (M) and standard deviations (SD) of number of weekly car trips, percent car use as driver, percent car use as passenger, and ratings of motives for car use related to household types. Single without children Single with children Cohabiting without children Cohabiting with children # weekly car trips 9.1 (7.3) 13.3 (7.7) 9.6 (7.8) 13.0 (9.5) % car use as driver 73.1 (33.9) 81.7 (27) 66.5 (35.9) 72.3 (29.7) % car use as passenger 8.5 (17.5) 8.2 (17.0) 15.5 (24.6) 12.3 (19.1) Affective-symbolic motives (0-6) Instrumentalindependence motives (0-6) 1.9 (1.3) 2.0 (1.2) 1.8 (1.2) 1.9 (1.2) 5.0 (1.0) 5.2 (0.7) 5.0 (0.9) 5.0 (0.9) Table 9. Mean (M) and standard deviations (SD) of number of weekly car trips, percent car use as driver, percent car use as passenger, and ratings of motives for car use related to sex. Men Women Number of weekly car trips 12.2 (9.4) 10.5 (8.1) % car use as driver 77.3 (29.4) 64.8(34.6) % car use as passenger 7.5 (14.9) 17.5 (24.5) Affective-symbolic motives (0-6) 2.0 (1.2) 1.7 (1.1) Instrumental-independence motives (0-6) 4.9 (0.9) 5.0 (0.9) Table 10. Mean (M) and standard deviations (SD) of number of weekly car trips, percent car use as driver, percent car use as passenger, and ratings of motives for car use related to different household types. Urban Semi-rural Rural # weekly car trips 8.2 (7.9) 11.7 (7.8) 11.7 (9.3) % car use as driver 53.1 (38.1) 72.1 (30.8) 74.5 (31.4) % car use as passenger 7.9 (14.2) 12.8 (20.2) 13.9 (23.1) Affective-symbolic motives (0-6) 2.0 (1.3) 1.9 (1.2) 1.8 (1.2) Instrumental-independence motives (0-6) 4.6 (1.0) 5.0 (0.83) 5.1 (0.9) In the first step of the regression analysis, the same socio-demographic variables used in the previous regression analyses were entered as independent variables. Table 11 shows that on the number of weekly car trips seven socio-demographic variables (sex, 55+ years vs years, having children, percent employment, residential area, and number of cars) are significant accounting for 9% of the variance. On the dependent measure percent car use as driver all the independent variables except age and percent employment are significant accounting for 15% of the variance. On the dependent measure percent car use as passenger sex, cohabiting, having children, percent employment and residential area (rural vs. urban) are significant accounting for 7% of the variance.

12 Table 11. OLS multiple linear regression analyses with the dependent variables number of weekly car trips, percent car use as driver, and percent car use as passenger and sociodemographic variables entered in the first step and the affective-symbolic and the instrumental-independent motives entered in the second step as the independent variables. Number of weekly car trips Percent car trips as driver Percent car trips as passenger Step 1 n β t P n β t p n β t p Sex (man 1, woman -1) < <.001 Age (36 54 years 1 vs years -1) Age (55 + years 1 vs years -1) Cohabitant (1 vs. single -1) marital status < Children (1 vs. no children -1) < University (1 vs. lower -1) education Employment (%) Income (1-6) Urban (-1) vs semi-rural (1) residential area < Urban (-1) vs rural (1) residential area < Number of cars < < Model R 2 adj =.09, F(11,996) = 9.51, p<.001 R 2 adj =.15, F(11, 996) = 16.97, p<.001 R 2 adj =.07, F(11,996) = 8.58, p<.001 Step 2 n β t P n β t p n β t p Affective-symbolic motive Instrumental-independence motives < Sex (man 1, woman -1) < <.001 Age (36 54 year 1 vs years -1) Age (55 + year 1 vs years -1) Cohabitant (1 vs. single -1) marital status Children (1 vs. no children -1) < University (1 vs. lower -1) education Employment (%) Income (1-6) Urban (-1) vs semi-rural (1) residential area Urban (-1) vs rural (1) residential area Number of cars < Increment R 2 adj =.02, F(2,994) =12.20, p<.001 R 2 adj =.03, F(2,994) = 21.55, p<.001 R 2 adj =.01, F(2,994) = 4.57, p<.001 Augmented model R 2 adj =.11, F(13, 994) = 10.11, p<.001 R 2 adj =.18, F(13, 994) = 18.28, p<.001 R 2 adj =.08, F(13, 994) = 8.01, p<.001

13 In the second step the measures of affective-symbolic and instrumental-independence motives were entered. Significant increases in accounted variance (2%, 3%, and 1% respectively) are observed. The regression coefficients for the instrumental-independence motives are significant on the dependent variables number of weekly car trips and percent car as driver, and close to significant for the dependent variable percent car as passenger (p =.051). The regression coefficients for affective-symbolic motives are only significant on the number of weekly car trips. Sobel tests showed that the affective-symbolic motives partially mediated the relationship between the number of weekly car trips and sex (z = 2.08, p =.038), that the instrumental-independence motives partially mediated the relationship between the number of weekly car trips and living in urban vs. rural residential areas (z = 2.06, p =.039), and that the instrumental-independence motives partially mediated the relationships between the percent car use as driver and the socio-demographic variables sex (z = 2.45, p =.014), urban vs. rural residential area (z = 2.40, p =.016), urban vs. semi-rural residential area (z = 2.08, p =.037), and the number of cars in household (z = 2.03, p =.042). Discussion The present study set out to investigate whether the effects of socio-demographic variables on daily car use in Sweden are mediated by affective-symbolic and instrumentalindependence psychological motives. It is concluded, firstly, that there were differences in daily car use between men and women, between households with and without children, and among households living in urban, rural, and semi-rural areas. Secondly, from 7% to 15% of the variance in weekly car use in travel related to frequently performed activities was accounted for by the socio-demographic variables (including also age, university education, percent employment, income, and number of cars). Thirdly, it was possible to construct reliable questionnaire measures of affective-symbolic and instrumental-independence psychological motives for car use. In accordance with the main aim of the present research, evidence was found that the psychological motives mediate and thus explain some of the differences in car use between different socio-demographic groups. The implication is that the psychological motives are proximal determinants of choices of car use. However, primarily the instrumentalindependence psychological motive mediated the relationships between daily car use and the socio-demographic variables, and apart from the effect of sex which was partially mediated by both psychological motives (implying that men drive more than women both due to affectivesymbolic and instrumental-independence motives), the mediated socio-demographic variables were restricted to residential area and the number of cars in the household. The partial mediation as well as failures of mediation implies that there are also direct effects of sociodemographic variables on daily car use. Yet, for this reason the results do not exclude other individual-level explanatory factors, not measured in the present study. A full mediation of residential area may require a more refined categorization than in urban, semi-rural, and rural. Within each of these types of areas, residential location is likely to vary substantially with respect to accessibility to different activity nodes. Even though partial mediation was found at the chosen aggregation level, it may change to full mediation at a more disaggregate level. Furthermore, the number of cars available to the household, that will increase accessibility to activity nodes, in part should have reduced the mediation of the instrumental-independence motive. At the same time, the mediation of the number of cars suggests that the versatility of the car is a similar benefit to all households, something that would otherwise not be revealed in analyses of individual differences. An issue that needs to be addressed further is why the affective-symbolic motive was rated to have little impact and only mediated the effect on daily car use of sex, not the other sociodemographic variables. One possible reason, pointed out by Steg et al. (2001), is that measures of affective-symbolic motives are influenced by social desirability. It is not socially desirable to express a positive affect for owning and driving a car and that owning a car is an

14 important status marker. Therefore, some indirect method is needed to assess its effect. Another perhaps more plausible reason, supported by the present mediation analyses, is that the instrumental-independence motive pertains to those segments of the population (people living in semi-rural and rural areas) dominating the sample. An affective-symbolic motive may be more important for urban-living, car-owning male and female singles without children. In addition, car use was measured for travel to frequently performed out-of-home activities, many of which are instrumental. Some previous studies have distinguished between affective-symbolic and instrumental psychological motives for car use (e.g. Steg et al. 2001), whereas other studies have concluded that it is not possible to identify different motives (e.g. Mann & Abraham, 2006). Still others (e.g. Steg, 2005; Gatersleben, 2007) have suggested the existence of more than two motives. The present study including a large number of statements covering the motives for car use, administered in a survey to a large population-based sample, revealed two distinct, correlated motives which were identified as affective-symbolic and instrumental-independence motives. Reliable scales measuring these motives were possible to construct. The present research thus takes a further step towards clarifying the motive structure of private car use, as well as to provide a measuring instrument that can be used in future research for assessing these motives. If required, additional refinements may be possible by distinguishing the affective, symbolic, instrumental, and independence components. Acknowledgements Financial support for this research was obtained by grant #AL90A 2006:16305 from the Swedish Road Administration through the program Social and economic impacts of the automobile ( to the Centre for Consumer Science, School of Business, Economics, and Law, University of Gothenburg, Sweden. References Anable, J., & Gatersleben, B. (2005). All work and no play? The role of instrumental and affective factors in work and leisure journeys by different travel modes. Transportation Research A, 39, Ben-Akiva, M., & Lerman, S. (1985). Discrete choice analysis: Theory and applications. Cambridge, MA: MIT Press. Dargay, J., & Hanly, M. (2007). Volatility of car ownership, commuting mode and time in the UK. Transportation Research A, 41, Dora, C., & Phillips, M. A. (2000). Transport, environment, and health (no. 89). Copenhagen: World Health Organization Regional Office for Europe. Gatersleben, B. (2007). Affective and symbolic aspects of car use. In T. Gärling & L. Steg (Eds.), Threats from car traffic to the quality of urban life (pp ). Amsterdam: Elsevier. Gärling, T., & Loukopoulos, P. (2007). Economic and psychological determinants of car ownership and use. In A. Lewis (Ed.), Cambridge handbook of psychology and economic behaviour (pp ). Cambridge: Cambridge University Press. Gärling, T., & Steg, L. (Eds.) (2007). Threats from car traffic to the quality of urban life : problems, causes, and solutions. Amsterdam ; Oxford: Elsevier. Groeger, J. A., & Rothengatter, J. A. (1998). Traffic psychology and behaviour. Transport Psychology F, 5, Hair, J. F., Anderson, R. F., Tatham, R. L., & Black, W. C. (2006). Multivariate data analysis (6th ed.). Hillsdale Hagman, O. (2003). Mobilizing meanings of mobility: Car user's constructions of the goods and bads of car use. Transportation Research D, 8, 1-9.

15 Hanson, S., & Schwab, M. (1995). Describing disaggregate flows: Individual and household activity patterns. In S. Hanson (Ed.), The geography of urban transportation (pp ). New York: Guilford Press. Jakobsson, C. (2007). Instrumental motives for private car use. In T. Gärling & L. Steg (Eds.), Threats to the quality of urban life from car traffic: Problems causes, and solutions (pp ). Amsterdam: Elsevier. Jakobsson Bergstad, C., Gamble, A., Hagman, O., Polk, M., Gärling, T., Ettema, D., Friman, M., & Olsson, L. E. (2009a). Impacts of routine out-of-home activities on subjective wellbeing (Göteborg Psychological Reports No. 39:4). Göteborg, Sweden: Department of Psychology, University of Gothenburg. Jakobsson Bergstad, C., Gamble, A., Hagman, O., Polk, M., Gärling, T., Ettema, D., Friman, M., & Olsson, L. E. (2009b). Subjective well-being related to satisfaction with daily travel (Göteborg Psychological Reports No. 39:5). Göteborg, Sweden: Department of Psychology, University of Gothenburg. Mann, E., & Abraham, C. (2006). The role of affect in UK commuters travel mode choice, British Journal of Psychology, 97, McFadden, D. (2001). Disaggregate behavioral travel demand s RUM side: A 30-year retrospective. In D. Hensher (Ed.), Travel behavior research: The leading edge (pp ). Amsterdam: Elsevier. Mokhtarian, P. L. (2005). Travel as a desired end, not just a means. Transportation Research A, 39, Polk, M. (1998) Gendered mobility: Women's and men's relations to automobility in Sweden. Ph.D. dissertation. Göteborg, Sweden: University of Gothenburg, Center for Interdisciplinary Studies. Polk, M. (2004). The influence of gender on daily car use and on willingness to reduce car use in Sweden. Journal of Transport Geography, 12, Swedish Institute for Transport and Communications Analysis (2007). RES : den nationella resvaneundersökningen. Stockholm: Swedish Institute for Transport and Communications Analysis. Steg, L., Vlek, C., & Slotegraaf, G. (2001). Instrumental-reasoned and affective-symbolic motives for using a motor car. Transportation Research F, 4, Steg, L. (2005). Car use: lust and must. Instrumental, symbolic and affective motives for car use. Transportation Research A, 39, Stradling, S. G. (2007). Determinants of car dependence. In T. Gärling & L. Steg (Eds.), Threats from car traffic to the quality of urban life (pp ). Amsterdam: Elsevier.