3.1. Survey method
Data collection in this research was based on the survey among Polish people, preceded by a pilot survey (437 respondents, non-random sampling, May 2020). The questionnaire was amended according to the comments of respondents. Then, the external company was recruited to perform the survey (September 2020). The data was collected using CAWI (computer-assisted web interview) approach in two time periods. A total of 3700 complete and reliable observations were collected, 1700 for October 2020 and 2000 for October 2021. For every round of the survey, the respondents were randomly selected by an external company, which was obliged to stratify the population of Polish citizens according to gender, age and place of living (ca. 60% of urban residents, ca. 40% of country residents including ca. 20% of suburban zones up to 20 kilometres from the city centre). If so, the stratified random sampling method was applied. In the case of age, only four generations were taken into consideration (Baby Boomers – people born between 1945-62, generation X – people born between 1963 and 1980, generation Y – people born 1981 and 1999, generation Z – people born after 2000; only persons aged 16 or more were respondents). The sampling process provided an error parameter at 3% with at least a 95% confidence interval.
The first step of the research involved calculating the intenseness of the millennial values and beliefs in a given person by summarising the number of positive responses on a scale suggested by Sapsford 31,32. The main point of focus was to grasp the inter-generational or inter-gender differences in car enthusiasm and catch if there are differences between the answers given during the second and fourth wave of the pandemic in Poland. The specifics of the millennial generation (generation Y) was mentioned many times in the literature and therefore, was one of the main areas of research in this study because today, this generation sets trends and has the most significant purchasing power 4,33,34.
The main body of the survey consisted of three parts: the characteristics of respondents' opinions, beliefs and attitudes, the mobility choices and individual characteristics about life choices (but anonymized). The modes of transport proposed in the questions were specified based on two chosen approaches 35,36 and referred only to everyday mobility. Together, the questionnaire had 23 closed questions, including 5 scale questions about the preferences and choices of transport modes.
3.2. Building the SEM model
The scale verification was achieved by performing the reliability analysis and calculating the Cronbach’s alpha parameter, which had a value of 0.71 thus making it reliable according to Dimitrov 37. All of the presented statements were described in the literature 38–42 as characteristic to generations Y and Z, significantly different from those presented for older generations: X and Baby Boomers 34,43−47. Because the alpha value was acceptable (see Table 1), as well as global Cronbach’s alpha (0.71), all of the proposed statements were included in further data processing, namely the sum of positive answers to the questions 1.1–1.35 was thus aggregated into one variable – StatMil.
Afterwards, reliability analysis was also performed upon variables KES1-KES6, representing the responses (on a five-point Likert scale) to the statements constituting the Kessler’s 6 item NSPD - non-specific psychological distress scale 48. The reliability analysis showed that this approach is highly acceptable. The Cronbach’s alpha for such a set of variables was 0.90 (see Table 2), warranting the proper use of the scale in further research. The statements were treated together as constitutive for latent variables representing the distress of the respondent at the moment of research.
Other independent variables in the model include:
-
QoL (self-declared perceived quality of life of the respondent on a scale of 1–10),
-
DOM_Car (a binary variable representing the dominant mode of transport used by the respondent: 1 – car, 2 – else),
-
Generation (a nominal variable representing the generation of the respondent: 1 – generation Z (born in 2000 or later), 2 – generation Y (born between 1981 and 1999), 3 – generation X (born between 1963 and 1980), 4 – Baby Boomers (born between 1945 and 1962),
-
Gender (a nominal variable representing the declared gender of the respondent: 1 – female, 2 – male, 3 – other),
-
Urban_01 (a nominal variable representing the area of residence: 1 – urban, 2 – suburban, 3 – rural),
-
Year (a binary variable representing the point of time of the response: 0–2020, 1–2021).
The independent variable was constructed based on the responses to questions 14.01–14.23, representing (on a five-point Likert scale) the respondent's attitude to mobility-related statements. An exploratory factor analysis (EFA) was then conducted to identify the latent variables that had been identified for this study to reduce the number of dimensions and diminish the effect of collinearity. In contrast to an observable variable, a latent variable means a variable that is not directly observed but inferred from other variables. According to Yong and Pearce 49 the “[...] purpose of factor analysis is to summarize data so that relationships and patterns can be easily interpreted and understood.”. Thus, it is often used to regroup variables and retrieve a limited set of factors based on shared variance.
Table 1
Results for reliability analysis – defining Millennials’ attitudes
Variable/statement
|
Mean if deleted
|
Var. If deleted
|
StDv. If deleted
|
Itm-Totl deleted
|
Alpha if deleted
|
1.01. I am in a hurry very often, and I want to have something now, not later
|
22.38
|
18.93
|
4.35
|
0.33
|
0.69
|
1.02. I want to balance my professional and personal life
|
22.23
|
19.64
|
4.43
|
0.21
|
0.70
|
1.03. I am open to people of other races and nationalities
|
22.12
|
20.08
|
4.48
|
0.15
|
0.71
|
1.04. I find myself tolerant
|
22.10
|
20.09
|
4.48
|
0.17
|
0.71
|
1.05. I use instant messaging (e.g. Whatsapp, Messenger)
|
22.11
|
19.85
|
4.46
|
0.26
|
0.70
|
1.06. I question authorities very often.
|
22.64
|
19.28
|
4.39
|
0.24
|
0.70
|
1.07. In life, having fun is more important to me than working.
|
22.77
|
19.22
|
4.38
|
0.29
|
0.70
|
1.08. I am well versed in technical innovations
|
22.40
|
19.29
|
4.39
|
0.24
|
0.70
|
1.09. I often send e-mails, short text messages, messages on social networks
|
22.28
|
19.08
|
4.37
|
0.34
|
0.69
|
1.10. I believe divorce is needed. There is nothing wrong with people getting divorced
|
22.29
|
19.54
|
4.42
|
0.21
|
0.70
|
1.11. I earn money to spend it. I wouldn't say I like saving.
|
22.65
|
19.18
|
4.38
|
0.26
|
0.70
|
1.12. My dream is to earn enough money to buy what I want.
|
22.19
|
19.53
|
4.42
|
0.28
|
0.70
|
1.13. I prefer to consult essential decisions rather than make them myself.
|
22.33
|
19.91
|
4.46
|
0.10
|
0.71
|
1.14. I'd rather call or chat or online messenger than meet face to face (even if I had time to meet)
|
22.74
|
19.29
|
4.39
|
0.26
|
0.70
|
1.15. I want to have access to information and communication channels (e.g. e-mail, messenger) all the time.
|
22.23
|
19.26
|
4.39
|
0.32
|
0.70
|
1.16. I like watching TV
|
22.29
|
19.91
|
4.46
|
0.11
|
0.71
|
1.17. I believe that getting a university degree is very important in life
|
22.34
|
19.79
|
4.45
|
0.13
|
0.71
|
1.18. I don't always follow the rules.
|
22.33
|
19.49
|
4.42
|
0.21
|
0.70
|
1.19. Work is not only an obligation; it is meant to be fun.
|
22.14
|
20.16
|
4.49
|
0.10
|
0.71
|
1.20. I prefer freedom and flexibility in my life than, for example, a safe but boring job.
|
22.39
|
19.38
|
4.40
|
0.22
|
0.70
|
1.21. I like to learn new things; I don't have to have the same job and tasks all my life.
|
22.14
|
19.95
|
4.47
|
0.18
|
0.70
|
1.22. I can do many things simultaneously (multitasking), e.g. drive a car and write a message on Facebook.
|
22.51
|
18.85
|
4.34
|
0.33
|
0.69
|
1.23. Preferably, I would rarely change my workplace.
|
22.29
|
20.14
|
4.49
|
0.05
|
0.71
|
1.24. I prefer reading paper documentation to an electronic one.
|
22.57
|
19.97
|
4.47
|
0.07
|
0.71
|
1.25. I am sceptical of most ideas (concerning various spheres of life).
|
22.63
|
19.50
|
4.42
|
0.19
|
0.70
|
Table 1
Variable/statement
|
Mean if deleted
|
Var. If deleted
|
StDv. If deleted
|
Itm-Totl deleted
|
Alpha if deleted
|
1.26. I have a close circle of friends.
|
22.25
|
19.72
|
4.44
|
0.18
|
0.70
|
1.27. It is difficult for me to keep a job.
|
22.83
|
19.53
|
4.42
|
0.24
|
0.70
|
1.28. I think I am unique and special.
|
22.50
|
18.95
|
4.35
|
0.31
|
0.70
|
1.29. Flexible working hours and private life are essential to me.
|
22.12
|
20.00
|
4.47
|
0.18
|
0.70
|
1.30. I believe that older generations, such as my parents, completely do not understand today's world.
|
22.53
|
19.07
|
4.37
|
0.28
|
0.70
|
1.31. I wouldn't say I like it when someone imposes something on me.
|
22.12
|
20.03
|
4.48
|
0.17
|
0.70
|
1.32. I demotivate myself quickly.
|
22.60
|
19.37
|
4.40
|
0.21
|
0.70
|
1.33. I'm online all the time.
|
22.47
|
18.85
|
4.34
|
0.33
|
0.69
|
1.34. I attach great importance to my image.
|
22.40
|
19.02
|
4.36
|
0.31
|
0.70
|
1.35. I believe that I don't need to have everything.
|
22.19
|
20.42
|
4.52
|
0.00
|
0.71
|
Table 2
Reliability analysis – NSPD scale
KES variable |
|
|
|
|
|
KES1. so depressed that nothing could cheer you up
|
13.93 |
26.47 |
5.15 |
0.77 |
0.88 |
|
13.77 |
27.21 |
5.22 |
0.71 |
0.89 |
KES3. restless or fidgety
|
13.93 |
26.02 |
5.10 |
0.76 |
0.88 |
|
14.08 |
26.10 |
5.11 |
0.77 |
0.88 |
KES5. that everything was an effort
|
13.80 |
27.09 |
5.20 |
0.69 |
0.89 |
|
14.37 |
26.35 |
5.13 |
0.68 |
0.89 |
Original, EFA was developed to reduce dimensions to prepare methods like regression and minimize the risk of collinearity. However, it is also common to use EFA to analyse the internal variability of a dataset before applying other methods, such as regression or structural equation modelling (SEM).
Typically, the Kaiser–Meyer–Olkin (KMO) criterion is verified as a first step of the EFA. Furthermore, Bartlett’s test of sphericity is conducted to verify the possibility of applying EFA to the given dataset. If the test is significant and the eigenvalues that are the basis for the criterion are higher than 1, the extracted factors can be taken into analysis. Classically, the principal component analysis (PCA) method was applied to extract the factors 50. In the next step, the extracted factors were analysed regarding their eigenvalues. Also, the total variance explained as well as the internal structure based on the factor loadings for each item within the factors were looked at. Within PCA, rotation methods are often applied to ease the interpretability; thus, each factor is associated with a reduced block of observed variables 51. The type of rotation applied is based on whether the factors are believed to be correlated (oblique) or uncorrelated (orthogonal). According to the literature, there are four main orthogonal methods: equamax, orthomax, quartimax and varimax, where varimax is the most common one. Therefore, we initially assume that the factors are uncorrelated (the factor analysis results imply no correlation) and apply the most common rotation method, the varimax rotation.
There is an ongoing discussion within the literature about the minimal acceptable value for the factor loadings for the item that is considered significant 52. Regarding this thesis, a factor loading for an item higher than 0.5 is a relatively good representation of a given item within the factor (see Table 3).
This first approach suggests a very dispersed set of responses regarding the mobility statements. Furthermore, the results suggested that the models of higher-order values were not well-represented by the data with the original item structure. As a consequence, two methods were used to improve the model fit until the CFI for the subsequent SEM reached a value of at least 0.90:
1) items with non-significant loadings were excluded from the factor analysis (variables 14.04, 14.14, 14.16, 14.19 were excluded at this point),
2) the item with the lowest factor load was deleted from the first four domains (variables 14.05, 14.23, 14.18, 14.21 were removed). Given that the last factor was fully represented by one-factor loading, its structure was left intact.
This led to a second recalculation of the factor analysis (see Table 4) which resulted in the extraction of 3 significant latent variables connected with mobility:
-
factor 1 representing the intensity of car fondness (the main dependent variable in the model)
-
factor 2 representing the attitude towards shared mobility
-
factor 3 representing the intensity of sustainable and “green” beliefs
The final structural equation model was estimated in SmartPLS 3 software. The list of variables included in the model is presented in Appendix A.
Table 3
The first approach to EFA in regards to the mobility-related statements
Variable
|
Factor Loadings (Varimax raw)
|
Factor 1
|
|
Factor 3
|
Factor 4
|
|
14.01. In my opinion, cars should be environmentally friendly (e.g. fueled by renewable raw materials)
|
0.05
|
0.07
|
0.74
|
0.09
|
0.11
|
14.02. Cars should be more long-lasting than now
|
0.27
|
-0.15
|
0.63
|
-0.05
|
0.12
|
14.04. The car gives unlimited travel opportunities only for drivers
|
0.30
|
0.20
|
0.05
|
0.10
|
0.48
|
14.05. A car makes it easier to visit family and friends than any other means of transport
|
0.65
|
-0.10
|
0.21
|
0.01
|
0.12
|
14.06. A car is just a means of transport for me, nothing else
|
-0.06
|
0.09
|
0.09
|
0.02
|
0.83
|
14.07. I prefer travelling by car than any other means of transport
|
0.66
|
-0.08
|
-0.04
|
0.17
|
0.09
|
14.08. Thanks to the car, you can get to the selected place much faster than by any other means of transport
|
0.62
|
-0.17
|
0.22
|
0.04
|
0.14
|
14.09. The car supports self-realization and the realization of one's own goals
|
0.74
|
0.09
|
0.16
|
0.06
|
-0.02
|
14.10. The car facilitates entertainment (cinema, theatre, pub)
|
0.72
|
0.10
|
0.00
|
0.13
|
-0.01
|
14.11. A car is associated with freedom and independence
|
0.71
|
-0.01
|
0.25
|
0.04
|
-0.09
|
14.12. I would like to travel cheaply by car, even if not comfortable
|
0.52
|
0.28
|
-0.04
|
0.11
|
0.29
|
14.13. Owning a car contributes to a higher level of happiness
|
0.75
|
0.28
|
-0.06
|
0.08
|
-0.13
|
14.14. The car you own shows your standard of living and material status
|
0.55
|
0.42
|
-0.15
|
0.03
|
-0.04
|
14.15. I like the concept of car sharing or car rental by minutes/hours
|
0.11
|
0.73
|
0.20
|
0.01
|
0.07
|
14.16. It is possible to live without a car
|
-0.17
|
0.22
|
0.34
|
-0.08
|
0.29
|
14.17. If the public carrier offered a shared car as one of the means of transport, I would use such a solution
|
0.11
|
0.77
|
0.08
|
0.15
|
0.03
|
14.18. I believe that introducing shared bikes is a good idea
|
0.12
|
0.35
|
0.61
|
-0.11
|
-0.12
|
14.19. Public transport works quite well in the area where I live
|
-0.04
|
0.46
|
0.34
|
-0.30
|
0.07
|
14.20. My main concern is the lack of punctuality of public transport
|
0.12
|
0.25
|
-0.01
|
0.74
|
0.03
|
14.21. The crowd disturbs me a lot in public means of transport
|
0.19
|
0.03
|
0.24
|
0.69
|
0.04
|
14.22. I would like public transport to be greener
|
0.05
|
0.14
|
0.67
|
0.36
|
0.07
|
14.23. If the public transport offer included a car rented for minutes, I would give up driving my own car
|
-0.01
|
0.73
|
-0.04
|
0.21
|
0.17
|
|
4.23
|
2.60
|
2.32
|
1.42
|
1.25
|
|
0.19
|
0.12
|
0.11
|
0.06
|
0.06
|
Table 4
The second approach to EFA in regards to the mobility-related statements
Variable
|
Factor Loadings (Varimax raw)
|
|
|
|
14.01. In my opinion, cars should be environmentally friendly (e.g. fueled by renewable raw materials)
|
0.03
|
0.10
|
0.83
|
14.02. Cars should be more long-lasting than now
|
0.25
|
-0.16
|
0.67
|
14.06. A car is just a means of transport for me, nothing else
|
-0.10
|
0.25
|
0.31
|
14.07. I prefer travelling by car than any other means of transport
|
0.70
|
-0.02
|
-0.02
|
14.08. Thanks to the car, you can get to the selected place much faster than by any other means of transport
|
0.63
|
-0.13
|
0.24
|
14.09. The car supports self-realization and the realization of one's own goals
|
0.76
|
0.08
|
0.15
|
14.10. The car facilitates entertainment (cinema, theatre, pub)
|
0.73
|
0.13
|
0.00
|
14.11. A car is associated with freedom and independence
|
0.73
|
-0.05
|
0.22
|
14.12. I would like to travel cheaply by car, even if not comfortable
|
0.52
|
0.35
|
0.07
|
14.13. Owning a car contributes to a higher level of happiness
|
0.75
|
0.23
|
-0.05
|
14.15. I like the concept of car sharing or car rental by minutes/hours
|
0.09
|
0.72
|
0.19
|
14.17. If the public carrier offered a shared car as one of the means of transport, I would use such a solution
|
0.12
|
0.80
|
0.06
|
14.20. My main concern is the lack of punctuality of public transport
|
0.23
|
0.43
|
0.00
|
14.22. I would like public transport to be greener
|
0.07
|
0.23
|
0.74
|
14.23. If the public transport offer included a car rented for minutes, I would give up driving my own car
|
-0.01
|
0.80
|
0.02
|
|
3.53
|
2.36
|
1.94
|
|
0.24
|
0.16
|
0.13
|