Project Blogs / South Africa

Project Blog: South Africa

SA_Overcrowding - Metrorail

Research Diary

To study mobility in South Africa we interviewed train users in the Western Cape, South Africa. In many ways, the mobility encounters of Metrorail users (shown in the above picture) represent the worst challenges a daily commuter could encounter and our study attempted to better understand what mobility means to people who are continuously confronted with difficulties including but not limited to delays, breakdowns, cancellations, as well as threats of violence, theft, and vandalism. An example of what such a reality looks like can be seen in the movie Tsotsi (2005). Here, also a link to an example of an interview: 20130123 SA interview transcript South Africa

After completing our study of Metrorail Commuters in the Western Cape, we published our findings in two peer-reviewed articles. The first was published on 15 March 2019 in the journal Frontiers in Psychology and the second on 18 May 2019 in the journal Sustainability. Under the Creative Commons License, we reproduce them here with links to the original publications.


The following article can be cited as follows:

Bergman, Z., Bergman, M.M., & Thatcher, A. (2019). Agency and Bandura’s Model of Triadic Reciprocal Causation: An Exploratory Mobility Study among Metrorail Commuters in the Western Cape, South Africa. Frontiers in Psychology, 10.


Most studies on sustainable mobility focus on technological, socio-structural, or psychosocial influences while neglecting individual motivations and practices. In this study, we examine mobility motivations and practices as part of a complex interplay between psychosocial and socio-structural dimensions within the mobility infrastructure of Metrorail in the Western Cape. Drawing on Albert Bandura’s theory of personal agency and the model of triadic reciprocal causation, we interviewed 38 commuters (mean age 33 years, SD 11, 50% women/men) and analysed the data using hermeneutic content analysis and multidimensional scaling. Based on our analyses, we identified three pathways that describe the mobility practices of Metrorail users, each with its own purpose and function. We explore these pathways and their consequences for sustainable mobility in relation to daily commuter agency, motivations, and past experiences.

Keywords: Sustainable mobility, Albert Bandura, agency, triadic reciprocal causation, mixed methods, hermeneutic content analysis, content configuration analysis, Metrorail.


Despite decades of innovations and interventions, the transport sector still accounts for approximately one-sixth of greenhouse gas (GHG) and CO2 emissions (IPCC 2007, 2014). Consequently, mitigating environmental, health, and social risks caused by mobility practices remain a crucial challenge (Bandura 2008; Geerken et al. 2009; Guadagno 2016; Boas 2017; Yamamoto, Serraglio, & de Cavedon-Capdeville  2018). More sustainable mobility practices would mean “to reduce the need to travel (less trips), to encourage modal shift, to reduce trip lengths and to encourage greater efficiency in the transport system” (Banister 2008, p. 75). Such solutions tend to focus on either technical and socio-structural changes or psychosocial interventions.

Technical and socio-structural approaches aim to mitigate GHGs by increasing the efficiency of transport systems. Known as hard policy approaches, they seek to remodel transportation systems through technological and infrastructure development (Novaco 2001; Brög, Erl, & Mense 2004; Hunecke et al. 2007; Gehlert et al. 2013). Current green technology solutions include mass public transit, high-speed rail, shared and autonomous mobility systems, and electric vehicles.

Hard sustainable mobility policies tend to assume availability and affordability of raw materials, industrial capacity, and extensive investment for the development, implementation, and maintenance of disruptive technologies. Also assumed is that, for example through incentives or taxes, the public could be enticed to support the necessary economic, political, and cultural changes that are part of the disruptive technology.

Psychosocial approaches focus on individuals or groups to improve the sustainability of mobility (Brög et al. 2004; Stanton et al. 2013). At the centre of the so-called soft approach are individual or collective attitudes, values, norms, motivations, preferences, habits, and behaviours with the goal of creating modal shifts in why and how individuals or groups travel (Novaco 2001; Steg & Vlek, 2009; Gehlert et al. 2013). Public appeals and awareness campaigns are currently the main tool to increase knowledge and acceptance of sustainable mobility (Brög et al. 2004; Hunecke et al. 2007). Included under this rubric are approaches that focus on inequality (Titheridge, Mackett, Christie, Oviedo Hernàndez, & Ye 2014; Zhao & Li 2016), inclusiveness (Bergman et al. 2014; OECD 2016), access (Starkey & Hine 2014; Bergman & Bergman 2015; World Bank 2016), and ecological behaviour (Collado, Staats, & Corraliza 2013; Lokhorst, Werner, Staats, van Dijk, & Gale 2013; Pillemer et al. 2017; Landry, Gifford, Milfont, Weeks, & Arnocky 2018). Compared to hard policy measures, psychosocial interventions tend to enjoy a greater degree of political support because they can be implemented at significantly lower cost and with fewer systemic disruptions (Stanton et al. 2013). However, psychosocial interventions encouraging modal shifts have had limited success partly because initiatives and policies tend to emphasise how people travel, not why they travel (Cass & Faulconbridge 2016), and because they often neglect environmental constraints, such as shortcomings in mobility infrastructure or other structural barriers (Bergman et al. 2014; Bergman & Bergman 2015).

With few exceptions (e.g., van Wee et al. 2002; Poortinga et al. 2004; Collins & Chambers 2005; Hunecke et al. 2007; Steg et al. 2014), most mobility studies focus on either technical and socio-structural or psychosocial dimensions of mobility. Hunecke et al. (2007), for example, examined the effect of psychological, socio-demographic, and infrastructure influences on the ecological impact of mobility behaviour. Other notable studies include Poortinga et al. (2004), who analysed the impact of value dimensions concerning different psychological and environmental aspects on home and transport energy usage, and Steg et al. (2014), who studied the impact of values, situational cues, and goals to encourage pro-environmental behaviour. Finally, van Wee et al. (2002), Collins and Chambers (2005) and Hunecke et al. (2007), explored psychological, sociodemographic, infrastructural, or situational effects on mobility behaviour. Despite diverse foci, these studies arrived at similar conclusions: mobility practices are nested within a complex network of individual, social, and environmental factors, and the consensus seems to be that we need to better understand the synergy between individuals and their mobility environment in order to exploit the potential of behaviour change toward more sustainable mobility (Shepherd & Marshall 2005). Given the relative neglect of the interdependence between commuter motivations and practices in specific mobility environments, little is known about what a sustainable transportation system would look like, what criteria should be used to assess it, or what would make it socially and culturally acceptable (Steg & Gifford 2005). Charlton (2004, p. 165) argues that “[u]nless these complex, interrelated socio-economic and behavioral influences can be properly interpreted and, crucially, incorporated into policy and practice, genuine advances to greater sustainability will be elusive.”

In this article, we first present Albert Bandura’s triadic reciprocal causation as a suitable theoretical framework that can account for the complex interdependence among mobility intentions, practices, and the environment in which they are embedded. We then present an empirical application of this framework to analyse mobility practices of train commuters in the Western Cape, South Africa. The overall aim of this article is to contribute to a better understanding of individual agency within specific mobility environments in order to improve conceptualisations and implementations of sustainable mobility solutions.

Theoretical Background

We encounter multiple and constantly changing environments each day, requiring a vast array of choices. Despite ever-changing dynamics, we manage to negotiate a highly complex world because our behaviours are neither hardwired nor mere products of our environment. As active agents, we influence outcomes, we act upon others’ behaviour, and we coordinate behaviours with each other (Bandura 2006).

Social cognitive theory (SCT) as developed by Albert Bandura proposes that human behaviour encompasses core features that include not only internal behavioural predispositions, such as cognition, affect, or motivation, but also various environmental influences (Bandura 2001). SCT terms deliberative behaviour personal agency, which has been studied extensively in psychology (Bandura 2006, 2008), public health (Bandura 2004a), education (Rogers et al. 1999; Chapman-Novakofski & Karduck 2005), business and management (Schmutzler, Andonova, & Diaz-Serrano 2018; St-Jean, Radu-Lefebvre, & Mathieu 2018), and media studies (Bandura 2004b; Gibson 2004). Personal agency refers to an individual’s ability to “designedly conceive unique events and different novel courses of action [while choosing] to execute one of them” (Bandura 2001, p. 5). It includes complex processes of intra-personal cognitive processing, deliberation, and decision-making, motivated by a desire to achieve specific outcomes. According to Bandura, desires shape our intentions to act, thus preceding behaviour toward goals or aspirational ends. Subsequently, desires to achieve an end serve as the impetus for, and the intended outcome of, our actions. The process of turning intentions into goals involves a number of decision-making strategies. The first relates to three modes of agency: individual, proxy, and collective. According to Bandura (2001), individual agency entails the process whereby people deliberately guide their behaviour within an immediate environment. If the goal is to get to work, for example, we may elect to drive our car or ride bicycles. Individual agency has its limits because individuals may not always be able to act on their own behalf. Children, for example, are unable to drive cars, and they may not own or be allowed to ride their bicycle to school. In this case, agency without the assistance of others is impeded. Proxy agency involves enlisting others to act on our behalf to secure desired outcomes. Collective agency refers to collective efforts to achieve a desired outcome through interdependence and the activation of networks (Bandura 2008). This might entail organising a car sharing club or petitioning local politicians to fund a public transit system. Each mode offers a different way to achieve a goal, and despite cross-cultural variations, we rely on all three modes of agency to conduct our lives (Bandura 2001).

Agency is mediated by contextual and cultural influences, such as “activities, situational circumstances, and socio-structural constraints, and opportunities” (Bandura 1999, p. 6). It is preceded by an assessment of opportunities and constraints inherent in socio-structural or contextual environments (Bandura 2001). Environments are assessed and perceived to facilitate or hinder the ability to act. Car ownership and lack of access to public transport are examples of components of mobility environments that facilitate or obstruct mobility pathways. Bandura (2001) distinguished between three environments, namely the selected, the constructed, and the imposed environment. The selected environment provides the largest scope of behaviour and therewith the broadest agentive space. Here, individuals are agents of their realities, they have at their disposal a range of different behavioural options, and they can choose behaviours that best suit a desired outcome in a specific situation. By choosing “associates, activities, and milieus,” environments are selectively activated as individuals formulate appropriate courses of action and decide how to behave (Bandura 1999, p. 6). In terms of mobility practices, a selected environment may include access to mobility modes, such as a car, bus, or train. The modal choice reflects whatever is perceived to best achieve a desired outcome. The constructed environment requires concerted effort to become a viable agentive option. It restricts agentive practice because it requires “people to construct social environments and institutional systems through their generative efforts” (Bandura 1999, p. 6). Examples include arranging a ride in a car sharing club to get to work, campaigning for public transport systems to be extended into a township, or relocating to reduce the distance to a train station. The imposed physical and socio-structural environment narrows the scope of agency because it dictates the boundaries within which people behave and, although “they have little control over its presence, they have leeway in how they construe it and react to it” (Bandura 1999, p. 6). For example, walking long distances to school as the only form of available mobility reflects an imposed environment. Agency still exists in which the pupil may choose whether to attend school on a given day, or which route to take to avoid anticipated hazards.

The modes of agency and their environments are interdependent. According to Bandura (2006 p. 6), “internal personal factors in the form of cognitive, affective, and biological events, behavioural patterns, and environmental influences all operate as interacting determinants that influence one another.” Derived from SCT, Bandura’s model of triadic reciprocal causation (Bandura 1986; see also Bandura 1989, 1999, 2001, 2006) emphasises that personal agency is inherently psychosocial and functionally dependent on events. Accordingly, agency may be presented as follows:

Figure 1. Model of Bandura’s personal agency and triadic reciprocal causation.

Figure 1models individuals’ intentions to achieve desired outcomes. Through complex processes of intra-personal deliberation, individuals assess how various environments (selected, constructed, or imposed) facilitate or constrain their potential to act (action potential), as well as how different modes of agency (individual, proxy, or collective) enable them to achieve their goal. Based on deliberations within environments, individuals choose the mode of agency (individual, proxy, or collective) that will most likely secure a desired outcome in a specific context. An appropriate course of action is then selected and implemented as people adjust their behaviour accordingly.

Given the variability of options and conditions, it follows that no fixed, predictable pattern of reciprocal interaction exists (Bandura 2006). The uniqueness of the constellation of agency in a given environment makes agency inherently difficult to study. One of the main criticisms aimed at Bandura’s work relates to the relative looseness of the concepts and their interdependence (Tschannen-Moran & Hoy 2001; Garvis & Pendergast 2016). Others argued that the conceptual overlap between expectations (intentions) and outcomes limits the applicability of the theory (Eastman & Marzillier 1984). Consequently, most SCT studies reduced agency to individual agency, focusing predominantly on self-efficacy to model behaviour change, such as in the field of health, especially therapeutic research (Bandura 1997; Langlois et al. 1999), preventative health (Tougas et al. 2015), public health education (Ryerson 1994; Rogers et al. 1999; Bandura 2004a; Chapman-Novakofski & Karduck 2005), education (Bores-Rangel et al. 1990; Church et al. 1992; Hackett & Byars 1996), and media studies (Gibson 2004; Hill et al. 2009). The selective focus on intra-personal, cognitive dimensions of individual agency limits accounts of human behaviour in situations that transcend the confines of unidirectional modes of causation (Bandura 1999). The fact that most studies have applied only one type of agency from SCT (Carillo 2010) and the consequences of this one-sided application represents the most compelling critique against studies on personal agency.

We seek to expand the conventional, unidirectional application of SCT by exploring the multidimensional nature of personal agency as initially formulated by Bandura and by applying triadic reciprocal causation to mobility practices of Metrorail commuters in the Western Cape. The rationale for this study are three-fold. First, we aim to study human agency using the model of triadic reciprocal causation to expand the applicability of Bandura’s theory. This means situating individual, proxy, and collective components of agentive practice within psychosocial and socio-structural environments. Second, we apply this multidimensional concept of agency to study mobility practices. Specifically, we propose to analyse agency and mobility practices in the context of Metrorail commuters in the Western Cape, South Africa. Metrorail is the largest commuter train service in South Africa, transporting approximately 2 million people every day on 2228 km of track. The local Metrorail network in the Western Cape region has been operational since 1863. It consists of four main lines – a Northern Line, Southern Line, Cape Flats Line, and Central Line with 610 km of track and 119 stations that connect informal settlements, townships, suburbs, towns, and cities in the South Western Cape. Third, by exploring mobility with Bandura’s multidimensional approach to agency, we hope to contribute to a debate on sustainable mobility that goes beyond interventions, which focus on either technical and socio-structural or psychosocial manipulations. In other words, Bandura’s theory of agency will be used in a case study to argue for a reciprocal relationship between technical, socio-structural, and psychosocial effects on mobility behaviour. Our rationale translate into three research questions:

  1. Can we empirically identify the agency and environment dimensions outlined in Bandura’s model of triadic reciprocal causation in the narratives of Metrorail commuters?
  2. How do the dimensions of agency and environment interrelate in the reported mobility practices of Metrorail commuters?
  3. What are the implications of conceptualising agency accordingly on the understanding of sustainable mobility systems for Metrorail commuters?

Materials and Methods


This study is based on 38 narrative interviews with Metrorail commuters in the Western Cape. Three selection criteria assisted in identifying eligible participants: mobility type (use of Metrorail), frequency (week-day commutes during that past 2 years), and geographical location (multiple commutes per week in the wider Cape Town or Stellenbosch region). The interviewed men (n = 19) and women (age range 18–62 years, mean 33 years, SD 11) were multi-ethnic (black, white, and colored – In South Africa, the term ‘colored’ is colloquially used to denote people from a mixed race background.), multi-lingual (speaking predominantly Afrikaans, isiXhosa, and English, as well as Tswana, Sesotho, Sotho, and isiZulu), and pursued a variety of occupations (students, teachers, security guards, shop attendants, cleaners, drivers, administrators, couriers, repair men and women, managers, occupational therapists, personal assistants, and unemployed). All recruited participants took part in the interview. The coding of the interviews yielded 784 codes for the multidimensional scaling (MDS) analysis, which is regarded as an adequate sample size for a dimensional analysis (de Winter, Dodou, & Wieringa 2009). Considering the small size of the sample, we could not control for the effect of individual differences.


Before commencing data collection, we obtained permission to conduct the research from the University of the Witwatersrand Human Research Ethics Committee. Based on our sampling criteria, participants were recruited near train stations in Cape Town and Stellenbosch. Interviews were conducted immediately, or at an arranged time near the station (in public spaces or cafes), or at a venue negotiated between the interviewer and the participant. All requests for interviews were accepted. Interviews were conducted in English or Afrikaans (the two dominant languages in this region). The interviews averaged approximately 40 min. All interviews were recorded, transcribed, and anonymised for analysis.


The interview schedule was developed and refined during two pilot phases with members from the research population. The interview schedule included exploratory and semi-structured questions. Initially, exploratory questions aimed to elicit extended narrative responses from interviewees regarding their mobility experiences. Question included “Tell me everything that comes to mind when you think about trains” or “What is your best memory with a train?” These were followed by semi-structured questions aimed to prompt specific mobility preferences or to examine mobility dimensions in detail, such as “When, where, and how often do you take trains?” or “What do you think will happen with trains in the future?”


Data were analysed using hermeneutic content analysis (HCA; Bergman 2010), a three-step mixed methods approach. First, interviews were analysed using content configuration analysis (CCA; Bergman 2011; Bergman & Bergman 2011). CCA is a qualitative method used for the systematic analysis of non-numeric data, closely related to qualitative content and thematic analyses (Bergman, Bergman, & Gravett 2011). For this article, interview data were coded top-down, using the dimensions of personal agency as proposed by Bandura. Due to the mobility focus of our study, we also coded intentions and outcomes that pertained to mobility within mobility environments. The coding scheme was developed and applied in a research team. Two independent coders applied the coding scheme iteratively until the emergent coding taxonomy stabilised. The purpose of the initial CCA was to trace dimensions of agency and environment as outlined by Bandura in the context of Metrorail commuters in the Western Cape. In the second analytic step, we identified agentive pathways and mobility environments, using MDS, which enabled a geometric representation of co-occurrences between agency and environment dimensions. We calculated similarity matrices using the Jaccard Index based on thectar (Berger forthcoming) and smacof (Mair et al. 2015) in R. The unit of comparison was at code-level (n = 784), and the parameters included a non-metric procedure with a primary approach to ties. Stress was at 0.11, which is considerably lower than the stress level for a random sample of the same number of points in MDS, estimated at 0.24 (Spence 1979). A two-dimensional map was found to be the most parsimonious and interpretable solution. Adding an additional dimension did not significantly improve stress but worsened interpretability and parsimony. The third and final step of HCA consisted of a re-contextualising qualitative analysis to connect the MDS structures to the interview data, again using CCA. This step helped interpret the meaning of the MDS patterns by referring back to the interview data in which the MDS structures were embedded.


In our analyses, we identified the intra- and inter-personal, as well as socio-structural environmental dimensions that delineated mobility practices, explored the relations between dimensions of agency, and conceptualised “mobility as agency” from the perspectives of Metrorail users.

Dimensions of Personal Agency

In the first step, we coded and analysed the interview data deductively to explore Bandura’s tripartite agency concept. This entailed identifying intentions, types of agency, environmental facilitators and constraints, action potentials, and desired mobility outcomes. The following example illustrates this analytic step:

I just want to walk down there this afternoon and hope the train will be on time. … No, honestly, no man, it’s a headache I tell you. No, I don’t even want to think about it. Because you see actually I leave here at 17:15, right? The train is actually supposed to be there at half past 5, but there is no way that I will waste my time and walk quickly because I know it will either be late or it would have left already. Do you see? So then I rather take the 6 o’clock train. Even if that means I only get there by 18:30. On the other side, the whole fact of the matter is that you need to get home. It doesn’t matter what time you leave here, you simply need to get home. How you are going to get there, what time you will get there, that is simply your own damn problem. And it shouldn’t be like that. It really shouldn’t be like that. (MT3: 3)

Intentions and Desired Outcomes: In this excerpt, agency in relation to mobility began with intentions to be mobile. According to Bandura, intentions consist of the intra-personal cognitive processing of personal needs or desires in relation to anticipated contextual factors and potential desired outcomes. The intention “to get home” connected to what the interviewee perceived as the most significant contextual issue, the unreliability of the train, because it had implications on when he would get home. This assessment enabled him to identify the best course of action – to take a later train. The overarching goal here was to align his intentions with perceived outcomes. Bandura proposed that desired outcomes relate to the extent to which intentions may be realised. Deliberation and making decisions based on when to leave work, when to reach the station, and which train to take assisted this commuter in achieving his goal or desired outcome.

The Environment as a Facilitator or Constraint: The ability to be mobile is mediated by environmental factors. They dictate mobility boundaries and enable or prevent agentive practices. The excerpt above exemplified a late train as a contextual factor that represented an obstacle to the interviewee’s ability to be mobile. This situational circumstance illustrates facilitating or impeding environmental factors mentioned by most Metrorail users in this study. Table 1below summarises the environmental facilitators and constraints from our interviews.

Table 1.Examples of environmental factors mentioned by Metrorail users.

Two characteristics are noteworthy in this table. First, environmental facilitators and constraints lie on opposite ends of a dimension, for example, a train that was on time and a train that was cancelled, respectively. Second, interviewees identified many more constraints than facilitators, a predominant trend in the data since the constraints that restricted agentive practice were far more prevalent not only in frequency but also in terms of perceived significance and degree. This means that our interviewees focused overwhelmingly on experiences associated with constrained mobility environments. The previous and following excerpts illustrate environmental constraints:

I mean we pay, even though we pay less but we pay. There are so many commuters. We buy so many monthlies [monthly tickets]. How much money does Metrorail make? Why can’t they do that? Why can’t they give us a comfortable, convenient environment to sit in? (MT1: 6)

Another noteworthy dimension underlying environmental concerns is that facilitators tended to be aspirational, hypothetical, or future-oriented, such as a planned expansion of train line, in contrast to environmental constraints, which were presented as common experiences, such as unsafe and unreliable trains during rush hour.

Selected, Constructed, and Imposed Environments: According to Bandura, an important feature of environmental facilitators and constraints concerns gradation of variability. Contextual dimensions are imposed, constructed, or selected. The train delay from the first excerpt was an example of this: the commuter may have decided to take an earlier or later train – an instance of bounded agency – but he lacked alternative modal choices. He does not own a car and cannot afford alternative modes of transport. Consequently, he lacked the ability to select or construct a different mobility environment. This illustrated the impact of an imposed environment since his environment and access to resources dictated the boundaries of his action potential. In this way, environmental facilitators and constraints impose a range of variability within which individuals can respond. All three environments shape actual and potential mobility options.

Individual, Proxy, and Collective Modes of Agency: The modes of agency in the previous excerpts related to perceived abilities to be mobile based on a relational dependence. Bandura termed this proxy agency – when others act on the agent’s behalf. Here another example of proxy agency:

So even now, if the car were to break down, I wouldn’t even take a taxi. I would just call someone to come and fetch me. Like my nephew or someone. I wouldn’t walk or take a taxi, not unless I really have no other choice. But that just goes to show how convenient and comfortable my life has become. (MT9: 1)

The interviewee did not elect walking or using a taxi because she considered these modes inconvenient or unsafe in relation to another option. She constructed an alternative mobility option by enlisting someone else, a proxy agent (“I would just call someone to come and fetch me. Like my nephew or someone.”). In our data, the most frequently mentioned proxy agent was Metrorail. For example, an interviewee wished he could rely on this proxy to act on his behalf [“How you are going to get there, what time you will get there, that is simply your own damn problem. And it shouldn’t be like that. It really shouldn’t be like that.” (MT3: 3)]. Commuters frequently expressed their desire for Metrorail to act as a proxy agent to improve their train experiences, such as requests to increase the frequency and the reliability of trains, to enhance the convenience and comfort of trains, or to improve safety and security.

Individual agency relates to instances where commuters deliberately guided their behaviour via mobility options at their disposal. Some commuters reported that Metrorail was their only mobility option, while others were able to limit train use to weekday commutes and made use of alternative mobility modes in other life spheres. In some cases, the obstacles commuters encountered resulted in abandoning Metrorail. For most, Metrorail was the least preferred mode of mobility and the first to be replaced, if other modes became available. Here an example:

So, I take [Metrorail] regularly. Yea, yea, yes, I take it Monday to Friday, weekends I don’t bother with the trains at all, like I’ve told you. We prefer to take the vehicle on the weekend of course it is going to work out more expensive but you can do so much more with the vehicle because then you can do your shopping and things like that. You see, because I am actually one of those fortunate ones because those other people have to also do their whole shopping with the trains, right? They are not as fortunate as some of us. But of course it costs a lot of money. (MT3: 6)

Collective agency refers to acts of interdependent effort that enabled individuals via groups or a collective to achieve a goal. In the case of our Metrorail commuters, collective agency referred mostly to the activation of social ties, often based on religious, friendship, or work networks that developed during train commutes. Here, an example:

…the positive thing that I learned out of [being unable to afford a car] was that God wanted to place me among people because He knows my heart and He knew that I have a need that burns inside of me to serve Him, and this is why I was short of money. But I have become richer in Him because now I have a social group that I have every day, they can feed me, they can give me provisions for the road, they can comfort me and this is really the thing that stands out the most for me about taking the train because I learn every day and I realise every day and I become wiser every day through them because I take the train. (MT1: 4)

Table 2 summarises the modes of agency of Metrorail users.

Table 2.Summary of the modes of agency mentioned by Metrorail users

Action Potential: Modes of agency, environmental facilitators, and constraints and the ability of commuters to assess agency according to their environment relative to their capabilities, intentions, and desires contributed to action potential. The action potential manifested positively or negatively, depending on how these dimensions combined. Positively framed, the reciprocal interaction between agentive and environmental dimensions enabled commuters to achieve their mobility goals. Negatively framed, some commuters were unable to overcome obstacles to mobility, based on environmental constraints or a lack of agency. Their action potential was restricted and their mobility desires remained unfilled.

In this analysis, we connected Bandura’s proposed dimensions of agency and environment to the mobility practices of Metrorail commuters and found that all dimensions were present in the narratives of Metrorail commuters. Next, we examined the interdependence these dimensions.

Systematising the Interdependence of Dimensions of Personal Agency and Environment in the Reported Mobility Practices of Metrorail Commuters

The narratives on mobility experiences were composed of unique constellations of intentions and goals, facilitating or constraining environmental factors, and modes of agency. While the first set of analyses examined the presence of dimensions as outlined by Bandura’s triadic reciprocal causation model, mapping them systematically deepened our understanding of agency in a specific mobility environment. To do this, we used a dimensional analysis, specifically MDS, to map the relations between Bandura’s agentive dimensions and to visualise mobility structures in an n-dimensional space. The representation of relations was facilitated by dividing the dimensions’ action potential and desired outcome into positive or negative constituents – action potential positive and action potential negative, and desired outcome achieved and desired outcome impeded. Mapping patterns of agency in a specific mobility environment revealed distinct patterns of reciprocal interaction between agentive practices and environments. We present this in Figure 2.

Figure 2. MDS map of the agentive practices of the Western Cape Metrorail commuters.

Represented in Figure 2 are the dimensions of agency and environment as points in a two-dimensional space. The distance between points represents the relative frequency or co-occurrence of dimensions in the interview data. The closer the points are located to each other, the more frequently the dimensions co-occurred. Conversely, the further apart these are, the fewer the co-occurrences, and the more orthogonal the dimensions are to each other. Consulting the interview transcripts assisted the interpretation of this map.

According to Figure 2, the agentive practices of Metrorail commuters are divided into two main clusters: a small cluster on the left, which we will refer to as cluster 1, and a larger cluster on the right, cluster 2. Linking the top 30% of co-occurrences with a straight line visualises the two-cluster structure. The second notable feature relates to the shape of the clusters. The dimensions of cluster 1 are situated in close proximity to one another. Each component connects to all others in this cluster. Except for one dyad, the points in this cluster are roughly equidistant to each other. The elongated, crescent shape of cluster 2 on the right indicates that the dimensions in this cluster are connected. In contrast to cluster 1 there are some interpretable differences in this cluster, given the relative distance between the dimensions situated at the top and bottom of the crescent. Finally, the elongated shape of cluster 2 is approximately equidistant to cluster 1. Five notable findings can be inferred from this map.

First, cluster 1 includes four dimensions: environmental impediments, a negative action potential, an imposed environment, and the impediment of a desired outcome. This is interesting because the cluster contains all dimensions, which restrict agency (imposed, impeded, and negative). These dimensions are highly interdependent, given not only the geometric proximity of the points but also their connectedness. This cluster is geometrically and thus conceptually different from the other agency and environment dimensions. Based on the content and location of this cluster, we observed that the challenges and obstacles that impeded mobility agency and practices of the Metrorail users were intertwined. These included references to their imposed mobility environment, such as dilapidated infrastructure or inadequate services, as well as environmental impediments, such as delays and breakdowns in the system. The consequences connected to these obstacles were negated agentive practices and inhibited desired outcomes as they resulted in restricted agentive practice. Given the relative distance to personal, proxy, or collective agency, this cluster represents the opposite of agentive practice – the lack of agency. The following are two excerpts to illustrate the nature of cluster 1:

There were some days, sometimes when people have stolen the power cables. So then, people can’t go to work for at least a day or will be late by two or three hours. So yes, I think this is actually a terrible experience, especially if there is work to be done. They steal the cables a lot, yes. (MT2: 2)

For example, yesterday morning. My train is at 5:45AM. No announcements, nothing. The train arrives at 6:30AM. Do you see? Now I have to let the people at work know that the train is late, but they don’t understand. It’s very frustrating. (MT3: 1)

Second, individual agency is located at the top point of the crescent in cluster 2. It connects to intentions to be mobile and the selected environment. Both dimensions are connected to a positive action potential and achieving a desired outcome, while a selected environment is further connected to a facilitating environment. Three characteristics can be observed in this part of the figure: 1) Individual agency seems to be a cornerstone of a network of connections between mobility dimensions. This part of the constellation implies that individual agency involves a number of core features including intentions to be mobile and the ability to select an environment that facilitates the action potential of the individual to achieve a desired outcome. 2) The dimensions of this part of the cluster are active and positive manifestations of agency. 3) This agentive pathway is similar to the classical understanding of personal agency, where individuals (in our case, commuters) are full agents in the sense that they have different (mobility) options. Here an illustrative excerpt:

I mean, it’s so many commuters, I mean, me coming from the sustainable side, that’s what I’m all about that, I’m all about green living. I prefer using public transport. I have a vehicle but coming to work, I use public transport. Even if it means a taxi or bus, I use it because, I mean, more people in one vehicle, automatically we’ll be saving the environment slowly. (MT24: 1)

For this commuter, individual agency shaped mobility. It transcended commuting from point A to B based on imposed options. Instead, it included an overarching goal of living a greener, more sustainable life, resulting in a concerted effort to utilise mobility options that made this a viable agentive option. Accordingly, the commuter adapted her mobility choices and selected an environment that best suited her desired outcome.

Third, proxy agency is located at the lowest point of cluster 2, which is closely associated with a constructed environment. Proxy agency is also connected to environmental facilitators and achieving a desired outcome. Similar to the constellation located at the top, we identified an agentive pathway at the bottom of this cluster. In this part of the constellation, proxy agency is associated with the ability to construct an environment on behalf of commuters. Interestingly, this agentive pathway is not directly connected to the action potential of commuters, a point we will return to later. The examples mentioned previously that related to Metrorail acting on behalf of commuters, or reaching out to friends and families during emergencies, were indicative of this agentive pathway. Here another example:

Stick to time, and send out notices, like if they know people are using trains regularly, like if there’s a delay, send SMS’s to people, you know, be like “the trains will be delayed like an hour,” like give me a choice, help me decide if I want to take the train or a taxi, maybe I could’ve compromised or made another option, but now I don’t know, I get there, now I wait, and it’s like five minutes, then you wait, then it’s like forty minutes, you know, so yeah. Like, let people know. (MT32: 3)

Fourth, the final type of agency in Figure 2, collective agency, is located near the centre of cluster 2. It is connected to three agentive dimensions, namely the construction of an environment, a positive action potential, and achieving one’s desired outcome. This constellation implies that collective agency consists of a co-construction of mobility environments through interdependent efforts, which increase the potential for agency. This agentive pathway is interesting for several reasons. In contrast to individual and proxy agency, which are situated at opposite ends of this cluster, collective agency is located centrally, close to achieving one’s desired outcome. The proximity of these two dimensions indicates that this agentive pathway is closest to Metrorail users achieving their desired outcomes. Nearly equidistant from proxy agency and individual agency, it also implies that it shares some characteristics with these. It appears that collective agency is made up of individual and proxy agency while being more effective than the single individual or the proxy agent.

Fifth, and as a consequence of the above, it is more appropriate to think of this agency cluster not as a crescent but rather as a continuum, where collective agency represents the mid-point between individual and proxy agency. One way to make sense of this is to consider the locality of agency along this continuum. At the top, Metrorail commuters are directly involved as individual agents in determining their mobility outcomes. Here, agency resides in the individual and represents deliberate personal action (the “I” and “me”). At the bottom, people are only indirectly influencing the outcome as they rely on someone else to act on their behalf. Here, agency and the ability to achieve a desired outcome reside with a proxy (“they” and “the others”). In the middle, agency is shared through interdependent effort as part of the collective agency pathway (“we” and “us”). The reason for this elongated cluster to bend into a crescent around the smaller cluster 1 is that all three interconnected agencies are different from, and thus maximally distant to, the non-agentive cluster 1. Finally, despite their commonalities, the modes of agency are relatively distant from each other, which supports Bandura’s argument that individual, proxy, and collective agency are different forms of agency. This difference is due in part to the distinct patterns of reciprocal interaction between the psychosocial and socio-structural environmental dimensions of agency. The illustration of this difference in agency types and their differential relation to environments was one of the goals of this analysis. According to the MDS results, we found that mobility as agency from the perspectives of Metrorail users consisted of three distinct agentive pathways, which were differentiated not only in terms of the locality of agency but also in how they related systematically to different mobility environments.

Mobility as Agency: The Agentive Pathways of Metrorail Commuters in the Western Cape

In this analytic step, we re-contextualised key patterns in the MDS map according to HCA (Bergman 2010), which allowed us to better understand the meaning of the MDS patterns as described above.

The Individual Agentive Pathway: How did Metrorail commuters achieve individual agency, considering the many challenges inherent in the mobility system? Re-contextualisation revealed that commuters primarily used the individual agentive pathway to overcome or avoid Metrorail’s weaknesses, which included alternative modal choices to overcome delays or breakdowns, such as borrowing or buying a car, or using buses or taxis, where available. Here are some examples:

Um, people get into trouble at work because [the trains] are always late. Often I can go back and fetch my car and go with my car but many thousands of people don’t have a car that they could take. This is their only transport. I use it because it is cheaper and because I can read while on the train. I can’t read while I am driving. (MT5: 2)

It’s ridiculous. I mean I used to use the railway but you cannot get to work late. I mean half-an-hour late, more than three times within a month and blame it on public transport. I mean after a while, I’m very sure, that your employer thinks it’s your fault. (MT24: 1)

Interviewer: Why don’t you take trains more often?

Interviewee: Really? No way, no, no. The thing is, like, if there is a possibility, I would not take a train, if I had a substitute. I would rather take the substitute. But the thing is, this is the cheapest form of transport. But preferably I would rather go by the car. Unfortunately, which I don’t have, but that’s not, no. (MT20: 7)

An interesting variant of this agentive pathway related to commuters exploiting systemic weaknesses to achieve desired outcomes, as illustrated by this excerpt:

Cause if, say tomorrow, a better service than the trains were to come at an affordable price, I promise you, people would stop using the train. It’s just that it’s affordable and it’s easier to use a train, when you don’t even have money or a ticket for some people, because I know guys who live in my street and they wake up early and they leave at four to go to the train, they catch the train for free and then they come back after eight – there’s no guards or anything. So, they don’t buy tickets; they just ride for free. (MT15: 7)

Getting up very early and coming back late at night, or returning home to fetch one’s car when a train is cancelled are examples of how individuals used selected environments to adapt to challenging situations to create viable agentive options. However, such individual agentive practices were exceptional because only a small number of commuters in our sample had access to a car, or were able to commute early or late enough to avoid certain environmental obstacles. Most commuters travelled during peak hours and reported feeling trapped in a deficient mobility environment, since they lacked the means and access to viable alternatives. The majority of mentions connected to this pathway were related to expressions of preferences, wishes, and aspirations; they referred to what commuters wished they had.

Interviewee: I don’t know, if you know the movie “The Italian Job?”

Interviewer: Yes, yes…

Interviewee: Did you see that scene where they’re sitting in that train that looks like a spacious, expensive one. Wouldn’t you wish to be there, like to use that train as a form of transportation?

Interviewer: Yeah, of course. But that’s in the movies.

Interviewee: Yeah, that’s what I’m saying. You only see them on TV and you wish, why don’t we have that, you know? If they’re coming up with something new that you’d wish, that you’d use as a substitution for your trains. It’s either not around you, or it’s too expensive for you. For instance, the Gautrain, that’s a nice train, you know? But where is it? It’s only in Gauteng, not around South Africa. And certain people use it, not everybody is using it, you know.

Interviewer: What do you mean with certain people?

Interviewee: People of that region. People who can afford it. Because even there, not everybody is using it. It’s people who can afford it, because maybe the prices are high, I don’t know, but why do they not make it for all of us. Not like the trains that we’re using now. (MT25: 9)

The re-contextualisation revealed the limitations of the individual agentive pathway. The ability to choose from a variety of mobility options that facilitate agentive goals tended to be unrelated to daily commuting experiences. Instead, this agentive pathway was predominantly aspirational, confined to wishes or outcomes that manifested in an imagined present or distant future.

The Proxy Agentive Pathway: In the proxy pathway, commuters often looked to other agents to act on their behalf. The initial analysis revealed that proxies included, most prominently, Metrorail, but also close family members and friends, lift clubs, and members of personal networks. A re-contextualising analysis showed that common to all proxies was an ability to construct a new or different mobility environment for commuters. This was obvious when considering the facilitating effect of a lift club, a ride with a work colleague to or from work, or being rescued when stranded. In relation to Metrorail, however, the agentive pathway was obfuscated. While expectations toward the mobility environment were often clearly communicated, facilitating environments rarely materialised as commuters reported that Metrorail was either unwilling or unable to intervene on their behalf. Here are some examples:

Interviewee: Um, I think it will just become more and more neglected. The whole train network will just become more and more neglected.

Interviewer: Why do you say that? Interviewee: Because all the signs are there that there is no focus on maintenance. I don’t think that Transnet [the State holding company of Metrorail] or Metrorail or whoever has the ability to do it and I also think that they don’t have the money to maintain everything or to keep it at an acceptable level. So that’s the picture that I see – a negative picture… (MT21: 3)

So they really need to implement something of that kind to improve their service. They really need to, they really need to improve their service. Funny enough, I actually saw the other day in the Argus [local newspaper] that they are planning to, but you know they always make plans and make plans and make plans and nothing ever comes from it. The Minister of Transport has just the other day, there was an article in one of the newspapers, they are planning to do something but they never get so far as to actually deliver anything. So, it’s really, it’s really a problem you know. It is a big problem and unfortunately this is the way it is… (MT3: 3)

Interviewer: So, what do you think will happen in the future?

Interviewee: In this current state? Nothing. If nothing happened for two decades, what will change now and in the future? I believe it was two years ago you know our president? They bought new trains, but I think it was too high, I believe. Or couldn’t fit on the railways or something, but I believe it was the wrong trains or the trains were not engineered for our railways, whatever, something was wrong, I believe it was too high, I’m not sure. […] So, nothing happened, still the same. The only thing that changed is that they made the tickets more expensive. (MT20: 8)

While commuters’ hopes and expectations illustrated how the proximal agentive pathway ought to function, many examples from the data also indicated that their lived experience differed considerably. This helps to explain why the dimension of action potential is unconnected to this pathway. The breakdown in the function of this agentive pathway emphasised the challenges inherent in the mobility system.

The Collective Agentive Pathway: In relation to the challenges intrinsic to the Metrorail system, collective agency was perhaps the most informative and successful of the agentive pathways. Given that it is maximally distant to cluster 1 (containing all socio-structural environmental constraints and negative situational circumstance associations of mobility) in Figure 19, we can assume that it was the most functional of the agentive pathways. A re-contextualising analysis of this agentive pathway revealed why this is the case. Collective agency was most frequently associated with social networks that commuters activated. These “cliques” consisted of friends, colleagues, or religious circles formed by commuting together. Here an example:

Interviewee: And there are also these cliques that form on trains. So everyone knows when they get on that this is their group that they chat with until they get off.

Interviewer: Do you have a group?

Interviewee: Yes, we have a group that meets in the mornings and we have church services on the train. So we are a group that meets on the train in the mornings and then hold a nice church service until we reach Stellenbosch station, until we get off at our station. There are a lot of networking groups and social networking groups that have formed because of it. And as a group, we also make sure that we meet once a month and go out for something to eat. (MT1: 2)

The functionality of these groups linked to a supportive or protective role they played in the lives of commuters. Not only did they support social and cultural activities, but they also provided safety and comfort to commuters confronted with uncomfortable, unreliable, and potentially dangerous commuting environments. In this way, the collective pathway helped to construct a protective buffer between commuters and the uncomfortable or potentially hostile mobility environment. Here are some examples:

Nothing bad has happened to me personally because I always travel in a group. It is a lot better if you travel in a group. And that’s the other thing when you take the train a lot then you meet and get to know people. And then people know you travel at that time every day and then you can sit in the same carriage and then you develop relationships like this. I have a group that I take the train with every night. And like when someone isn’t there you would message them and say “Where are you?” or “Are you late?” and so on. (MT4: 4)

So there is usually the thing that if I travel on a train then I need to be in a large group, large group being five or more people, safety in numbers. (MT27: 4)

I also once, it happened in the morning. I was writing [exams] that morning. So the trains were delayed and there had been delays from early in the morning. I had no other option ‘cause it was internal exams. With internal exams my teacher shows no mercy. If you’re late, you’re late, you’re not gonna write. And it was June so I needed the marks to apply to University. People started [she claps her hands loudly], the train came and people started to get on. I tried to get on, I tried, I fought and I fought. Then I could, one foot was on but the other foot was not. My bag was outside, my face was inside. I was holding on by the doors there, you know, onto the frame. I was holding by the door frame, so when the train was about to approach Bellville, it makes a turn but like a huge turn. I almost fell. If it was not for the person that was next to me, but a bit to the inside, I would have fallen. ‘Cause this guy saved my life, he just grabbed me by my shirt and tie and held onto me. And then I couldn’t breathe because I have asthma. I had already given up, I was going to die. But he pulled me in and other people also noticed that I was fainting. There was, I don’t know what happened, I don’t know where the people went, there was space, they made space. I was able to lie down and then they gave me a space to breathe, but I almost died. (MT15: 8)

The Function of Agentive Pathways: As stated earlier, individual, proxy, and collective agentive pathways lie on an agentive continuum. We found that these pathways varied according to the socio-structural, environmental impositions commuters experienced. Individual agentive pathways, for example, allowed individuals to respond to environmental constraints by selecting different mobility options to overcome or avoid problems and therewith created viable alternatives. Another strategy involved activating the proxy agentive pathway, which aimed to secure the help of more powerful actors, such as Metrorail or the government, to improve the mobility environment and to overcome socio-structural environmental impositions on behalf of commuters. While both strategies should have theoretically enabled someone to address, overcome, or avoid environmental constraints, few commuters were able to effectively implement personal agentive strategies to avoid Metrorail and most attempts at proxy agency seemed to fail at least in the short run to activate Metrorail or the government to improve regular commutes. It is within this context that the function and relative success of collective agency became most apparent. While collective efforts may not have been able to change the environment – they cannot prevent trains from being late, or services from being disrupted, or skollies and tsotsis (loosely translated, gangsters or criminals) from boarding trains – through interdependent effort, they provided a protective buffer that enhanced the action potential of commuters. By constructing an environment that offered resource and information sharing, coping mechanisms, and strength in numbers, this collective effort often provided the most functional agentive pathway of Metrorail users during their mobility encounters.

Discussion and Conclusions

The purpose of this article is three-fold: to explore empirically Albert Bandura’s dimensions of agency and environment using the model of triadic reciprocal causation, to examine mobility as agency among Metrorail users in the Western Cape from this theoretical perspective, and to explore ways in which conceptions of mobility need to integrate technical, socio-structural, and psychosocial components in order to offer a context- and culture-sensitive approach to sustainable mobility. In applying Bandura’s framework, we identify empirically all intra- and inter-personal, as well as psychosocial and socio-structural dimensions that are part of his theory. We find that Bandura’s agency concept serves as a suitable analytic framework to systematise commuting experiences and practices among our Metrorail users. Mapping patterns of reciprocal interaction between agency and environment dimensions to study the interdependence between agentive dimensions enables us to visualise how mobility as agency unfolds along distinct pathways relating to individual, proxy, and collective agency. Agentive pathways lie on a continuum as agency moves from the individual to the proxy, with the collective occupying a central position. Another way to understand these agentive pathways refers to the function they serve in relation to the types of mobility environments. While the individual agentive pathway is closest to the classical understanding of personal agency or self-efficacy, and therefore representative of mobility achieved, in the context of Metrorail users, it remains largely aspirational, given that few of our commuters have access to alternative mobility modes. Proxies, such as Metrorail and the government, are critical to creating and mediating the mobility environment, and their failure to do so contributes to the restrictions and frustrations associated with the mobility system. In the context of Metrorail commuters in the Western Cape, it is the protective buffer of collective agency that enables commuters to achieve most consistently mobility as agency.

With regard to our first objective – to expand the concept of personal agency beyond the confines of unidirectional modes of causation adopted in studies on personal agency and self-efficacy, our application of Bandura’s framework of reciprocal causation shows that mobility as agency is inherently psychosocial and functionally dependent on technical and socio-structural dimensions. While our study supports evidence for the one-dimensional, intra-personal individual agentive pathway conventionally pursued in studies on personal agency (Bores- Rangel et al. 1990; Church et al. 1992; Ryerson 1994; Hackett & Byars 1996; Bandura 1997, 2004a; Langlois et al. 1999; Rogers et al. 1999; Gibson 2004; Chapman-Novakofski & Karduck 2005; Hill et al. 2009; Tougas et al. 2015), we also identify other agentive pathways. To our knowledge, this is the first empirical exploration of the model of triadic reciprocal causation as proposed by Albert Bandura, and our study provides evidence for a more nuanced understanding of agency as distinct and systematic patterns of reciprocal interactions. Although our study is limited to a specific context – Metrorail commuters in the Western Cape, Bandura’s framework of triadic reciprocal causation and the mixed methods framework we adopt here serve as effective analytic tools to examine empirically this theory. Future research on agency and mobility in this vein ought to examine train systems and populations that differ in agency, environment, or region, or to study personal agency beyond mobility contexts to systematise how agentive pathways function more generally.

In contrast to some of Bandura’s critics, we find that the relative looseness of his concepts and their interdependence (Tschannen-Moran & Hoy 2001; Garvis & Pendergast 2016) prove to be an advantage. It enables us to use the model of triadic reciprocal causation to examine the interdependence between commuters and their environment without imposing a priori relationships. While Bandura defined the three types of agency and three types of environments, he did not define or operationalise how they are connected, arguing that this would vary according to context, culture, and other behavioural predispositions. Consequently, we could use the experiences of Metrorail commuters to identify how they and their environment shape mobility as agency, and how this functions in the context of Metrorail in the Western Cape. Given that agency is mediated by a constellation of contextual and cultural influences, the variability embedded in Bandura’s model provides an excellent framework to study personal agency in different settings, something future studies should pursue further.

The situated application of the model of triadic reciprocal causation expands what sustainable mobility may mean in a specific mobility context. To date, most studies in the mobility domain are limited to either intra-individual or structural concerns (Bergman et al. 2014; Bergman & Bergman 2015; Cass & Faulconbridge 2016), focusing on either infrastructure (Novaco 2001; Brög et al. 2004; Hunecke et al. 2007) or commuter preference and behaviour (Novaco 2001; Steg & Vlek 2009). When we examine the links between intra- and inter-personal, as well as socio-structural environmental dimensions of the Metrorail commuters we interviewed, our study concurs with Shepherd and Marshall (2005) findings that mobility practices are nested within an interdependent network of individual, social, and environmental factors. The reciprocal interactions between these dimensions have consequences on the day-to-day practices of commuters, and they highlight the weakness of policy approaches that fail to take this into account (see also Charlton 2004; Steg & Gifford 2005). Our study makes an empirical contribution toward systematising the distinct patterns of reciprocal interaction between preferences and behaviours in conjunction with a specific context of a mobility environment. Based on our analysis of Metrorail commuters, we suspect that sustainable mobility policies aimed solely at individual behaviour change or environmental and structural barriers are likely to have only limited success because the Metrorail environment is insufficiently aligned with different types of agency. They tend to lack the necessary attributes to enable a positive action potential, and in their current state, they do not connect sufficiently with the context and culture of commuters. Accordingly, we can make two policy recommendations: Individual commuter preferences and behaviours need to be conceptualised and understood in relation to a specific context and culture of mobility environments when formulating mobility solutions. And mobility interventions need to carefully blend hard and soft policy approaches while considering agency in a specific environment. Whereas most mobility approaches rightfully stress the importance of safe, reliable, and affordable public transport, they neglect what these three characteristics mean in situ, for example for employed or unemployed women living in townships or informal settlements. A mobility system that integrates contextual and cultural sensitivities would present a formidable baseline for agency beyond transportation. Our study invites policy makers to think in more complex ways about mobility systems. For example, in the South African context, the low prestige of train travel in relation to the high status of car ownership, particularly for males, needs to be considered when developing mobility solutions that integrate technological as well as motivational and emotional components. The agentive pathways we studied here reflect the constellation of psychosocial and socio-structural environmental dimensions, which make up the mobility context of Metrorail commuters in the Western Cape. This context is characterised by extreme environmental constraints: overcrowded, dilapidated, outdated, and often unsafe trains and train infrastructure. Perhaps the most significant limitation to our study is its small scale and its specific Western Cape context, and future studies with a larger sample size could examine in more detail structural, contextual, and individual differences. While our study is thus not generalizable to a research group or geographic region, it nevertheless reveals how a theoretical framework serves well to illustrate different types of agency and their association with different types of mobility environments. Often, the ineffectiveness of a policy approach is best understood by transposing general policy assumptions into a specific context. Thus, in addition to this accomplishment, our study presented a thick description (Geertz 1973) of everyday experiences of commuters in the Western Cape along a sophisticated psychological framework. Finally, our study highlights a promising approach for improving sustainable mobility systems beyond hard or soft policies. Removing obstacles that prevent agentive practices and taking into consideration different types of environments represent important steps toward developing context-specific and culture-sensitive sustainable mobility strategies.


Bandura, A. (1986). Social foundations of thought and action: A social  cognitive theory. Englewood Cliffs, NJ, US: Prentice-Hall, Inc.

Bandura, A. (1989). Human agency in social cognitive theory. Am. Psychol.

44, 1175–1184. doi: 10.1037/0003-066X.44.9.1175

Bandura, A. (1997). Self-efficacy: The exercise of control. New York, NY, US: W H Freeman/Times Books/Henry Holt & Co.

Bandura, A. (1999). “A social cognitive theory of personality” in Handbook of personality. L. Pervin, and O. John (Ed.) (New York, US: Guilford Publications), 154–196.

Bandura, A. (2001). Social cognitive theory: an agentic perspective. Annu. Rev. Psychol. 52, 1–26. doi: 10.1146/annurev.psych.52.1.1

Bandura, A. (2004a). Health promotion by social cognitive means. Health Educ. Behav. 31, 143–164. doi: 10.1177/1090198104263660

Bandura, A. (2004b). “Social cognitive theory for personal and social change by enabling media” in Entertainment-education and social change: History, research, and  practice. LEA’s Communication  Series. eds. A. Singhal, M. J. Cody, E. M. Rogers, and M. Sabido (Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers), 75–96.

Bandura, A. (2006). Toward a psychology of human agency. Perspect. Psychol. Sci. 1, 164–180. doi: 10.1111/j.1745-6916.2006.00011.x

Bandura, A. (2008). “An agentic perspective on positive psychology” in Positive psychology: Exploring the best in people, Vol 1: Discovering human strengths. Praeger perspectives. ed. S. J. Lopez (Westport, CT, US: Praeger Publishers/ Greenwood Publishing Group), 167–196.

Banister, D. (2008). The sustainable mobility paradigm. Transp. Policy 15, 73–80. doi: 10.1016/j.tranpol.2007.10.005

Bergman, M. M. (2010). “Hermeneutic content analysis: textual and audiovisual analyses within a mixed methods framework” in SAGE handbook of mixed methods in social & behavioral research. eds. A. Tashakkori and C. Teddlie (Thousand Oaks, California, United States: SAGE Publications, Inc.), 379–396. Bergman, M. M. (2011). Content configuration analysis. StudyCube. St. Gall: University of St. Gall.

Bergman, M. M., Bergman, Z., and Gravett, S. (2011). The development and application of the explanatory model of school dysfunctions. S. Afr. J. Educ. 31, 461–474.

Bergman, Z., Bergman, M. M., and Pirie, G. (2014). Die Zukunft Der Bahnmobilität in Südafrika: Vom Service Public Zur Sozialen Exklusion. Afr. Bull. 156, 6–7.

Bergman, Z., and Bergman, M. M. (2015). The imaginary of the train of future in South Africa: from public access to social exclusion. World Transp Policy Pract. 21, 23–34. doi: 10.1007/978-94-017-8911-0

Boas, I. (2017). Environmental change and human mobility in the digital age. Geoforum 85, 153–156. doi: 10.1016/j.geoforum.2017.07.022

Bores-Rangel, E., Church, A. T., Szendre, D., and Reeves, C. (1990). Self-efficacy in relation to occupational consideration and academic performance in high school equivalency students. J. Couns. Psychol. 37, 407–418. doi: 10.1037/0022-0167.37.4.407

Brög, W., Erl, E., and Mense, N. (2004). “Individualised marketing: changing travel behaviour for a better environment” in OECD (Ed.) Communicating. (Paris: OECD Publications). 83–97.

Carillo, K. D. (2010). “Social cognitive theory in IS research – literature review, criticism, and research agenda” in Information systems, technology and management. Communications in computer and information science (Berlin, Heidelberg: Springer), 20–31.

Cass, N., and Faulconbridge, J. (2016). Commuting practices: new insights into modal shift from theories of social practice. Transp. Policy 45, 1–14. doi: 10.1016/j.tranpol.2015.08.002

Chapman-Novakofski, K., and Karduck, J. (2005). Improvement in knowledge, social cognitive theory variables, and movement through stages of change after a community-based diabetes education program. J. Am. Diet. Assoc. 105, 1613–1616. doi: 10.1016/j.jada.2005.07.010

Charlton, C. (2004). Social change and sustainable transport. J. Transp. Geogr.12, 165–167. doi: 10.1016/j.jtrangeo.2004.02.003

Church, A. T., Teresa, J. S., Rosebrook, R., and Szendre, D. (1992). Self-efficacy for careers and occupational consideration in minority high school equivalency students. J. Couns. Psychol. 39, 498–508. doi: 10.1037/0022-0167.39.4.498

Collado, S., Staats, H., and Corraliza, J. A. (2013). Experiencing nature in childrens summer camps: affective, cognitive and behavioural consequences. J. Environ. Psychol. 33, 37–44. doi: 10.1016/j.jenvp.2012.08.002

Collins, C. M., and Chambers, S. M. (2005). Psychological and situational influences on commuter-transport-mode choice. Environ. Behav. 37, 640–661. doi: 10.1177/0013916504265440

de Winter, J. C. F., Dodou, D., and Wieringa, P. A. (2009). Exploratory factor analysis with small sample sizes. Multivariate Behav. Res. 44, 147–181. doi: 10.1080/00273170902794206

Eastman, C., and Marzillier, J. S. (1984). Theoretical and methodological difficulties in Bandura’s self-efficacy theory. Cogn. Ther. Res. 8, 213–229. doi: 10.1007/ BF01172994

Garvis, S., and Pendergast, D. (Eds.) (2016). Asia-Pacific perspectives on teacher self-efficacy. Rotterdam Netherlands: Sense Publishers. https://www.springer. com/gp/book/9789463005210

Geerken, T., Vercalsteren, A., and Borup, M. (2009). “Review of the mobility domain” in System innovation for sustainability 2 case studies in sustainable consumption and production – Mobility. Vol. 2 (London: Routledge).

Geertz, C. (1973). The interpretation of cultures. 1st edn. New York: Basic Books. Gehlert, T., Dziekan, K., and Gärling, T. (2013). Psychology of sustainable travel behavior. Transp. Res. Part A Policy Pract. 48, 19–24. doi: 10.1016/j.tra.2012.10.001

Gibson, S. K. (2004). Social learning (cognitive) theory and implications for human resource development, social learning (cognitive) theory and implications for human resource development. Adv. Dev. Hum. Resour. 6, 193–210. doi: 10.1177/1523422304263429

Guadagno, L. (2016). Human mobility in the Sendai framework for disaster risk reduction. Int. J. Disast. Risk Sci. 7, 30–40. doi: 10.1007/s13753-016-0077-6 Hackett, G., and Byars, A. M. (1996). Social cognitive theory and the career development of African American women. Career Dev Q. 44, 322–340. doi:10.1002/j.2161-0045.1996.tb00449.x

Hill, J. R., Song, L., and West, R. E. (2009). Social learning theory and web- based learning environments: a review of research and discussion of implications. Am. J. Distance Educ. 23, 88–103. doi: 10.1080/08923640902857713

Hunecke, M., Haustein, S., Grischkat, S., and Böhler, S. (2007). Psychological, sociodemographic, and infrastructural factors as determinants of ecological impact caused by mobility behavior. J. Environ. Psychol. 27, 277–292. doi: 10.1016/j.jenvp.2007.08.001

IPCC. (2007). “Climate Change 2007: synthesis report” in IPCC fourth assessment report (AR4). (Geneva, Switzerland: Intergovernmental Panel on Climate Change). assessment_report_synthesis_report.htm

IPCC. (2014). “Climate change 2014: synthesis report” in Fifth assessment report of the intergovernmental panel on climate change. (Geneva, Switzerland: Intergovernmental Panel on Climate Change).

Landry, N., Gifford, R., Milfont, T. L., Weeks, A., and Arnocky, S. (2018). Learned helplessness moderates the relationship between environmental concern and behavior. J. Environ. Psychol. 55, 18–22. doi: 10.1016/j. jenvp.2017.12.003

Langlois, M. A., Petosa, R., and Hallam, J. S. (1999). Why do effective smoking prevention programs work? Student changes in social cognitive theory constructs. J. Sch. Health 69, 326–331. doi: 10.1111/j.1746-1561.1999. tb06423.x

Lokhorst, A. M., Werner, C., Staats, H., van Dijk, E., and Gale, J. L. (2013). Commitment and behavior change: a meta-analysis and critical review of commitment-making strategies in environmental research. Environ. Behav. 45, 3–34. doi: 10.1177/0013916511411477

Mair, P., De Leeuw, J., Borg, I., and Groenen, P. J. F. (2015). Multidimensional scaling in R: SMACOF. R Package version 1.9–6.

Novaco, R. W. (2001). “Psychology of transportation” in International Encyclopedia of Social & Behavioral Sciences. (London: Elsevier). books/international-encyclopedia-of-social-andampamp-behavioral-sciences/smelser/978-0-08-043076-8

OECD (2016). Making cities work for all, October. Paris: OECD Publications. Pillemer, K., Wells, N. M., Meador, R. H., Schultz, L., Henderson, C. R., and Cope, M. T. (2017). Engaging older adults in environmental volunteerism: the retirees in service to the environment program. Gerontologist 57, 367–375. doi: 10.1093/geront/gnv693

Poortinga, W., Steg, L., and Vlek, C. (2004). Values, environmental concern, and environmental behaviour – A study into household energy use. Environ. Behav. 36, 70–93. doi: 10.1177/0013916503251466

Rogers, E. M., Vaughan, P. W., Swalehe, R. M., Rao, N., Svenkerud, P., and Suruchi, S. (1999). Effects of an entertainment-education radio soap opera on family planning behavior in Tanzania. Stud. Fam. Plann. 30, 193–211. doi: 10.1111/j.1728-4465.1999.00193.x

Ryerson, W. N. (1994). Population communications international: its role in family planning soap operas. Popul. Environ. 15, 255–264. doi: 10.1007/ BF02208459

Schmutzler, J., Andonova, V., and Diaz-Serrano, L. (2018). How context shapes entrepreneurial self-efficacy as a driver of entrepreneurial intentions: a multilevel approach. Entrep. Theory Pract. doi: 10.1177/1042258717753142

Shepherd, A., and Marshall, E. (2005). Timeliness and task specification in designing for human factors in railway operations. Appl. Ergon. 36, 719–727. doi: 10.1016/j.apergo.2005.05.005

Spence, I. (1979). A simple approximation for random rankings stress values. Multivariate Behav.  Res. 14, 355–365. doi: 10.1207/s15327906mbr1403_5 Stanton, N. A., McIlroy, R. C., Harvey, C., Blainey, S., Hickford, A., Preston,

J. M., et al. (2013). Following the cognitive work analysis train of thought: exploring the constraints of modal shift to rail transport. Ergonomics 56, 522–540. doi: 10.1080/00140139.2012.718366

Starkey, P., and Hine, J. (2014). Poverty and sustainable transport: How transport affects poor people with policy implications for poverty reduction. London, UK: UN Habitat. sustainable-transport-how-transport-affects-poor-people-with-policy- implications-for-poverty-reduction#citation

Steg, L., Bolderdijk, J. W., Keizer, K., and Perlaviciute, G. (2014). An integrated framework for encouraging pro-environmental behaviour: the role of values, situational factors and goals. J. Environ. Psychol. 38, 104–115. doi: 10.1016/j. jenvp.2014.01.002

Steg, A., and Gifford, R. (2005). Sustainable transportation and quality of life. J. Transp. Geogr. 13, 59–69. doi: 10.1016/j.jtrangeo.2004.11.003

Steg, L.,  and  Vlek,  C.  (2009).  Encouraging  pro-environmental  behaviour:  an integrative review and research agenda. J. Environ. Psychol. Environmental Psychology on the Move 29, 309–317. doi: 10.1016/j.jenvp.2008.10.004

St-Jean, E., Radu-Lefebvre, M., and Mathieu, C. (2018). Can less be more? Mentoring functions, learning goal orientation, and novice entrepreneurs’ self-efficacy. Int. J. Entrep. Behav. Res. 24, 2–21. doi: 10.1108/IJEBR-09-2016-0299 Titheridge, H., Mackett, R. L., Christie, N., Oviedo Hernández, D., and Ye, R. (2014). Transport and Poverty: A Review of the Evidence. Report. London, UK: UCL Transport Institute, University College London. Available at: https://

Tougas, M. E., Hayden, J. A., McGrath, P. J., Huguet, A., and Rozario, S. (2015). A systematic review exploring the social cognitive theory of self- regulation as a framework for chronic health condition interventions. PLoS One 10:e0134977. doi: 10.1371/journal.pone.0134977

Tschannen-Moran, M., and Hoy, A. W. (2001). Teacher efficacy: capturing an elusive construct. Teach. Teach. Educ. 17, 783–805. doi: 10.1016/ S0742-051X(01)00036-1

van Wee, B., Holwerda, H., and van Baren, R. (2002). Preferences for modes, residential location and travel behaviour: the relevance for land-use impacts on mobility. Eur. J. Transp. Infrast. Res. 2, 305–316. doi: 10.1016/ S0742-051X(01)00036-1

World Bank Group. (2016). Measuring Rural Access. Other Papers. World Bank. Yamamoto, L., Serraglio, D. A., and Cavedon-Capdeville, F. d. S. (2018). Human mobility in the context of climate change and disasters: a South American approach. Int. J. Clim. Change Strateg. Manage. 10, 65–85. doi: 10.1108/IJCCSM-03-2017-0069

Zhao, P., and Li, S. (2016). Restraining transport inequality in growing cities: can spatial planning play a role? Int. J. Sustain. Transp. 10, 947–959. doi:10.1080/15568318.2016.1191693


The following article can be cited as follows:

Bergman, Z., & Bergman, M.M. (2019). A Case Study of the Sustainable Mobility Problem–Solution Paradox: Motility and Access of Metrorail Commuters in the Western Cape. Sustainability, 11(10), 2842.

Agency and Bandura’s Model of Triadic Reciprocal Causation: An Exploratory Mobility Study Among Metrorail Commuters in the Western Cape, South Africa


Public transport in general and passenger trains in particular are often advertised as solutions to mobility challenges due to their relatively low carbon footprint, high commuter load, high public safety, and contribution to reduced road congestion. But, how do these advantages apply to contexts characterised by inequality, poverty, and exclusion, and where train infrastructure is underdeveloped and poorly maintained? In this study, we examine the imaginaries and their associated transport predispositions of Metrorail users in the Western Cape province of South Africa. Based on 31 interviews conducted with Metrorail users, we explored how they conceptualise access to and use of mobility. The conceptual framework for this is provided by the Motility concept as developed by Kaufmann, Bergman, and Joye. Findings show that the context and culture defining the daily lives of Metrorail users reflect a reality, which is far removed from the way we theorise sustainable mobility. The limitations of spatial and social inequality, which create the mobility boundaries of Motility for these commuters, reveal a significant gap between their lives and the policies aimed to foster our sustainable mobility future. Despite this, the commuters of our study are highly mobile, and we end this article with an attempt to align these conflicting domains of dysfunctional contexts, mobility practices, and sustainability ideals.

Keywords: Sustainable mobility, trains, mobility access, Metrorail, motility, Western Cape, South Africa, Content Configuration Analysis (CCA).

“We must all learn … to think differently. We need to learn how to transform our policies and strategies to address the challenges of sustainability. To reach the poor and vulnerable, we need targeted policies, active outreach, and integrated information to inform decision-making. We need to recognise and understand the multiple dimensions of poverty and vulnerability, and how they interconnect. And we need to break down silos … between the economic, social, and environmental aspects of development.”

(Ban Ki-Moon, former UN Secretary General, 19 July 2016)

Look, I think that the service Metrorail provides is really terrible. I started using the trains in 2006. It was sad back then and now it is even sadder. It gets worse every year. (J, 1) (Each interview was assigned a unique identifier, denoted by a letter, followed by the page number of the transcript from which the quote was taken.)


Our mobility reflects the best and worst of human development. The displacement of goods, information, technology, and people has alleviated extreme poverty for billions of people and increased the quality of life and wellbeing for many more. It has also enabled unprecedented access to education, work, health, family and friends, and consumer goods. However, it also harms our environment, health, and wellbeing. The costs of persistent and increasing road, water, and air traffic congestion, the continued dependence on fossil fuels and associated destruction, as well as the millions of victims of accidents and pollution result in the “Triple C” challenge: How to reduce CO2, congestion, and casualties (Geerken et al. 2009)? More precisely, how should societies reduce mobility’s social, environmental, and economic costs, while maintaining the potential for development (Holden, Linnerud, & Banister 2017)?

At its broadest and most abstract level, the literature on sustainable mobility borrows the Brundtland framework of consolidating economic growth, social development, and environmental protection with a particular focus on inter-generational responsibility or fairness (Greene, 2001, see also Holden, Linnerud, & Banister 2013; Holden et al. 2017). Accordingly, sustainable mobility is defined as “the ability to meet today’s transportation needs without compromising the ability of future generations to meet their transportation needs” (Richardson 2005, p. 30). By decreasing the dependence on private vehicles (Charlton 2004), encouraging modal shifts to public transport (IPCC 2007), and reducing the ecological impact of individual mobility behaviour (Holden & Høyer 2005; Hunecke et al. 2007), the main goals are ostensibly to “reduce the need to travel (less trips) … reduce trip lengths, and to encourage greater efficiency in the transport system” (Banister 2008, p. 75; see also Holden 2004; Holden & Gilpin 2013). Many cities and regions across the globe are adopting some of these ideals, such as Baltimore and Portland in the USA, Vancouver and Calgary in Canada, Canberra in Australia, Birmingham and Windsor in the UK, Beijing and Chengdu in China, parts of New Zealand, the Philippines, and the Gauteng and Western Cape provinces in South Africa.  Many cities and regions have embraced implicitly or explicitly the green transportation hierarchy, which ranks modes of transportation according to their degree of sustainability (Figure 1).

  Figure 1. Green transportation hierarchy (Bradshaw 1992).

According to this individual transportation hierarchy, walking is the most sustainable form of mobility, followed by cycling, then public transport, such as buses, trams, and especially trains. High-occupancy vehicles or HOVs, which include car-pooling and car-sharing initiatives, are ranked next, and, finally, the least sustainable form of individual mobility encompasses single-occupancy vehicles (SOV) or private cars. This hierarchy is presented in numerous policy documents, urban planning forums, and government and city websites around the world, indicating not only the growing importance of sustainability more generally, but also an emergent normative agreement on how mobility and its sustainability consequences ought to be addressed: carbon-free cities in which daily activities are reorganised into decentralised clusters that favour human-propelled modes of locomotion.

Despite some forerunners, most cities and regions around the globe are far removed from these ideals. Nevertheless, many politicians, non-governmental and non-profit organisations (NGOs and NPOs), academics, city planners, and engineers aim to “unsprawl”, “decongest”, and “decelerate” cities by fostering public because large-scale public mobility systems, especially trains, are considered sustainable, given their reduced carbon footprint, high commuter load, increased safety, and contribution to reducing road congestion. However, large-scale public mobility systems also create new problems or exacerbate existing ones (Bergman & Bergman 2015; Bergman, Bergman, & Pirie 2014; OECD 2015; Young & Kiel 2010). Kaufmann, Bergman, and Joye (2004, p. 735) described, for example, how access to large-scale mobility systems, such as the “train à grand vessel” (TGV) in France or EasyJet in Europe, not only revolutionised people’s perception of time and distance by creating new ways of living and working through “multi-residentiality”, “multi-locality”, and “multi-occupationality”, but they also created new forms of social and spatial segregation, since only a limited number of people are able to leverage these advantages. In this way, mobility systems may inadvertently aggravate inequality and unsustainability, and the major challenge to sustainable mobility remains identifying viable strategies to balance the moral imperatives of “satisfying human needs, ensuring social equity and respecting environmental limits” (Holden et al. 2017, p. 224). While the idea that large-scale public mobility systems contribute to new forms of spatial and social segregation is not new (e.g. Cass, Shove, & Urry 2005; Dupont 2004; Friedmann 1986; Grengs 2001; Power 2012), it seems to be largely absent from most debates on sustainable mobility. These predominantly Eurocentric and urban mobility solutions tend to have a strong impact on emerging economies, such as South Africa. Yet, how feasible, context-sensitive, and culture-aware are these approaches in developing economies, where 95% of future urban expansion will take place (UNSDG11 2016)? This is the central focus of our study as we aim to explore the suitability of normative, Eurocentric mobility models of sustainable mobility in a stratified South African context. We do this using a qualitative approach based on 31 in-depth, exploratory interviews with Metrorail commuters in the Western Cape, South Africa.

Theoretical Background

The concept of accessibility, conventionally used in the fields of transport planning, urban planning, and geography divides mobility access into land-use, transportation, temporal, and individual components (Geurs & van Wee 2004), and it defines access in terms of individuals’ potential to interact with mobility infrastructure (Hansen 1959). While focusing on road networks, travel speeds, congestion levels, and individual or household activity programs makes accessibility easy to measure and interpret (Geurs & van Wee 2004), this approach cannot account for the relationship between social determinants, such as inequality, exclusion, encultured mobility preferences, and mobility access.

To emphasise the socio-cultural dimensions of mobility access, Kaufmann, Bergman, and Joye (2004, p. 750) proposed that “spatio-social mobility may be realised differently or have different consequences across varying socio-cultural contexts”. To systematise this notion, Kaufmann and his colleagues expanded on the concept of Motility, which denotes the actual and potential capacity of people, goods, technology, and information to be mobile. Motility is understood as a form of capital that enables people to access and utilise other economic, social, and cultural assets. It consists of interrelated components, relating to “access to different forms and degrees of mobility, competence to recognise and make use of access, and appropriation of a particular choice, including the option of non-action” (2004, p. 750). Given that our aim is to examine what access to sustainable mobility, such as trains, means to people in the stratified, developing context of South Africa, we focus primarily on access. According to Kaufmann, Bergman, and Joye (2004, p. 750), access refers to “the range of possible mobilities according to place, time, and other contextual constraints, and may be influenced by networks and dynamics within territories”. This range is defined by access to mobility options and conditions. On the one hand, mobility options consist of the range of transportation means and communication available to commuters, as well as commuters’ access to mobility services and equipment. Mobility conditions, on the other hand, refer to how accessible these options are in relation to the spatial distribution of infrastructure, the sedimentation of spatial policies, and the socio-economic position of individuals.

In this article, we use the access component of motility to examine train mobility in the Western Cape province of South Africa. Specifically, we analyse the perspectives of Metrorail users from two cities, namely Cape Town and Stellenbosch. As the largest commuter train service in South Africa, Metrorail transports approximately two million people per day (Bergman, Bergman, & Thatcher 2019). In the Western Cape region, it consists of four lines—the Northern, Southern, Cape Flats, and Central Lines. The network is made up of 610 kilometres of track and 119 stations, which primarily serve to connect settlements, townships, suburbs, towns, and cities (Bergman, Bergman, & Thatcher 2019). Using the access component as proposed by the theory, we examine how commuters navigate and conduct their daily mobility practices in relation to options and conditions, which define or constrain their mobility access. By examining these components, we aim to better understand what role mobility plays in their lives, as well as the implications this has on the greater mobility context of the Western Cape. The theoretical framework outlined by the motility concept lends itself to this study because South Africa, similar to many other developing economies, is characterised by high levels of inequality and exclusion, and because it faces a variety of structural problems relating to individual and public mobility systems (Bergman & Bergman 2015; Bergman, Bergman, & Pirie 2014; Pirie 2013; Pirie 2009). A study on mobility access from the perspective of users in a developing context can highlight potentials and challenges for a global mobility future, especially because current and future mobility practices in emerging contexts will be a significant determinant of how not only sustainable mobility but especially sustainability itself will take shape globally. Accordingly, this article has three aims: (1) to systematise the options and conditions of mobility access as reported by the Metrorail users we interviewed; (2) to examine how these interactions shape commuters’ potential to be mobile not only in relation to Metrorail, but also to the greater mobility context of the Western Cape region; and (3) to explore how the perspectives of these Metrorail users on mobility in general and trains in particular connect to notions of sustainable mobility.

Materials and Methods

This qualitative study is based on 31 interviews conducted with Metrorail users in the Western Cape, South Africa. Our sampling criteria included location (working or living in Cape Town or Stellenbosch), mobility mode (Metrorail), and frequency of use (weekly use in the past two years). Consequently, most interviewees lived or worked in the vicinity of Stellenbosch or Cape Town. These two locations, approximately 50 km apart, provided a large geographic catchment area of people from varied backgrounds and introduced a diverse range of mobility characteristics. Some interviewees, for example, came from affluent suburbs in Stellenbosch and Cape Town, such as Claremont or Tygerberg, while others came from middle- to low-income communities, such as Brackenfell, Belhar, and Eerste Rivier. Some travelled from as far as Paarl and Wellington (an additional 40 to 50 km inland, which corresponds to a two- to three-hour train commute, one-way), and many lived in surrounding informal settlements and townships, such as Khayelitsha, Bonteheuwel, Langa, Mitchells Plain, and Kayamandi. Although the majority of our interviewees were daily commuters, some commuted only occasionally. Among the men and women we interviewed were students, teachers, security guards, shop attendants, cleaners, drivers, administrators, couriers, repair men and women, managers, occupational therapists, personal assistants, and unskilled labourers. Our interviewees were Black, White, Asian, and Coloured (Coloured is an official term used in South Africa to denote individuals from a mixed ethnic background. Different ethnic typologies exist but the statistical office and most recent population census use the labels Black (80.5%), Colored (8.8%), White (8.3%), and Asian or “Indian or other Asian” (2.5%). It should not be confused with pejorative and racist usages in other countries, referring to non-whites), and, although the interviews were conducted in English or Afrikaans, the two most widely spoken national languages (among 11), some of our interviewees also spoke Xhosa, Sotho, North Sotho, and isiZulu. Although this is not a representative sample, our interviewees reflected South Africa’s multi-cultural and diverse context. The interviews consisted of exploratory and semi-structured questions. Examples of exploratory questions included “Tell me everything that comes to your mind when you think about trains” or “What is your best memory with a train?” (Bergman, Bergman, & Thatcher 2019). These questions aimed to elicit extended narrative responses regarding interviewees’ mobility experiences. Semi-structured questions included “When, where, and how often do you take trains?” or “What do you think will happen with trains in the future?”. These questions aimed to prompt specific mobility needs, preferences, and aspirations.

We analysed the interviews using content configuration analysis (CCA; Bergman 2011; Bergman, Bergman & Gravett 2011). CCA is a type of qualitative analysis related to qualitative content and thematic analyses. It can be used on all non-numeric data including written and visual sources (Bergman, Bergman, & Gravett 2011). It is applied in the fields of education, health, business studies, psychology, economics, philanthropy studies, sociology, and sustainability studies. The main strength of this method lies in its flexibility, since analytic strategies can be adapted to research foci or researcher needs (Bergman, Bergman, & Gravett 2011). In this study, for example, we used CCA to conduct a quasi-deductive analysis based on theory-guided top-down coding, using the access component of the motility framework as a structuring principle and coding framework (Kauffman, Bergman, & Joye 2004). As a first step, we identified all mobility-related elements in the data. Next, we coded and classified these using the dimensions of motility access. This analysis enabled us to (a) identify if the dimensions of motility as outlined in the theory were present in the interviewees’ narratives, and (b) how these dimensions relate to the particular mobility experiences of commuters in the context of Metrorail in the Western Cape. Finally, we analysed each dimension of motility access to systematise what they mean from the perspective of the Metrorail commuters. Systematising the range of motility options and conditions that characterise the daily mobility experiences, and how they constrain actual and potential capacities of commuters to be mobile allowed us to link mobility preferences and behaviours with the mobility context and environment.


We sorted the data according to these six dimensions across a macro-, meso-, and micro-level, as illustrated in Figure 2.

Figure 2. Motility access from the perspectives of Metrorail interviewees.

At the macro-level, motility access relates to the greater mobility environment, specifically the spatial distribution of mobility infrastructure and the sedimentation of spatial policies. At the meso-level, motility access concerns Metrorail services and infrastructure. The micro-level connects Metrorail commuters to their means of transport and socio-economic position.

Macro-Level: The Greater Mobility Environment

Spatial distribution of mobility infrastructure: Metrorail commuters described their greater mobility environment predominantly in terms of the limitations of public transport infrastructure. These limitations manifested on multiple levels. Regionally, public transport was reported as restricted or unavailable, modal choices as limited, and existing infrastructure as too far away. Two examples are provided below.

When the trains are late, what are you supposed to do? You must just sit and wait because you don’t have other alternatives. There isn’t a taxi here that you could take, or there isn’t a bus service in place. (A, 3)

You see, because when it is off peak, even for us who need to walk home from the station. For women it is actually more dangerous. And you see, in the area where we live, where I live in Kleinvlei, the police station is right next to the train station, which means they are very visible so the people can walk and so on. But see now its winter and it is dark early. So already by 6 o’clock, its dark and then people still have to walk. So, no way. It is really very dangerous. For a woman to walk that time of the morning or evening is very dangerous. (W, 4)

From the perspectives of the Metrorail commuters we interviewed, the spatial distribution of mobility infrastructure was defined in terms of its lacunae. Commuters reported that mobility networks were either underdeveloped, dilapidated, or non-existent. Examples included that the entire catchment area of the Stellenbosch region was serviced by a single-track rail line, which caused frequent and long delays, or the differential levels of service between Northern and Southern lines or between the so-called “business trains” and the “normal trains”, the former receiving preferred treatment at the expense of the latter in terms of scheduling, solving delay issues, and overall service. Also mentioned was the absence of a feeder bus system, which resulted in long, inconvenient, or dangerous journeys to reach train stations. In some instances, commuters walked for more than one hour, sometimes starting well before 5:00 a.m., to reach the nearest station. Consequently, this lack of mobility infrastructure placed significant constraints on their mobility access.

Sedimentation of spatial policies: Public transport policies were described in similar terms since commuters reported that policies where either lacking or promised improvements were not implemented. Some examples are provided below.

They say that by next year they will launch a new system but I know that nothing is going to come of it. I don’t think it’s the trains that are the problem. I think it is the way it is being managed. […] Because a management that is focused on service delivery will not provide a poor service. They would maintain things well. They would ensure that the trains are running on time. Um, they don’t really care about the people. (J, 2)

That’s why, that’s why, their planning was really very, very bad for these lines. […] They really need to, they really need to improve their service. Funny enough, I actually saw the other day in the Argus [Cape Argus, a local newspaper] that they are planning to. But you know they always make plans and makes plans and make plans, and nothing ever comes from it. The Minister of Transport has, just the other day. There was an article in one of the newspapers: They are planning to do something but they never get so far to actually deliver anything. So, it’s really, it’s really a problem you know. It is a big problem and unfortunately this is the way it is. (W, 2/3)

When interviewees spoke of mobility options, they lamented the inadequate mobility planning of their mobility environment. According to interviewees, politicians often promised infrastructure investment, better services, or more alternatives, which rarely materialised. Many commuters believed that regional spatial policies were either insufficiently formalised or non-existent. Consequently, many commuters felt trapped in a deficient mobility environment characterised by the absence of a vision for current or future public transport. The stasis of their mobility environment and the lack of policy intervention to remedy it left commuters feeling abandoned by policy-makers, who, according to our interviewees, were unconcerned for their welfare. In this context, many of the commuters we interviewed had access to only one public mobility source, Metrorail, while the only other mobility option was walking.

Meso-Level: Inside the Metrorail System

Mobility options relating to the Metrorail system were located at the meso-level, particularly with services and infrastructure. Inadequacies of services manifested in several ways. According to our interviewees, commuter demand often outstripped carrying capacity, especially during peak hours, which resulted in severe overcrowding. Commuters also reported frequent delays and breakdowns, such as when “the train simply stops in the middle of nowhere” for long periods of time (J, 6), which contributed to discomfort, given the lack of facilities and summer heat or winter cold. They also mentioned that replacements for cancelled services were rarely provided, which further exacerbated overcrowding. Some examples from our transcripts are given below.

Yes, you see their name is not really Metrorail; their name is Metro-fail. No really, because you pay R133 [just over United States dollars (USD) $9] for a [monthly] train ticket, which is fair-and-square, which is nice; the cheapest transport that exists. But their service is pathetic … pathetic. Overcrowded trains, train delays, no announcements, you know. (W, 1)

Metrorail … Terrible service. Dirty trains. Um, affordable. Waiting. In the past thirteen days there was just one day when all the trains, and I use four trains every day, there was just one day when all the trains were on time, all of them. (J, 1)

It happened in the morning. I was writing [exams] that morning. So, the trains were delayed, and they had been delays from early in the morning. I had no other option ‘cause it was internal exams. Internal exams, my teacher shows no mercy. If you’re late, you’re late, you not gonna write. And it was June so I needed the marks to apply to university. People started clapping [she claps her hands loudly], the train came and people started to get on. I tried to get on, I tried, I fought and I fought. Then I could, one foot was on but the other foot was not. My bag was outside, my face was inside. I was holding on by the doors there, you know, onto the frame. I was holding by the door frame, so when the train was about to approach Bellville, it makes a turn but like a huge turn. I almost fell. If it was not for the person that was next to me, but a bit to the inside, I would have fallen. Cause this guy saved my life, he just grabbed me by my shirt and tie and held onto me. And then I couldn’t breathe because I have asthma. I had already given up, I was going to die. But he pulled me in and other people also noticed that I was fainting. There was, I don’t know what happened, I don’t know where the people went, there was space, like they made space. I was able to lie down and then they gave me a space to breathe, but I almost died. (Y, 8)

Frequently connected to the inadequate service was the dire state of Metrorail’s infrastructure. According to interviewees, trains were old and dilapidated, and the associated infrastructure was not maintained. Windows and seats were dirty, broken, or missing, and carriage doors did not function properly. Also reported were dilapidated train stations and platforms, malfunctioning communications and signal systems, and stolen copper cables that would bring the train system to a standstill. Some examples are provided below.

The trains themselves, look, there are some of the trains that are so dirty you can’t see out of the windows and they don’t tell you at which station you are arriving. So, if you don’t know the route and its dark then you just guess. Sometimes you can’t even see what’s going on outside unless you open the window and, in the winter, this can become very uncomfortable when it is cold. So, this is really a bad service. (J, 5)

Well, it can’t get any worse or much more anyway because they are just falling apart. I once reported a carriage that was so dilapidated that I thought it was unsafe. It felt like a wheel was going to come loose. These days it happens regularly that a train just stops in the middle of nowhere and then they say the train is broken. It can’t go any further. I don’t know what they do with it then, but it is just the service that becomes even more disrupted. (J, 6)

While services and infrastructure accounted for the majority of data located in this access cluster, we identified a third dimension that related to the way people acted as a constraint to other peoples’ access and use of trains. Some examples are provided below.

There were some days, sometimes when people have stolen the power cables. So the people can’t go to work for at least a day or will be late by two or three hours. So yes, I think this is actually a terrible experience, especially if there is work to be done. They steal the cables a lot, yes. (N, 2)

Interviewee: I remember it clearly. I was on the train with my mother, and this guy who came in like a normal guy, he then approached this pregnant woman and asked for her earrings, ‘cause she was wearing diamond earrings, beautiful earrings. Then, he kindly asked for the earrings, so obviously the lady said “no”. You know, like any other person. So, then he took out a gun, and then we all then realised, “oh, my gosh, this is a robbery”. Then he told us to get down on the floor. So, we all got down and then this lady was just refusing. Guess what? She took the earrings and threw them outside, and said, “go, fetch them if you want them”. Then the guy took her and threw her out [of the open door]. Yes, he threw her off. Like, that was the last time, that I used trains in my life. The last time. So, imagine with that trauma. (R, 3)

Based on the experiences of the commuters we interviewed, we identified three types of human constraints: a threat to commuters, a threat to train infrastructure, or a threat to the functioning of the mobility system. Examples of the latter included commuters accidentally falling out of open doors, from the roof of trains, from riding illegally between carriages, or from accidents caused by pedestrians or vehicles failing to obey railway signals. Threats to train infrastructure occurred, for example, when angry and frustrated commuters vandalised trains, or when cables were stolen and sold for scrap metal. The most frequently mentioned human constraint to access, however, related to personal safety because of “skollies” (Afrikaans slang for thugs or gangsters). While most reported incidents related to theft or armed robbery, the vast majority of the Metrorail commuters we interviewed had been affected by gang-related violence, assault, murder, rape, vandalism, or drunk and disorderly behaviour.

For all interviewees, unsatisfactory services, dilapidated infrastructure, and the possibility of becoming a victim or witness to a crime are part of Metrorail commuting. The experiences associated with services, infrastructure, and crime shape mobility system expectations, as system expectations shape mobility experiences. In this context, motility conditions such as the spatial distribution of mobility infrastructure and the sedimentation of spatial policies (or lack thereof) reveal how challenges constraining the mobility of the Metrorail users were not limited to Metrorail services but extended to the greater mobility environment in which Metrorail is embedded.

Micro-Level: People and Their Environment

The actual and potential capacity of commuters to be mobile at the micro-level of motility access connected to the commuters themselves and was mainly associated with the means of transport and their socio-economic position. The characteristics defining macro- and meso-levels intersected with commuters’ mobility needs and aspirations. To examine how this manifested, we sorted interviewees’ mobility preferences according to means of transport and the meanings they attach to them. Based on our analyses, we found that preferred mobility choices formed a distinct hierarchy, as depicted in Figure 3.

Figure 3. Preference of means of transport as ranked by Metrorail users.

Private cars occupy the top of this pyramid, whether or not the interviewees had access to a car. Cars were valued for their convenience, flexibility, comfort, reliability, and status. Instances in which Metrorail users resorted to car use, often shared, included important meetings, early-morning appointments, late-night events, weekly grocery shopping, special celebratory occasions, and transporting children. Indeed, most commuters reported that they would not take children on a train. Some examples are provided below.

Interviewer: Why don’t you use trains more often?

Interviewee: Because using the car is more convenient. If I don’t have to sit and wait for transport then I am not going to do it. When I can just get in and go, that is what I would do, that will always be my first choice. (E, 4)

If you miss a train, then you have to wait. So how do you deal with that when you need to get to work? With your car, you can make up some time, you can give the throttle a little bit. That’s the only thing with trains, yes. […] Now these days, I don’t use it anymore because I have to do a lot of things so I have to save time. But I also have to save money but I can’t save money. Because with our trains you can’t save time so you actually waste time so that’s why I don’t use the trains anymore because I have to use my car to move around. (N, 1/3)

Other forms of public transport, such as buses and especially 8–30-seater minibus “taxis”, occupy the second tier of the pyramid, as approximately 14 million South Africans use taxis daily. Reported as less convenient than private cars but still faster, safer, more efficient, and in closer proximity to where commuters lived, taxis have a fixed route but are flexible where they pick up and drop off passengers along their route. Taxis are also used to get to and from train stations. Thus, a work commute usually consists of walking before catching taxis or trains—nearly 10% of South Africans commute to work or school for three hours or more per day. Our interviewees tended to limit their train use to work commutes and used alternatives, especially taxis, on weekends. Overall, commuters reported that they prefer buses and taxis to trains. Some examples are provided below.

Interviewer: And do you think trains will be part of our future?

Interviewee: It better. We need it. We have too many cars on the road. Public transport is really terrible. I think the city of Cape Town is trying to do something about it with this new fast-tracked bus service. Um, unfortunately, it is not in the area where I live but what I have heard about it is that it is a very good service. (J, 6)

Interviewee: No way, forget it [he shakes his head]. No, I have now, like I said, only take it to work and back. […] When I am home, I would rather take a [minibus] taxi or a bus. Interviewer: Why would you rather take a taxi when you are at home?

Interviewee: Because it is quicker. The taxi is quicker and it picks you up near your home, and it drops you off near your home, and so on. So, if you have to go quickly, 10, 15 minutes, if you want to be somewhere quickly, then you take a taxi and of course if you want to go to the city it a naturally a completely different story. A completely different story. (W, 5)

I prefer the bus. […] It’s safer. You are almost guaranteed that you’ll get a seat. And it doesn’t really run late, and you get to school on time. And when it’s raining, you don’t worry about it delaying or being cancelled. (Y, 9)

Metrorail, located at the bottom of the pyramid, was considered an affordable but least preferred mobility option. Affordability was reported as its most positive attribute, mentioned by nearly all interviewees. Some examples are provided below.

But we need the trains. There are thousands of people that use it every day and they don’t have any other transport. (J, 6)

Their price is really reasonable. It’s much cheaper that what it would be with a car or a bus. As the price of petrol continues to rise more and more, people will become dependent on the train. (A, 3/4)

[Metrorail] is inconvenient versus cost-effective. In the sense that it is always late, it is always overcrowded, and it is always wet or it is always too hot. (G, 1)

Metrorail’s affordability, however, was also considered a trap toward dependence on the most underserved, inconvenient, and unreliable form of transport as stagnating wages, high unemployment rates, and rising petrol and food prices drained household budgets and made it less likely for the state to invest in, and commuters to afford alternatives.

By ranking mobility types according to preferences, we identified characteristics that related to how mobility modes are defined and separated. Privately owned cars were valued for their convenience, flexibility, comfort, reliability, and status. Buses or taxis offered no social status and much less flexibility and autonomy, but provided reliability and convenience due to geographic proximity, availability, and speed. Metrorail lacked these qualities and was, thus, unable to fulfil commuters’ needs. Cost was an additional quality associated with this hierarchy. At the top of the pyramid, Metrorail users with access to a car frequently referred to trade-offs between cost and convenience, or cost and necessity. In the middle of the hierarchy, commuters reported how increasing fuel prices made alternative modes of public transport an attractive albeit inaccessible alternative, while most commuters at the bottom of this pyramid depended on Metrorail because it was the only or cheapest mobility option (W, 1). Thus, economics and status further defined this pyramid. While mobility preferences shaped the content of this hierarchy, it was the purchasing power and, by implication, status aspirations that ultimately maintained clear boundaries between the three mobility modes, evidenced by how our interviewees separated people into three categories: the poor, who walk or are stuck with Metrorail; “normal” people, who, depending on situation and context, may flexibly select from a variety of mobility modes; and the rich, who own and use cars.

Metrorail commuters who owned cars deliberately and selectively used train travel as an occasional and alternative mobility option, as the following two examples illustrate:

Um, I think it is really nice to take the train. I enjoy it. It’s relaxing, it’s a little bit more personal as opposed to when you sit in your car on your own. You know, it’s actually nice to be part of the community to go in and out of the community; to do some people-watching. And it is a, I don’t come from a background where people use a lot of public transport, and I find it interesting to see how other groups of people make their way through the day. I definitely prefer taking the train than driving. (M, 1-2)

Um, people get into trouble at work because they are always late. Often, I can go back and fetch my car, and go with my car. But many, thousands of people, don’t have a car that they could take. This is their only transport. I use it because it is cheaper and because I can read while on the train. I can’t read while I am driving. (J, 2)

Car owners who commute by Metrorail often took the more expensive “business express”, reporting that they enjoyed the train commute because it was considered relaxing; they utilised travel time to read, sightsee, or prepare for the workday. They also mentioned that trains offer a cheaper mode of transport although they emphasised that they did not use trains to save money. Among this group, train cancellations were considered an annoyance that could be overcome by using their private vehicles. For important events or when they needed to travel during rush hour, they preferred using their cars.

Using the terminology of our interviewees, the “normal” people were located between the rich and the poor. However, the “poor” is a misnomer because people who must travel are not the poorest in South Africa. Metrorail commuters tend to be employed, are actively seeking employment, or they have resources to access education, apprenticeships, distant family and friendship network members, and so on. Even a dysfunctional mobility system is an improvement over not having a system at all. Living near a station is substantially better compared to communities without Metrorail and its feeder system. Metrorail offers a low transportation stratum in a hierarchy, but many people would be unable to go to school or work, see a doctor, or visit friends or family without it. Thus, when our interviewees refer to the poor, they mean those who have funds and a reason to take a train. Due to limited financial resources, the poorer commuters use Metrorail as their main form of transportation. The poor nevertheless reported access to alternatives, such as borrowed or shared cars, buses, taxis, or lift clubs. Their purchasing power was severely constrained and the costs associated with using alternatives required sacrifice and careful calculation. While the so-called rich and normal people were more flexible with regard to their mode of transportation, the former selected trains only for carefully selected and controlled trajectories; otherwise, they would use a car. The latter created a transportation patchwork based on specific contexts and motility needs, underpinned by a cost–benefit analyses of mobility options. Two examples are provided below.

Yea, yea, yes, I take it [the train] Monday to Friday, weekends I don’t bother with the trains at all, like I’ve told you. We prefer to take the vehicle on the weekend of course. It is going to work out more expensive but you can do so much more with the vehicle because then you can do your shopping and things like that. (W, 6)

And I have really realised that it is very good to take the train. Because you can save so much money, especially if you want to go out more and if you want to … Like in the past, I couldn’t even afford to go out for breakfast or lunch. You know, on weekends, if I want to relax, then I take my baby out or we go on an outing. And before I always had to think about the petrol money and, now, I can just do it. Because I know that I just have to buy a [train] ticket and then I know I don’t have to worry throughout the month, if I have to go to this place or that place. (G, 7)

The poor were associated with the bottom of the mobility hierarchy. They lacked the purchasing power to afford alternatives and found themselves locked into an ailing and unreliable train system. Some examples are provided below.

You see, I am actually one of those fortunate ones because those other people have to also do their whole shopping with the trains right? They are not as fortunate as some of us … Just put yourself, try to put yourself in those people’s shoes, that person who lives in Bonteheuwel [a township near Cape Town]. He has to go to Cape Town and now he has his whole months’ worth of groceries or whatever. I mean when you have a vehicle you just load it into the vehicle and there you go. But now you sit with seven or eight bags that you now have to hold onto and oh no. No, thank you. Forget it. (W, 6)

We pay, even though we pay less, but we pay. There are so many commuters. We buy so many monthlies [monthly tickets]. How much money does Metrorail make? Why can’t they do that? Why can’t they give us a comfortable, convenient place to sit in? (G, 7)

These narratives illustrate the significant relationship between economics and mobility choice. As commuters’ purchasing power diminished, so too did their access to mobility modes and variability in mobility choices. A growing dependence on this cheap but unreliable mode of transport was associated with a decline in autonomy, security, status position, and life options. Finding themselves at the mercy of a dysfunctional mobility system, interviewees often expressed despair or resignation, as the following interviewees recounted:

Um, they don’t really care about the people. Sometimes, when we are late, there are students on the trains who are supposed to be writing their exams. Once it happened that someone had the keys to open the bank and they are sitting with the keys, stuck at Muldersvlei station and can’t take the train. Um, people get into trouble at work because they are always late. (J, 2)

For example, yesterday morning. My train is at 5:45 a.m. No announcements, nothing. The train arrives at 6:30 a.m. Do you see? Now I have to let the people at work know that the train is late, but they don’t understand. It’s very frustrating. (W, 1)

And there are some people who rely on the train to get to get them to work. And there are some employers who don’t understand that sometimes the trains are just late and then they start to become difficult. And, so, it is possible for you to lose your job just because you use the train every day and because you need to trust it because it is your only transport that can get you to work. (A, 1)

It was this sense of capitulation and precariousness at the bottom of the mobility pyramid that was particularly striking. Although the relationship between socio-economic status and mobility access is not new or surprising, we observed a relationship between economic position and access that was not limited to how poor people lacked access or choice due to diminished purchasing power. Rather, we identified a poverty trap at the bottom of this mobility pyramid that locked commuters into the Metrorail system. In a vicious cycle, people are trapped in an under-priced, affordable mode of transport that is unreliable and dangerous in ways that render their work and lives precarious, as outlined below:

  • Metrorail cannot raise prices because its main customer base cannot a
    afford a price hike and, worse, its infrastructure could be sabotaged because some customers would consider a price hike unjustified and unfair, given Metrorail’s unsatisfactory service.
  • Metrorail, thus, cannot modernise or expand services, which leads to further dilapidation, unreliability, and infrastructure decline.
  • Given the high unemployment rate and availability of unskilled or semiskilled workforce in South Africa, employers have little patience with the poor who show up late or miss work due to a transportation breakdown. Long commutes, often in anticipation of delays or cancellations, in association with compromised diets and a poor health status among the most disadvantaged, further drain the energy and capacity of workers. Similar conditions and consequences are faced by commuting high-school or university students from the lowest income bracket.
  • Especially commuters without alternative mobility options are, thus, caught in a trap because they are too poor to afford alternative and more reliable forms of transportation, causing the inability for Metrorail to improve its services and infrastructure.

In sum, the dysfunctional public transportation system available to the poor keeps them locked in an unreliable transport system, which locks them into precarious education, work, and health situations, which keeps them locked in a precarious life situation, which keeps them locked in an affordable but unreliable and declining mobility system, which is unable to raise standards or increase service offers due to a lack of funding because their main customer base, the poor, are unable to afford higher prices.

An Exception to the Rule

As with most complex social phenomena, there are always exceptions to the rule. A good example here was the business train Metrorail provides in the Western Cape region. If we tried to fit this train into our mobility hierarchy, it would be placed at the top of the middle tier, above taxis and buses, but below private vehicles. This service provides a luxurious commuting alternative for business people at a premium price. This sub-division of Metrorail fits between cost and mobility mode and its association with comfort, convenience, reliability, and safety. Ironically, by removing affluent commuters from the first class section of a dysfunctional train system, and by developing a parallel train infrastructure, the business train side-steps the dysfunctionality of Metrorail’s other trains at the cost of regular passengers. Worse, an affluent commuter class with a strong political clout and voice is, thus, eliminated from participating in reforming Metrorail. Affluent commuters are successfully pushing for an expansion of a privileged service, often subsidised by government funds, at the cost of mobility reforms that would benefit the majority of the population. The following examples illustrate the parallel train system:

You see there has always been this thing that has been bothering me. ‘Cause I also took trains from Kuilsriver to Mitchell’s Plein, so that line every time, every day, Monday to Friday at seven o’clock, there was a business class train that went past. In a month that train went delayed about two to three times maximum. That’s the only time it would delay. But, and the ticket they pay like, it’s huge, I think it’s R1000 [approximately 70 USD] a month and everyone in that train gets a seat. It’s seven o’clock, every day, everyone has a seat, everyone has a cup of coffee, and everyone has newspapers. I understand that they can afford it but why is it that at, seven o’clock, the business class train is always on time, every day, and then the other trains just go whenever and however. That’s always been one of the things bothering me. Like is it because we pay R150 [approximately 10 USD] a month? Is it because, I never understood, that dilemma, because if they say, “good morning Metrorail users, trains are delaying due to cable theft, maybe in so and so, Metrorail apologises for the inconvenience”. Seven o’clock, the [business] train comes! Ahh, wasn’t there cable theft? And I’m like “Whaaat?! What’s happening? It’s not right”. […] Why should there be a business class train that comes at this time for such people, and then our train delays all the time. (Y, 5-6)

And these days they have these business class trains that provide a service to the business people, but why can’t they also provide this business class to us normal people? Because see now, these people get there in the morning, then they get a cup of coffee. They come and sit there and then they get a newspaper. But you pay extra for that train. You pay R1000 a month or something, I’m not sure. (W, 7)

Discussion and Conclusions

In this article, we analysed the access component of the motility framework, but also the greater mobility context within which access is embedded. This allowed us to link mobility preferences and behaviours with a mobility context and environment. That Metrorail is not fulfilling most commuters’ expectations is well documented, and we embarked on this study with the understanding that the problems and challenges Metrorail commuters experience would be a central characteristic of our data. Our aim was not to catalogue characteristics of a dysfunctional system, but to better understand how commuters connect to, and integrate dysfunctionality into their mobility needs and aspirations. The ability to be mobile is inseparably linked to needs and aspirations, whether they are personal, social, educational, economic, and beyond.

The motility framework we draw on proposes that people’s actual and potential capacity to be mobile is best understood by examining the space where spatial and social mobility intersect. A defining feature of mobility is access, a necessary but insufficient condition for mobility. Our study revealed a context- and culture-sensitive mobility pyramid, whose tripartite nature is formed by access, economic position, and the resulting degree of flexibility. This pyramid is not a mere representation of interlocking mobility networks in the Western Cape region. It reveals how dimensions of safety, convenience, dependence, reliability, status, and autonomy converge in a hierarchical mobility structure.

Also striking in our study was the emphasis on economic resources, rather than ethnicity or race, with which commuters differentiate themselves and others. While ethnicity is mentioned occasionally, it was not used as a consistent marker of differentiation. Over the past two decades, the lines of demarcation between class and ethnicity somewhat diverged in South Africa, and it was more difficult to differentiate narratives about the rich, normal, and poor people versus White, Asian, Coloured, and Black South Africans. There were rare instances where rich signified white, and everyone else (non-whites) was labelled as either normal or poor. However, this euphemism was exceptional in our data.

In some of our previous work, we emphasised the dangers of dismantling Metrorail and replacing it with a modern mobility system, such as the Gautrain (a modern commuter rail system linking, among other things, central business districts, universities, shopping malls, upmarket residential areas, airports, and, still to be realised, some townships between Johannesburg and Pretoria). In many ways, the Metrorail’s business express trains seem to mimic an exclusionary, parallel mode of train mobility—it is indeed comfortable, reliable, and safe, but it is also prohibitively expensive for the majority of South Africans, and, although the situation is improving, it still does not serve enough areas where the majority of South Africans live and work. Upmarket train systems risk creating and maintaining new forms of social and spatial segregation, shifting a segregated society based on race to a segregated society based on class, while concurrently reinforcing and possibly exacerbating poverty and exclusion for the majority.

In this sense, it is the dysfunctionality of Metrorail that will keep it going because it remains the most affordable form of public transport for the masses. It will remain affordable, even to poor people with jobs, and it is its cheapness of this mode of transportation that, in the absence of major structural reforms, imposes dysfunctionality on Metrorail. Thus, Metrorail is an enabler for poor people to reach distant jobs; however, given the strong relation between spatial and social mobility, it concurrently represents a hurdle to overcome in order to escape the bottom of the pyramid.

We began this article by introducing some of the main tenets of sustainable mobility as represented by the green transport hierarchy. This hierarchy, implicitly or explicitly, serves as the guiding framework for how sustainable mobility ideals are conceptualised in mobility studies and policy documents. Comparing the mobility hierarchy of our Metrorail users with the upside-down pyramid of the green transportation hierarchy, we observe how experiences and aspirations in the Western Cape are remarkably different from those propagated by the sustainable mobility paradigm. Comparing the mobility preferences of our interviewees with the green transportation hierarchy, we find that they are antithetical. At first glance, we may conclude that our commuters’ mobility choices and aspirations are unsustainable because of their reliance on, or preference for fossil-fuelled and individual-based mobility modes, especially cars. However, is this really the case? What would happen if we used sustainable mobility ideals as the benchmark to evaluate the mobility context in the Western Cape? What would the implications be for the Metrorail commuter context?

One of our interviewees lives in Khayelitsha, an informal settlement on the eastern fringe of the city of Cape Town, just over 30 km from the central business district. She shares a 6-m² shack with her infant and her aunt. Her aunt is 53 years old and unemployed. She cannot read or write, and she depends on whatever income her niece generates to survive. She neither owns a bicycle, nor does she use public transport, and she certainly has no access to a private vehicle. She rarely leaves the township, except for Sundays, when she walks to the nearby veld to attend an outdoor church service. During the week, she spends most of her time taking care of our interviewee’s son and household. Every morning, she walks to the nearest water point to collect water for the day, which she carries back in a bucket balanced on her head. Some days she may do this several times. Occasionally, she walks to a nearby spaza shop, where she may buy phone credit or maize meal to make mieliepap for dinner. According to the green transportation hierarchy, and given that her mobility practices are limited to walking, she would be considered far more sustainable, compared to our interviewee, who uses taxis and trains to get to work a few times per week.

Although this is only an illustrative example, it exemplifies the challenges in applying Eurocentric concepts that are embedded in value systems and infrastructure availabilities that are difficult to translate into other contexts. That the aunt of our interviewee is attributed with a high level of sustainability in terms of her mobility practices is problematic, especially given that South Africa’s official unemployment rate is at 27.2% (StatsSA 2018) and youth unemployment at 54.3% (Mcgregor 2017). People living in circumstances such as these cannot be thought of as sustainable, even from the narrow perspective of the green transportation hierarchy. Thus, sustainable mobility must be adapted to reflect regional contexts and cultures in order to avoid being reduced to the wishful thinking of an educated, liberal, East Coast or European urban elite. While there is a lot of mileage in the sustainable mobility hierarchy, it needs to be adapted to reflect different contexts and cultures. In our case study, for example, sustainable mobility theory needs to take into account the lack and skewedness of mobility infrastructure in this developing and highly unequal society. Of course, if the first step toward sustainable mobility is to create an equal society and a well-funded and maintained mobility infrastructure, then the green transportation hierarch mobility model would work. However, if we want to pursue a greater degree of sustainability when considering people’s mobility in the interim (and, by extension and based on the motility framework, make the mobility of goods, information, and technology more sustainable), then we have to revisit the green transportation hierarchy to take into consideration, first, regional contexts and cultures and, second, the extent to which some groups, regions, and countries need to develop economically in order to become more sustainable—across mobility modes and well beyond.

This case study provides insights into the challenges and opportunities defining a specific, encultured mobility landscape, thus providing important considerations for the theoretical assumptions upon which we base our sustainable mobility aims. This study could serve as an important impetus from which to develop context- and cultural-relevant, large-scale studies on mobility practices and mobility development. There are multiple, inter-locking contextual and cultural characteristics that shape people’s actual and potential capacity to be mobile. The effectiveness of sustainable mobility interventions will depend on policies that can adequately account for these variations. In the Metrorail context, this would mean creating policies which simultaneously address contemporary and developmentally inspired socio-spatial dynamics. Given the poverty trap and resulting dependency cycle, an obvious recommendation would be to initiate structural reforms, including a baseline mobility access to ensure that commuters have access to a functioning, accessible, safe, and affordable public transport infrastructure. Improving the status quo, especially in relation to safety and reliability, could be funded by eliminating a parallel “business train” system. In addition, it is necessary to reformulate sustainable mobility theory and policies, which currently tend to privilege reducing environmental impact over the necessity of socio-economic development. The massive expansion of mobility of goods, information, technology, and people in the near future will take place mainly in the poorest and unequal societies. Sustainable mobility will only be attained through a global, national, and local commitment, through public and private partnerships, and through a careful balancing of socio-economic and environmental concerns. If sustainability privileges environmental protection over social and economic development, it risks being considered a luxury concern of educated, urban elites, especially in developing economies.

References (in order of citation)

  1. Ban, K.-M. Welcome Address. 38th Meeting, High-Level Segment—High-Level Political Forum on Sustainable Development 2016. Available online: high-level-segment-high-level-political-forum-on-sustainable-development-2016/5043498338001/?term= &lan=french&page=5?term (accessed on 17 December 2018).
  2. Geerken, T.; Vercalsteren, A.; Borup, M. Review of the mobility domain. In System Innovation for Sustainability 2: Case Studies in Sustainable Consumption and Production—Mobility; Geerken, T., Borup, M., Eds.; Greenleaf Publishing Limited: Sheffield, UK, 2009; pp. 6–26. ISBN 978-1-906093-23-5.
  3. Holden, E.; Linnerud, K.; Banister, D. The Imperatives of Sustainable Development. Sustain. Dev. 2017, 25, 213–226. [CrossRef]
  4. Greene, D.L. Sustainable transportation. Int. Encycl. Soc. Behav. Sci. 2001, 15335–15339.
  5. Holden, E.; Linnerud, K.; Banister, D. Sustainable passenger transport: Back to Brundtland. Transp. Res. Part A Policy Pract. 2013, 54, 67–77. [CrossRef]
  6. Richardson, B.C. Sustainable transport: Analysis frameworks. J. Transp. Geogr. 2005, 13, 29–39. [CrossRef]
  7. Charlton, C. Social Change and Sustainable Transport. J. Transp. Geogr. 2004, 12, 165–167. [CrossRef]
  8. IPCC. The IPCC Fourth Assessment Report on Climate Change. Synthesis Report; Intergovernmental Panel on Climate Change (IPCC): Geneva, Switzerland, 2007.
  9. Holden, E.; Høyer, K.G. The ecological footprints of fuels. Transp. Res. Part D Transp. Environ. 2005, 10, 395–403. [CrossRef]
  10. Hunecke, M.; Haustein, S.; Grischkat, S.; Böhler, S. Psychological, sociodemographic, and infrastructural factors as determinants of ecological impact caused by mobility behavior. J. Environ. Psychol. 2007, 27, 277–292. [CrossRef]
  11. Banister, D. The sustainable mobility paradigm. Transp. Policy 2008, 15, 73–80. [CrossRef]
  12. Holden, E. Ecological footprints and sustainable urban form. J. Hous. Built Environ. 2004, 19, 91–109. [CrossRef]
  13. Holden, E.; Gilpin, G. Biofuels and Sustainable Transport: A Conceptual Discussion. Sustainability 2013, 5, 3129–3149. [CrossRef]
  14. Bradshaw, C. The Green Transportation Hierarchy; Ottawalk and the Transportation Working Committee of the Ottawa-Carleton Round-table on the Environment: Ottawa, ON, Canada, 1992.
  15. Bergman, M.M.; Bergman, Z. The imaginary of the train of the future in South Africa: From public access to social exclusion. World Transp. Policy Pract. 2015, 21, 23–34.
  16. Bergman, Z.; Bergman, M.M.; Pirie, G. Die Zukunft der Bahnmobilität in Südafrika: Vom Service Public zur sozialen Exklusion. Afr. Bull. 2015, 156, 6–7.
  17. OECD. All on Board: Making Inclusive Growth Happen; Oecd Publishing: Paris, France, 2015; ISBN 978-92-64-21849-9.
  18. Young, D.; Keil, R. Reconnecting the disconnected: The politics of infrastructure in the in-between city. Cities 2010, 27, 87–95. [CrossRef]
  19. Kaufmann, V.; Bergman, M.M.; Joye, D. Motility: Mobility as capital. Int. J. Urban Reg. Res. 2004, 28, 745–756. [CrossRef]
  20. Cass, N.; Shove, E.; Urry, J. Social Exclusion, Mobility and Access. Sociol. Rev. 2005, 53, 539–555. [CrossRef]
  21. Dupont, V. Socio-spatial differentiation and residential segregation in Delhi: A question of scale? Geoforum 2004, 35, 157–175. [CrossRef]
  22. Friedmann, J. The World City Hypothesis. Dev. Chang. 1986, 17, 69–83. [CrossRef]
  23. Grengs, J. Does Public Transit Counteract the Segregation of Carless Households? Measuring Spatial Patterns of Accessibility. Transp. Res. Rec. 2001, 1753, 3–10. [CrossRef]
  24. Power, A. Social inequality, disadvantaged neighbourhoods and transport deprivation: An assessment of the historical influence of housing policies. J. Transp. Geogr. 2012, 21, 39–48. [CrossRef]
  25. United Nations. Transforming our world: The 2030 Agenda for Sustainable Development; United Nations: Geneva, Switzerland, 2016.
  26. Geurs, K.T.; van Wee, B. Accessibility evaluation of land-use and transport strategies: Review and research directions. J. Transp. Geogr. 2004, 127–140. [CrossRef]
  27. Hansen, W.G. How Accessibility Shapes Land Use. J. Am. Inst. Plan. 1959, 25, 73–76. [CrossRef]
  28. Bergman, Z.; Bergman, M.M.; Thatcher, A. Agency and Bandura’s model of Triadic Reciprocal Causation: A case study of mobility among Metrorail commuters in the Western Cape, South Africa. Front. Psychol. 2019, 10, 411–425. [CrossRef] [PubMed]
  29. Pirie, G. Sustainable urban mobility in ‘Anglophone’ sub-Saharan Africa. In Planning and Design for Sustainable Urban Mobility: Global Report on Human Settlements 2013; UN-HABITAT: Nairobi, Kenya, 2013; pp. 1–44. ISBN 978-92-1-132568-3.
  30. Pirie, G.H. Virtuous mobility: Moralising vs measuring geographical mobility in Africa. Afrika Focus 2009, 22.[CrossRef]
  31. Bergman, M.M. Content Configuration Analysis; StudyCube; University of St. Gall: St. Gall, Switzerland, 2011.
  32. Bergman, M.M.; Bergman, Z. Perspectives of learners and teachers on school dysfunctions in South Africa.Educ. Chang. 2011, 15, 35–48. [CrossRef]
  33. StatsSA. Quarterly Employment Statistics (QES), seCond Quarter 2018; StatsSA. Available online: http://www. (accessed on 20 January 2019).
  34. Mcgregor, M. Labour Law Rules! 3rd ed.; Siber Ink: Cape Town, South Africa, 2017; ISBN 978-1-928309-14-7.