Koen van der Meulen | 10831347
S H A R E D S P A C E I N A M S T E R D A M
Successful innovation or dangerous experiment?
Supervisor: Danielle Chevalier
10 July 2017
University of Amsterdam
Urban sociology
Content
Summary ... 3
1. Introduction ... 5
2. Theoretical framework ... 8
2.1 History of shared space ... 8
2.2 Interactions in public space……….8
2.3 production of space ... 12
Hypothesesis ... 15
3. Methodology ... 17
3.1 Outline of research ... 17
3.2 Strength and weaknesses of methods ... 19
4. Operationalization of key concepts ... 19
4.1 Structured observations ... 19 4.2 Interviews ... 21 5. Data ... 21 5.1 Quantitative ... 23 5.1.1 Dependent variables ... 21 5.1.2 Independent variables ... 21 5.1.3 Analytical strategy ... 23
5.2 Qualitative ... Fout! Bladwijzer niet gedefinieerd. 6. Results ... 24
6.1 Quantitative results ... 24
6.1.1 Interactions and locations……….……….24
6.1.2 collisions and near collisions ... 31
6.1.3 Sex, age and mode of transportation ... 33
6.1.4 timeframes ... 38
6.1.5 dismounting ... 41
6.2 Interview results ... 41
6.2.1. Respondents on lay out of space ... 41
6.2.2 Safety ... 42
6.2.4. How respondents used space... 45
6.2.5 planned space versus lived space ... 46
6.3 Field notes ... 47 6.3.1 Shared Space ... 47 6.3.3 Muntplein ... 49 6.3.4 comparing spaces ... 49 7. Analyses ... 50 7.1 Quantitative analyses ... 50 7.2 Interview analyses ... 53
7.3 Field notes analyses ... 54
8. Conclusion ... 55 9. Discussion ... 56 9. Literature ... 59 10. Appendix 1 ... 62 Appendix 2 ... 63
Summary
This thesis researched the in 2015 finished Shared Space behind the Amsterdam central station. A Shared Space is an area were all clear traffic regulation and boundaries are removed and people have to direct their own way through that space without the normal regulation to tell them how. The aim of this research was to find out if and in what way this new space changes the interactions between people, their experience in that space and if this was different in comparison with two other comparable locations in the city. These locations were picked because the first location, the Haarlemmerstraat, was moderately regulated. And the second, the Muntplein, was heavily regulated.
There were three methods used to research these three locations. The first being a quantitative comparison by conducting highly structured observation which compared the interactions between the three locations numerically. The second were semi structured interviews with five respondents on each location which compared the location in a more qualitative manner. And lastly, qualitative observations were kept to record any additional information that could not be registered by the other two methods.
This thesis starts with an introduction to the topic and the research locations. Here earlier research to the concept of the Shared Space will be explained and how it motivates and directs this thesis. It ends with the research question and a number of sub research questions.
After the introduction the theoretical framework will follow. The first part consists of the history of the Shared Space. And an explanation of why the municipality chose for the Shared Space, what theory is behind the idea and what goals the municipality has for the area. After that the theories that will be used in the rest of the research will be described and will then end with a number of hypotheses.
Next the methodology will outline in depth were the research locations are, why they were chosen and what methods will be used to research them. Here the strengths and weaknesses of the methods will also be discussed.
In the operationalization of the key concepts that follows these methods will be further
explained and the theory. This chapter will explain how these methods are operationalized by using the theory and in what form they aim to research the topic.
The following chapter describes the data. Beginning with the quantitative data end ending with the qualitative data. Detailing what the dependent and dependent variables are and how many
observations each variable has. And it specifies how the interviews will be coded and processed. The results section will show the outcome of these three research methods. Beginning with the quantitative part, than the interviews and ending with the field notes.
research questions. But also by comparing the theory with the results and checking whether the hypotheses should be rejected or not.
In the conclusion the main research question shall be answered. In the discussion the methods will be discussed. But there will also be a review of what the research achieved and what it is lacking. Lastly the direction future research on the Shared Space should take based on the results of this thesis will be discussed and some recommendation will be made.
1. Introduction
Since 21 November 2015 the space behind the Central station in Amsterdam has been transformed into a so called Shared space (V&OR, 2016, p. 7). A Shared space is a space were all kinds of clear traffic notifications like traffic lights, boundaries, traffic signs and zebra crossings that would normally separate traffic have been removed. By doing so users of this (Shared) space share all available space and cannot depend on conventional regulations which tells them where to be and how the behave (Moody & Melia, 2012, pp. 1-2). The hypotheses this concept is based on is that it makes people more observant, less prone to risk taking behavior, communicate more, experience greater risk and therefore, paradoxically, make that space just as safe, or safer as a conventionally regulated area (Hamiltion & Balie, 2008, p. 171). This is not the first shared space in the Netherlands. As there are shared spaces in the provinces of Groningen and Friesland (Van Veen & Nota, 2011). This is however the first of its kind in the city of Amsterdam.
Before the introduction of the shared space the area behind central station was used by cars, bikers, pedestrians and other motorized vehicles like scooters and mopeds. There was a road strictly meant for cars and a bicycle track next to it. Three zebra crossings where put over the cycle track and car road and all of them had traffic lights for pedestrians that used the ferry to go from the north to the center or vice versa (V&OR, 2016, p. 8). With the increasing popularity of the north of Amsterdam both as a residential zone and recreational area the strain on this space grew yearly. The municipality
eventually chose to transform the area into a shared space because they hoped it would lessen the train on the area and as a test to see if this could work in Amsterdam so that, if proven successful, it might be introduced on other squares (Litjens, 2015).
At first there where many who opposed the idea of the shared space. Both the media and
scholars expressed their doubts. They felt that Amsterdam was too busy for something as experimental as a Shared space. The critique ranged from headlines in the media like: ‘’crazy traffic civil servants will kill cyclists in Amsterdam’’ (van Rossom, 2015), to more nuanced critiques by scholars like UvA
communications teacher Reindert Rutsema who started a petition against the shared space. He believed that the area was too busy and that the planned introduction of the shared space would only create chaos (AT5, 2015). Yet despite some of these early negative reactions the construction of the shared space was finalized in 2015 (V&OR, 2016, P.7) . As seen in the map above the car road was completely removed and the only road for car traffic is the road underneath the station and the half circle on the left part which is strictly meant for taxis. The shared space constitutes the large circled area between the ferry’s and cycle tracks. There is no clear structure in that area for people that arrive to board and disembark from the ferry’s. Nor are there clear signs or boundaries for pedestrians, cyclists and motorized vehicles to pass through the shared space area (V&OR, 2016. P.8). Cyclists, pedestrians and small motorized vehicles thus share al available space within the circle and have to plan their own movement trough this part of public space in order to avoid collision.
After one year of shared space in Amsterdam the negativity by the media seemed to have largely disappeared. The newspaper Parool and branded the experiment as a success. There was no increase in accidents and they reported that people in general seemed to have accepted the new shared space without any problems (Kruyswijck, 2016). The NOS also reported its success and used a small time laps video of a few moments to prove this (NOS, 2016). However, lately there have been two instances of reported accidents on the shared space that have reignited some of the initial concerns. With one hospitalized scooter driver as the result (At5, 2017).
While this initial change in media discourse seemed like a good indications that the shared space is not the disaster the media initially claimed it to be, none of these claims were based on actual local scientific proof. Early research has shown that the introduction of the Shared space does lead to feelings of unsafety, chaos, uncertainty and indecisiveness in users (De Haan, 2012, p. 5). Yet according to De Haan (2012) it is still too early to make definite claims about user experiences of the Shared space because these researches have been far and few between and were often not very extensive. Furthermore there has been research by a social psychology student in Groningen which seems to confirm the hypotheses that people in shared spaces become more observant and communicate more. And that this might be the reason why the shared space is successful. They arrived at this conclusion by comparing it with conventionally regulated spaces and observing 300 respondent per location (de Vries-Koopmans, 2012). Yet such extensive research has not yet happened in Amsterdam. So far there has
been one local research done by the municipality of Amsterdam. Were they measured the amount of collisions in the beginning of the shared space, after 4 weeks and 10 weeks. Their conclusion was that the number of (near)collisions did not increase and that the shared space would not have to be altered (V&OR, 2016, pp. 19-20). While this result is certainly interesting, for it proves that people do not lose focus over time (which was one of the initial fears), based on these results one cannot conclude anything about this space in comparison to other spaces. Is the shared space really safer than a conventionally regulated space and if so why? Does the shared space make people more observant and communicative? And how do people experience the shared space? Trying to fill the lack of knowledge about this new way of planning space in Amsterdam will thus be the aim of this thesis research. To answer this the following research question will be answered:
‘’ How does the introduction of the shared space behind Amsterdam central station influence the interactions in that space, the way in which people experience it and does this differ from other conventionally regulated spaces?’’
This main research question will be answered by a number of sub research questions:
1. ‘’ Is there a significant difference in interactions between the shared space and similarly regulated spaces?’’
2. ‘’ Is there a significant difference in collision and near collisions between the shared space and the conventionally regulated spaces?’’
3. ‘’ Are there significant differences between sex, age and modes of transportation and the amount of interactions used by respondents on the Shared Space?’’
4. ‘’Are there differences between specific timeframes and the way in which the shared space is used’’? 5. ‘’How does the experience people have in the shared space differ from the conventionally regulated spaces?
2. Theoretical framework
2.1 History of shared space
The believe in the city’s ability to organically regulate itself can be traced back to the battle between the famous Keynesian urban planner Robert Moses and his biggest critic Jane Jacobs. Robert Moses believed that New York was chaotic and needed to be thoroughly regulated. He believed that this could, and should, be done by large urban renewal projects within the city. Moses is seen as one of the more famous top down urban planners. He believed that by regulating and redesigning large parts of the urban landscape chaos could be transformed into an orderly and logical city. Jane Jacobs opposed this idea. She believed that the city was much more able to organically regulate itself than Moses thought (Gratz, 2010, pp. 21- 23). Although Jane Jacobs never spoke of shared spaces, it is easy to see the
similarities between the notion that the city that can organically self-regulate, and the concept of shared space.
The first real notions of shared space like concepts where In the Buchanan (1963) rapport. Although it still never actually mentioned the concept of the shared space, the idea that traffic would be able to self-regulate was mentioned. This report stated that when traffic volume increased the different types of traffic had to be separated. This way of thinking has been very influential. It can be seen in the form of bicycle tunnels, barriers, traffic lights, and all other well-known traffic regulation methods. However, Buchanan (1963) believed that in some cases a mix where pedestrians and some forms of motorized vehicles would share a space could be viable.
The Buchanan rapport is sometimes seen as the precursor of the ‘’woonerf’’(which roughly translates to home zone) concept, which is in turn seen as an early form of a shared space
(Karndacharuk, Douglan & Dunn, 2014, p. 191). The ‘’woonerf’’ concept was first introduced in the sixties by urbanist Joost Vahl. it is an integration of traffic and residential activities. The pedestrian gets priority over motorized transportation, which has to adapt to the speed of pedestrians. There are no traffic regulations signs, no sidewalks or boundaries (Hammilton-Baillie, 2008, p. 163). According to Karndacharuk et al. (2014) this was the first application of a shared space like concept.
Hans Modderman, a city planner for the province of Groningen, started the first true shared space experiments in Drachten and Oostellingenwerf and is seen as the ‘’inventor’’ of the concept. He believed that urban planners in the Netherlands were ‘’over’’ regulating. And that when given the right circumstances people could often regulate themselves much more effectively. He believed that the Shared space experiment (Although he never named it as such) was particularly effective for slow traffic
like pedestrians, bikes and scooters. These two experiments all became permanent as it was believed that they allowed traffic to move through that space faster and more efficient, while not increasing the amount of accidents by a noticeable amount. The shared space was originally developed with three main goals in mind. While some goals of specific shared spaces may vary according to the needs of that space, they all share these three larger goals. The first shared goal is giving back responsibility for traffic
behavior to the users of the space. Which is thought to lead to the second goal of improving livability . And the last goal is combatting traffic cluttering by making traffic move through that space more effectively (De Haan, 2012, p. 2).
However it was not until much later that British urban planner Ben Hamilton Ballie coined the term ‘’Shared space’’ that the experiments got the name it is known by today (De Haan, 2012, p. 4). The shared space in Drachten was researched by the aforementioned Hamilton Ballie. He concluded that the introduction of the Shared space there made the traffic in the area more fluent and less congested. He credited this to respondents using their anticipatory and communication skills much more than they normally would on more regulated spaces. Potential conflicts are mitigated by falling back to informal protocols and interactions which are stimulated by these new spatial circumstances (Hamilton-Baillie, 2008, p. 171) .
Nowadays the concept of shared space in gaining in popularity. As mentioned earlier it has been introduced inside of the Netherlands in the province of Groningen (Drachten) and later in Frysia (Van Veen & Nota, 2011), but also outside of the Netherlands. With examples in New Zealand (Karndacharuk et al., 2014) and Great Britain (Moody & Melia, 2014). Especially in Great Britain the idea is very popular with the in 2012 finished Exhibition road in London being one of the biggest and most expensive Shared Spaces to date (De Haan, 2012 p.4). Yet until now the usage of the shared space in the Netherlands seemed to be limited to the northern provinces of the Netherlands. This has off course changed with the introduction of the Shared space behind the Amsterdam central station.
The implementation of the Shared space in Amsterdam is connected to a number of broader policy implementations. Firstly the city of Amsterdam wants to encourage people to use either their bike or public transport and not their cars. In the case of the shared space this was done by removing the car road. Secondly there is a trend in the policy plans of the municipality to create more efficient and safe traffic by influencing traffic behavior (Gemeente Amsterdam, 2013, p. 17). This seems in line with what Hamilton Ballie (2008) described about the Shared space and the way it affects traffic behavior. All these policy considerations were important contributors for the decision to implement the Shared space. But what also seemed like an important reason was the suitability of the area because of the relatively slow
traffic behind the central station. Which ,as mentioned earlier, is the type of traffic that is also seen as most effective for a successful Shared space by the municipality (Litjens, 2015).
In the policy letter to the Municipality by Alderman Litjens (2015) the goals of the Shared space were described as being threefold. Firstly the area was to become a shared area. Pedestrians, cyclists and slower motorized vehicles like scooters and mopeds must effectively and safely share that space. This presumably to enhance the flow of traffic, although this was not specifically mentioned as such. Secondly the speed of cyclists, scooters and mopeds must be slowed down in order to improve safety. And lastly the focus must be changed to all users of that space, not just the cyclists and scooters. How and why this must be done remains unclear in the Alderman’s note. The broader line at least seems to be enhancing safety by reducing speed and making people more observant/communicative. This in line with what Modderman (de Haan, 2012, p.4) and Hammilton Ballie (2008) say about the Shared space. Yet there is one minor difference. They put more emphasis on giving back responsibility to people that use that space and thereby increasing livability (De Haan, 2012, p. 5). Unlike the alderman who states that increasing safety is also an important goal (Litjes, 2015). This however makes sense because the interests of the Alderman are more inclined to direct and provable results. And it is easier to measure accidents (or the lack of) than increased user responsibility or livability.
2.2 Interactions in public space
One of the main theories the shared space was based upon is the ‘risk-homeostasis theory’ . This theory states that people in relatively ‘safe’ situations display more risk taking behavior, while people in
relatively unsafe situation show less risk taking behavior. This is because people constantly search for the right mixture of safety and behavior. By removing visual traffic regulations and boundaries people would feel less safe and would therefore take less risks (Wilde, 1982, p. 210-215). By removing these
regulations and boundaries people are forced to regulate their way through these squares. This self-regulating would lead to more observant and communicative behavior and thus one would expect more interactions on these shared spaces.
From a sociological perspective it is this hypothetical increase in interactions in the shared space that is most interesting. Is there truly an increase in interactions? If so how does this manifest itself and how is this different from a normal space? To research this it is important to delve a deeper in theory about interactions in public space.
Erwin Goffman(1966) wrote extensively on these interactions in public space in his book that is adequately called: behavior in public space . He broadly distinguishes two types of interactions. Focused
interactions and unfocussed interactions and coined the term civil inattention. While his work was later expanded upon by Lyn Lofland (1996) in her book called: the public realm. Where she further described how people transgress space without (usually) crashing in to each other. She describing this by inventing the concept of cooperative motility. Furthermore she wrote extensively on how the build environment affects interactions in public space.
A unfocused interaction is an interaction that occurs with people who are in the same space. They do not engage in lengthy face to face conversation, notice each other or look at each other for lengthy periods of time. They only notice the other in quick moments that are only enough to make the other known he has been noted, but nothing more. These can range from anything like a quick stare, a look around or stepping aside for someone to let him pass. The most important part is that the moment is quick and the focus is thereafter immediately removed (Goffman, 1966, pp. 33- 63). Goffman (1966) also calls these moment civil inattention. He describes it as the way in which people make bodily co presence possible without true interaction. Lofland (1998) describes it as courtesy without conversation. The other must not feel like a target of special curiosity or design. If that were to happen Goffman(1966) points out that there is no longer an unfocused interaction, but a focused interaction.
A focused interaction, or face engagement, is different from the unfocused interaction because participants join each other openly in maintaining a focused visual attention to a social situation. These focused interactions are not by definition verbal and can be non-verbal too. Focused interactions have to be more than just a short acknowledgement of co-presence. They can be in the form of spoken words, gestures or moves (Goffman, 1966, pp. 63-124). A conversation between a tourist and a local is obviously a focused interaction. But there are other moments which also qualify. Think of a tram conductor giving a passenger a ticket or two cyclists that look at each other for a longer period to see who takes priority . Focused interactions are interactions where participants do more than just quickly acknowledge each other’s presence. The types of focused interactions one would expect in an area like the shared space are people that strike up a conversation to successfully transgress through traffic. But also an angry cyclist that makes a hand gesture to a tourists who returns the hand gesture, or a verbal fight that breaks out between cyclists that clash in to each other. In short, the types of interactions may vary wildly, but they all have in common that there is a focused visual attention to each other and the social situation (Goffman, 1966, pp. 89-91).
As mentioned earlier there is another sociologist that wrote extensively on behavior in public space: Lyn Loffland (1998). One aspect of her book that is particular interesting in regard to the shared space is cooperative motility. This stands for the innate ability of individuals to transgress public space by
cooperating their movement and thereby avoiding things like collisions (Lofland, 1998, p. 92). Within the shared space there is more freedom for participants to choose their own way of transgressing space. There is thus more room for cooperative motility. Lofland (1998) sees cooperative mobility as an almost choreographed dance. She described that even though animate objects (or people) and their animatedly propelled inanimate objects (in our case bicycles, scooters and mopeds) are harder to read, people very rarely collide into each other and traffic is usually very uneventful (Lofland, 1998, p.92). Lofland (1998) points out this is because humans cooperate, and so interact, to make it uneventful. she feels that civil inattention ensures that people do not become too focused in the doings with strangers that they lose the needed focus for cooperative motility. People have to find the right balance between noticing others and letting them know they are noticed, without losing sight of the intricate way of moving through public space (Lofland, 1998, p. 92).
Furthermore she wrote extensively on the built environment and how it effects interactions. In her own words:
‘’The built environment certainly does not determine exactly how people are going to interact with one another, but it does amplify or constrain the range of interaction possibilities’’ (Lofland, 1998, p. 181)
She uses the example of large Le Corbusier inspired high rise housing projects to illustrate this. Environment where people, that live together, meet each other so rarely that longer periods of interaction are almost ruled out. Yet she describes this as a loss, because space can be a place that structures how communications will occur, who will communicate and what the content of those communications are. While pre industrial old market squares where usually places everybody would meet architecture of the 19th and 20th century often tried to separate classes. Lofland (1998) uses the example of Trafalgar square, that was intended to separate upper classes from the working class. While the area before the shared space was obviously not meant to separate the rich from the poor, it was meant to separate the pedestrian from the cyclists, the scooter from the car or the moped from the Canta. Regulatory control in cities was and is everywhere (Lofland, 1998, pp. 185-188). The shared space is clearly a breakaway of this regulatory control. A experiment in releasing some of this control and returning it to the public. The question is, does this indeed restructure the manner in which people communicate and if so in what way?
2.3 production of space
linear Idea of how space should be planned. Public space planned in such a way that it accounts for every possibility by regulating wherever possible. In the shared space planners give people more freedom to use it in their own particular way. French urban philosopher Henry Lefebvre pointed out that public space is formed partly by how it is planned, but also by how it is used and perceived. Lefebvre sees space as an interlinkage between the manner in which it is planned, perceived and lived. They form a triad that produces space. Space is, according to Lefebvre, not just a linear result of planning and designing, but as much a result of how that space is lived, and the way in which space it is perceived by its users (Lefbvre, 1974/1992, pp. 73-88)
Planned space is the most noticeable and easily understood of the three categories. It is the space as envisioned by its planners, which are usually architects and government bureaucrats. In the example of the shared space planners envisioned that traffic would self-regulate if boundaries and visual regulations were removed. So although the shared space might at first look like negligent planning, there is a clear idea and plan behind it. Planned space does not come about in a natural way, but is in a sense imposed upon its users. It encompasses the way both the physical features of that space are planned and how the rules (or regulations) that are applied to that space are implemented (Chevalier, 2015, p. 125-126). One of Lefebvre’s main points is that in some cases planned space does not match how space is actually used. For example architects can plan a square, or apartment complex with the intent of stimulating community, yet can sometimes fail in this attempt (Lefbvre, 1974/1992, p. 292) Take the example of the Heesterveldbuurt in Amsterdam Zuid Oost. A number of apartment blocks built in the 1980’s as a reaction to the high rises that are common in that area. It consisted of low-rise apartment blocks on all sides and a large communal space in between. This space in between was supposed to stimulate interactions between residents, yet it ended up being a hotbed for crime and drug users (de Jong, 2013) . This discrepancy between how space as planned and how it ended up being used(or lived) by its everyday users brings us to the second dimension of Lefebvre’s triad, lived space.
Lived space refers to the routine of everyday life. Within the shared space this could be the habit of people to dismount their bikes, people displaying more social behavior, an increase in collisions or people that use objects as benches that were not designed as such. As mentioned earlier this can be markedly different than what planners had envisioned (Lefbvre, 1974/1992, p. 292). In the earlier used example of the Heesterveld neighborhood the lived space might entail youths that use the communal space to use soft drugs. Another example might be cyclists in Amsterdam. The city is especially renowned for its cyclist that do not adhere by the regulations that are imposed from above. Anybody that has cycled in Amsterdam has stories about people going through red lights(and has very probably
done so themselves). In this regard one might almost speculate that the Shared space was implemented because cyclist wont adhere by regulations anyway (although this is merely speculation as it is not mentioned in any of the policy documents). It is however interesting to see whether people start behaving differently when this important faced, namely visible regulations, is taken away. How people shape space by their experiences and the habits that people form when they encounter a specific space are central to this dimension. This dimension is often described as a more passive part and not wholly consciously experienced. Not many people usually think of why they sit on a sculpture on a square or cycle through a red light, they just do (Chevalier, 2015, pp. 126-127). In the example of the Shared space it is interesting to compare this unconventional way of urban planning to conventional spaces and see how it affects the relationship between planned and lived space. To see if the plan for that space matches the way in which it is actually used
The last and (in my opinion) most elusive dimension of the production of space is perceived space. This is the symbolic and mental part of space. It is the way in which people mentally experience space. This is not to be confused with the more practical unconscious experience people have in the lived space dimension. The perceived space is the more overarching and conscious perception of that space. It does not necessarily coincides with the realities in that space (Although it usually does) (Chevalier, 2015, pp. 127-129). For example people might come to see this new shared space as very unsafe, even if statistically speaking, it is not. If we return to the example of the Heesterveld
neighborhood there was a general sentiment by many of the inhabitants and the rest of the city that that area was very unsafe and crime infested. While nowadays it is full of artists and students and has a more positive perception for a lot of people (De Jong, 2013) . This change in perception does not necessarily mean that the way in which space is actually used has also changed (although in many cases, and this one, it usually is). Yet it is an important contributor to the whole production of space because the perception of space can ultimately change plans (planned space) or habits (lived space). In this
triad and not a linear model. All three categories influence each other (Lefebvre, 1974/1992).
According to Lefebvre (1974/1992) the production of space is happing continuously. The cyclists that dismounts his bike because he perceives space as unsafe. The bureaucrat that changes regulations in a space because of a negative perception. Or the pedestrian that can use space more effectively because of a new set of regulations implemented in that space. Yet the production of space is not always in harmony between planners and users. Public space can be planned to accommodate families, yet when they are mainly used by soft drugs using youths there is a discrepancy between planned and lived space. In the case of the shared space the planners removed regulation because they theorized that users could self-regulate traffic. If for example the shared space would result in more collisions and a negative perception about that space for users there would also be a discrepancy between how that space was planned and the manner in which it is lived and perceived by its users.
Based on the policy notes, theories on behavior in public space and the theory the production of space three hypotheses have been constructed. The first hypotheses is aimed at testing if people do indeed interact more to transgress the shared space safely. The second hypotheses is aimed at
discovering if (near) accidents are not more frequent in the Shared space and so is aimed at measuring safety. And the final third hypotheses is aimed at testing how this changes their experience of this space. Hypotheses 1:
H0: people on the Shared space do not interact significantly more in both focused and unfocused interactions than people on either the Haarlemmerstraat or the Muntplein square.
H1: People on the Shared space interact significantly more in both focused and unfocused interactions than people on the Haarlemmerstraat or the Muntplein square.
Hypothesis 2:
H0: The shared space has no significant difference in (near)collisions compared with other similar regulated spaces.
H1: The shared space has significantly more (near)collisions compared with other similar regulated spaces.
Hypotheses 3:
therefore become more observant in order to safely transgress(lived space) space.
3. Methodology
3.1 Outline of research
To find out how the shared space differs from other similar conventionally regulated spaces this research conducted structured observations that looked for statistical differences, qualitative interviews to compare in what way peoples experiences may differ between these spaces and qualitative observations that registered other important information not measurable in numbers or interviews.
Next to the shared space behind the central station there where observation and interviews on the crossroads on the Haarlemerstraat and on the Muntplein square. The crossroads on Haarlemerstraat was chosen because it is conventionally regulated. As can be seen on the image below it has a sign for cars which signals their speed limit, shark teeth to indicate who has priority, and a cycling line that indicated where the cyclist should stay. It however lacks traffic lights and zebra crossings. It can therefore be seen as a moderately regulated space. The other space where the observations were conducted was the Muntplein area. This is also a conventionally regulated space, but it is more heavily regulated than the crossroads on the Haarlemmerstraat. As can be seen on the image below it has zebra crossings, traffic lights, boundaries that separate motorized vehicles from bicycles and like the
Haarlemmerstraat shark teeth. Muntplein is clearly the most regulated of the three spaces. Of all tree spaces only the Haarlemmerstraat has car traffic(however not a great deal of). The observations there will only be conducted when there is no car present so the influence of a car can be ruled out.
(Haarlemmerstraat, Google maps)
(Muntplein, Google maps)
The observation where be done from a distance. And so where not participant observations. This simply because it is hard to observe traffic while partaking in it. Because the data was be used to look for statistical differences a highly structured observation schema was used. See appendix 1 for the schema. The number of observed respondents was 300 per square.
The methodology of the observation’s that will be used in this research will have many
similarities with the famous urban research ‘’the social life of small urban spaces’’ by William H. Whyte (1980). He observed many squares in New York and would quantify large numbers in charts. In these charts he would specify how many people would use a specific area and how they used it. He would then use these numbers to explain why some squares where more popular as meeting places than others. Why for example some spaces were used frequently to sit in by some groups of people and others were not. He meticulously described every detail in these squares like for example number of benches, of trees and fountains, and their effect on the social life on these squares. Whyte (1980) had many students at his disposal and much more time. So while the scale and numbers of his research are out of reach, his style of methodology is not and will be therefore be an example this thesis will be based upon.
The second part of this thesis research entailed qualitative interviews and observations. These interviews where semi structured and were conducted in the same areas as the observations. The
interviews were all conducted with people on location. five interviews were conducted per location. The qualitative observations were done as an addition to the structured observations. Keeping a notebook close at hand to write down anything of interest that was not covered by the structured observations.
3.2 Strength and weaknesses of methods
The design of this research is, as mentioned earlier, a mixed method design. The strength of the mixed method design lies in the fact that is measures both quantitatively and qualitatively. By doing so it can avoid the some of the inherent weaknesses of these methods on their own. Quantitative research is often weak in understanding anything outside the parameters it measures. And it is often is weak in explaining the context in which people behave. Yet qualitative research can lead to biased
interpretations because the sample size Is much smaller which makes it hard to generalize findings to the larger group the researcher intends to research.
However there is one obvious weakness of combining the two methods and that is time. Because of time limitations using tree methods will mean that they cannot be conducted as comprehensively as would have been possible with a quantitative or qualitative only research. This means that the sample size of both methods is lower than a research method that just uses one. However in this case it was deemed better to approach the problem from both sides because of the nature of the research question. Which looks for numerical differences, but also for respondent experiences.
4. Operationalization of key concepts
4.1 Structured observations
The observations were focused on interactions between the people that partook in traffic on the selected squares. The manner in which these interactions were structured in the observation schema was based on the earlier described theories on interactions in public space. Focused and unfocused interactions were therefore important categories.
To construct an accurate schema for the observations a preliminary observation has been conducted. Tree hours from 12:00 to 15:00 on a Saturday were spend on the shared space observing traffic. Most interactions on the shared space were short and fast, as one would expect in traffic. Bell ringing was observed as where quick hand gestures. Most hand gestures were used to give someone priority or to thank someone for the priority that was given. But one hand gesture was observed performed by a pedestrian against a scooter who had cut him off. Although this gesture was probably not noticed by the scooter as he sped away too soon. Almost all the longer interactions where all
observed after (near) confrontations. In the three hours there were four near confrontations and three actual confrontations(or small collisions). All of these confrontation where relatively benign and nobody got hurt, and all of these confrontations led to longer interactions. one collision was with a cyclist and two (what were presumed to be) tourists. The cyclist yelled something and the tourists made what looked like apologetic hand gestures. Another collision was between two cyclists, one who was getting on the ferry the other of it. They seemed to note each other, both seemed to apologize and walked on. The last confrontation was between a woman on a bike and a man on a scooter. The man sped past her and the woman, who looked startled, dismounted her bike. She starting talking with another cyclist and afterwards continued walking through the shared space with her bike in her hands and only mounted her bike when she was on the cycling track.
Based on this short preliminary observation it was expected that both unfocused interactions and focused interactions would be present, but unfocused interactions much more so. Civil inattention in the form of fleeting small moments of fast interactions are expected to be most common. Quick looks of acknowledgement, hand gestures, the ringing of a bel, the sound of a scooter claxon or some other quick but unfocused form of interaction. These moments of unfocused interaction were expected to
sometimes lead to focused interactions. These can be when a traffic confrontation are avoided, or have already happened. These were expected to consist of long moments of eye contact, conversations, or hand gestures where both side use multiple gestures and focus their attention on each other through eye contact. Off course these focused interactions do not always have to be the product of a
confrontation, and can also be someone helping a tourists, or two friends meeting during biking. The main point is that there is a longer visual attention on each other and the situation and some form of reciprocal communication is used. The observation schema thus registered if there was a unfocused, or focused interaction, and what type of unfocused or focused interaction it was. next to this the
observation schema also registered the gender, age group, time of observation and vehicle type of the respondent. Lastly, the schema registered weather there was a near collision, or collision.
Sex was registered by either a 0 for male or a 1 for female. The mode of transport was registered as either P for pedestrian, C for cyclist or S for scooter or moped. The age group was either 0 for under 30, 1 for between 30 and 50 and 2 for above 50. Time of day will be either 0 for morning rush hour, 1 for late morning, 2 for midday, 3 for afternoon rush hour and 4 for evening. The categories unfocused and focused interactions were simply registered either by a 0 if present and a 1 if not present. Lastly the categories near collision and collision were simply have a cross when they are observed. The observation schema can be found in appendix 1.
Location of observations
4.2 Interviews
The interviews were conducted after the observations and are meant to measure whether the results from the observations matched the way that space is lived and perceived by the people that use it. The focus of the interviews is in answering the second part of the main research question. How the
experience people have differs from the conventionally regulated spaces. The topic list is influenced by Lefebvre’s spatial triad. The way people experience (or live and perceive) this unconventionally planned way of public space is central in this. The topic list can be found in Appendix 2.
5. Data
5.1 Quantitative
300 observations were made on the three distinct locations and so in the end 900 respondents were observed. Each respondent was observed for eight dependent variables and four independent variables. See appendix one for the observation schema that was used to conduct the observations.
5.1.2 Dependent variables
The dependent variables is the amount of interactions and collisions that where observed. The
interactions where put in two groups, unfocused interactions and focused interactions. Within these two groups further categories were added as can be seen in appendix one. These categories were quick glances, small hand gestured and bell (or claxon) ringing. For focused interactions the categories where, conversations, reciprocal hand gestures and longer period of eye contact. Both groups had another option for interactions that did not fit any of the prescribed categories. All these categories would be registered with either a 0 if the interaction happened and a 1 if it did not. Making all dependent variables in this research binary variables. Near collisions and collisions were also dependent variables register with either a 0 for present or 1 for not present.
5.1.3 Independent variables
5.1.3.1 Locationthe variable location registered on what location the
respondent was observed. This was either the shared space, which was marked by an C, the Haarlemmerstraat, marked by a H, or the Muntplein, marked by a M. Making this variable a nominal variable with 3 distinct values. Each location had 300
observations each as can be seen in the pie chart to the right. 5.1.3.2 Time
The next independent variable was the moment the observation took place. These were morning rush hour, late morning, midday, late rush hour and evening. 187 observations were done during morning rush hour, 143 during late morning, 210 in midday, 255 during afternoon rush hour and 105 in the evening. Like the variable for location this is a nominal variable. The number of observation per square can be found in the table below.
Time Location
Morning rush hour
Late morning Midday Late rush
hour Evening Shared space 59 31 60 105 45 Haarlemmerstraat 68 82 90 30 30 Muntplein 60 30 60 120 30 Figure 2 5.1.3.3 Sex
The sex of the respondents was also registered. the observations ended up with an unequal amount of males and females. There were 509 males observed and 391 females. This varibale registred a f for female and a m for male making this varibale a binary varibale. In the table below the amount of females and males opbserved per location can be seen.
Location Sex
Shared Space Haarlemmerstraat Muntplein
Female 133 127 127
Male 169 167 173
Figure 3
5.1.3.4 Age group
The age of the respondents was registered as either young, middle ages, or older. With younger being everybody until 30, middle aged anybody above 30 but under 50, and older everybody above 50. It was registered as either 0 for young, 1 for middle aged, or 2 for old. the amount of younger people observed was much higher than the amount of middle aged and older people observed. With 597 young people, 150 middle aged and 153 older people. Like the variable for time this is an ordinal variable. The observed ages of the respondent per location can be found in the table below.
Location Agegroup
Shared Space Haarlemmerstraat Muntplein
Young people 195 202 200
Middle aged 57 46 47
Older people 48 53 53
Figure 4
5.1.3.5 Type of transportation
The last independent variable that was registered was the type of transportation the respondent had. This was either a b for bicycle, a p for pedestrians and a s for scooters or mopeds. Like most independent variables there was a unequally large amount of one variable. As expected in the city of Amsterdam those where bicycles. There were 673 cyclists observed, 116 pedestrians and 111 scooters. This is also an nominal variable. The numbers per location can be found in the table below.
Location Agegroup
Shared Space Haarlemmerstraat Muntplein
Cyclist 205 244 224
Pedestrians 46 25 45
Scooters 49 31 31
Figure 5
5.1.4 Analytical strategy
Because this research want to compare the interaction variables separately on one independent variable with more than two levels(or groups) there are tree obvious candidates. The one way Anova test, the Kruskal Wallis and the Chi-squared test. All tests can compare three or more groups. So in that sense they are all adequate. However the Anova test is a parametric test that requires that the dependent variable is at least interval. These requirements are not met by the either variables used in this research. The Kruskall Wallis test is a non-parametric test and thus seems like a good alternative, however it requires the dependent variable to be at least ordinal. Which again the dependent variables are not. They are all nominal variables that only registered the presence or absence of an interaction. This leaves the Chi-squared test as most suitable method of testing if the amount of interactions and (near)collisions is dependent on the location (de Vocht, 2015, pp.119-157).
The chi square takes all the values for the observations and calculates the expected frequencies if there was no difference between the prevalence of interaction between the tree locations. It then puts
these against the actual observation numbers and indicated if the difference is significant or not.
However on its own the chi square test only tells us that one of the groups has a significant effect on the dependent variable, it does not specify which one. This than can be discovered by using the adjusted standardized residuals to get the z scores of the groups in between. Which than can be used to find the significance p values between the locations (de Vocht, 2015, pp.119-157).
5.2 qualitative
In the end fifteen different people were interviewed on the three different locations. The interviews were coded using Atlas.ti. five codes were created based on the topic list that was used during these interviews. These topics were in turn based on the last sub question which was concerned with the experience people had on the shared space and how this compared to the other three locations. These five codes were: Opinion of lay out of that space, how Safe respondents felt that space was, changes respondent would make to that space, how respondents used that space and if respondents thought that space functioned as it was meant by its planners.
Furthermore next to the quantitative structured observations a notebook was kept were more qualitative observations were written down. These were meant for observations that could not be added in the schema. This entailed information like where in that space people would wait, what effect high traffic had on the way that space was used, if some form of organization happened that was not in the schema or people dad did something unexpected. In short it was meant for observations that stood out, yet could not be taken into account by the schema.
6. Results
6.1 Quantitative results
6.1.1 Interactions and locations
The results show that there are no significant differences between the amount of unfocused interactions on all three locations when not controlling for the different time frames. As shown in figure seven there is no significant difference between the prevalence of any of the categories that represent unfocused interactions .
The Chi-squared test shows that the prevalence of quick glances is not significantly different per location. As seen in figure six the prevalence is reasonably close to being perfectly equal on all squares. Unsurprisingly the p value is 0.539. Which makes it safe to assume that there is now difference between the three locations. They have almost the same number of observations on all tree location. Which
means that about half of every respondent had a quick view around him or herself to see who or what was around them. This is in line with the scores in the table being all close to 150(or half). Because of this the p values are, unsurprisingly, high for all locations. The difference between the Haarlemmerstraat and the Shared space is not significant with a p value of 0,845. It is the same for the difference between Muntplein and the Shared space and Muntplein and Haarlemmerstraat with p values of 0,350 and 0,684 respectively.
The second category that represented unfocused interactions was small hand gestures. Here there is also no significant difference. With a p value of 0.06. This does however mean that the
differences between means is only barely not significant. In figure six it becomes clear that the observed small hand gestured of the shared space and muntplein are both very close to each other with 39 and 36 small hand gestures respectively. Because these scores are so similar heir between groups p value is unsurprisingly high with a score of 0.93. This means that around 12% of the respondents at both these locations used a small hand gesture to communicate. At the Haarlemmerstraat there were 55 small hand gestured observed. Which means that respondent here use small hand gestured around 20% of the time. As these scores are more different the p values between the Haarlemmerstraat and the other two location are much smaller. However this differences is still not statistically significant and so one may not conclude that there is a difference between the tree locations and the amount of small hand gestured used by the people on these locations.
The final category that represented unfocused interactions was the ringing of a bel(or scooter claxon). Here the same image arises as in the two previous categories with a p value of 0.378. There seems to be no significant difference between the prevalence of ringing a bel or claxon per location. The observed scores are all very low between either 9 and 4. Which means that only 0.3 % of the time people ringed their bell. And because these means are so similar the p values between locations are high. With the p value between Haarlemmerstraat and the Shared Space 0.84, between Muntplein and the Shared space 0.35 and Muntplein and Haarlemmerstraat 0.68. All clearly not significant.
Location
Unfocused interaction
Shared Space Haarlemmerstraat Muntplein
Quick glance 161 (153) 151 (153) 148 (153)
No quick glance 139 (146) 149 (146) 152 (146)
Small hand gesture 39 (43) 55 (43) 36 (43)
Ringing of bell 9 (7) 7 (7) 4 (7)
No bell ringing 291 (293) 293 (293) 296 (293)
Figure 6 osberved scores of interactions and expected chi2 score between brackets if the prevelance of interactions where perfectly independent of location
Significance levels between groups and overall
* p < 0.05, ** p < 0.01, *** p < 0.001
Significance levels
Between Shared
Space &
Haarlemmerstraat
Between Shared
Space &
Muntplein
Between
Muntplein &
Haarlemmerstraat
Overall
Significance
Quick glance
0.8
0.35
0.68
0.53(not sig)
Small hand gesture
0.15
0.93
0.07
0.06 (not sig)
Bell ringing
0.84
0.35
0.68
0.37 (not sig)
Observations
Total
observations: 900
When the all the categories that represented focused interactions are compared the results are very different. Firstly, the prevalence of conversation between respondents on the locations is
significantly different, with a p value of 0.017. This is reflected in the scores in figure eight were eight conversation happened on the Shared Space, four on the Haarlemmerstraat and zero on the Muntplein. Looking at these scores it becomes clear that at none of the location a lot of conversations took place in traffic. This is hardly surprising as one would expect most people to just move through that space without having a conversation with a stranger. This is clearest on the Muntplein where no single
conversation between respondents that did not know each other took place. On the Shared Space three percent of the respondents had a conversation and at the Haarlemmerstraat one percent. Not
surprisingly this means that only the difference between Muntplein and the Shared Space is significant. With a p value of 0.012. While the other two p values are both around 0.3 and clearly not significant.
Another category that represented focused interactions was longer period of eye contact. The scores of this category also seem to differ significantly as there is a p value of 0.000. The difference seems to be largely between the Shared space and the other two locations. With a observed score of nineteen for the Shared space and six and tree for the other two. This difference in scores would suggest that the only significant p value are the ones between the two other locations and the Shared space. The p values show that this is indeed the case. With a p value difference between the Haarlemmerstraat and the Shared space of 0.001 and Muntplein and the Shared space of 0.000. While Muntplein and the Haarlemmerstraat do not differ significantly with a p value of 0.86.
The last category that represented focused interactions was reciprocal hand gestures. Here too the scores differ significantly with a p value of 0.000. Again the main difference in means seems to be between the Shared space and the other two spaces. With the observed scores at the Shared space being eighteen and the other two scores four and two respectively. As expected by these means the only significant difference is the p value between the Shared space and Haarlemmerstraat with a p value of 0.001 and the Shared space and Muntplein with a p value of 0.000. While the p value between
Muntplein and Haarlemmerstraat is nog significant at 0.86.
Location
Focused interaction
Shared Space Haarlemmerstraat Muntplein
Conversation 8 (4) 4 (4) 0 (4)
No conversation 292 (296) 296 (296) 300 (296)
No eyecontact 281 (290) 294 (290) 297 (290) Reciprocal hand gestures 18 (8) 4 (8) 2 (8) No reciprocal hand gestures 282 (292) 296 (292) 298 (292)
Figure 8 osberved scores of interactions and expected chi2 score between brackets if the prevelance of interactions where perfectly independent of location
Significance levels between groups and overall
*
p < 0.05, ** p < 0.01, *** p < 0.001
Significance levels
Between Shared
Space &
Haarlemmerstraat
Between Shared
Space &
Muntplein
Between
Muntplein &
Haarlemmerstraat
Overall
Significance
Conversation
0.3
0.012*
0.3
0.017(*sig)
Longer period of eye
contact
0.01**
0.00***
0.86
0.00 (***sig)
Reciprocal hand
gestures
0.01**
0.00***
0.86
0.00 (***sig)
Observations
Total
observations: 900
Figure 96.1.2 collisions and near collisions
After conducting another Chi-squared test it seemed that the variable near collision is not independent of the locations with a p value of 0.001. as can be seen in figure eleven he differences are most
significant between Muntplein and the Shared space with a p value of 0.001. But there is also a significant difference between Muntplein and Haarlemmerstraat with a p value of 0.03. The scores as seen in figure ten suggest that near collision happened about 05 percent of the time (or fourteen times at the shared space and ten times at the Haarlemmerstraat) at both the Shared space and the
Haarlemmerstraat. Yet did not happen once at the Muntplein area. This makes the p values in figure eleven of 0.56 between the Shared space and the Haarlemmerstraat logical as their means do not really differ.
The amount of registered actual collision is very low. As seen in figure ten only two observed collisions on the Shared space, two on the Haarlemmerstraat and none on the Muntplein. This means naturally that there is no significant difference between the three locations as they all have very few actual collisions. As can be seen in figure eleven the p values is not significant between any of the tree locations.
Location (near) collisions
Shared Space Haarlemmerstraat Muntplein
Near collision 14 (8) 10 (8) 0 (8)
No near collision 286 (292) 290 (292) 300 (292)
Collision 2 (1) 2 (1) 0 (1)
No collision 298 (299) 298 (299) 300 (299)
Figure 10 osberved scores of interactions and expected chi2 score between brackets if the prevelance of interactions where perfectly independent of location
Significance levels between groups and overall
* p < 0.05, ** p < 0.01, *** p < 0.001Between Shared
space &
Haarlemmerstraat
Between Shared
space &
Muntplein
Between
Muntplein &
Haarlemmerstraat
Significance
Near collision
0.5
0.00 ***
0.03 **
0.001(**sig)
Collision
1
0.43
0.43
0.36 (not sig)
Observations
Total
observations: 900
6.1.3 Sex, age and mode of transportation
Sex does not seem to differ over any of the interaction variables. Neither unfocused nor focused interactions seem to differ between males and females. All the chi-square test scores are not
significantly different from what they would be if there was no difference between men and women. If the same tests are done while controlling for the locations the results do not change and remain not significant. We can thus conclude that sex is independent of the prevalence of interaction on all three locations and on the Shared Space. Or in other words, sex does not influence the amount of interactions people use. It is worth noting that the scores look rather different, but this is because the observations ended up with a larger amount of men (see chapter 5.3.3). The chi-2 test holds this in consideration as can be seen by the expected scores between brackets in figure twelve.
Unfocused interactions Focused interactions
Type of interaction Observed quick glance Observed small hand gesture Observed ringing of bell Observed conversation Observed reciprocal hand gestures Observed longer period of eye contact Women 212 (200) 58 (57) 11 (9) 2 (5) 6 (10) 10 (12) Men 248 (260) 72 (73) 9 (11) 10 (7) 18 (14) 18 (16)
p value 0.10(not sig) 0.7(not sig) 0.29(not sig) 0.06(not sig) 0.4(not sig) 0.06(not sig)
* p < 0.05, ** p < 0.01, *** p < 0.001
Figure 12 osberved scores of interactions and expected chi2 score between brackets if the prevelance of interactions where perfectly independent of sex
Different types of transportation do make a significant differences in prevalence of interactions. Firstly when the whole dataset is used it becomes clear that most interaction variables are dependent on the type of transportation. As can be seen in figure fourteen only the prevalence of reciprocal hand gestures and longer period of eye contact were statistically independent of the type of transportation. The differences between expected scores and observed scores leads to the expectation that scooters and pedestrians use significantly less interactions than cyclists. The variable bell ringing was not incorporated for pedestrians cannot do this.
Focused interactions Unfocused interactions Type of interaction Observed quick glance Observed small hand gesture Observed ringing of bell Observed conversation Observed reciprocal hand gestures Observed longer period of eye contact Cyclists 355 (344) 118 (97) X 5 (9) 17 (18) 19(20) Pedestrians 67 (60) 7 (17) X 5 (1.5) 6 (3) 7 (4) Scooters 38 (57) 5 (16) X 2 (1.5) 1 (3) 2 (3.5)
Figure 13 osberved scores of interactions and expected chi2 score between brackets if the prevelance of interactions where perfectly independent of type of transportation
The scores suggest that in general cyclist use much more interactions than scooters and pedestrians. The p values between the groups confirm this. As can be seen in figure fourteen the amount of quick glances are significantly less common with scooters and pedestrians. As are small hand gestures. While at the categories for focused interactions these changes are not as large. With only pedestrians having significantly less conversations than cyclists. As the observed values versus the expected values suggest scooters and to a lesser extend pedestrians seem to use significantly less interactions than people on bikes. If only the observations for the Shared Space are tested the results remain relatively similar. As can be seen in figure fourteen scooters still use unfocused interactions significantly less. With one minor difference. Pedestrians used significantly less small hand gestures and conversations when all three datasets are used, yet on the Shared Space they use them about the same as the other two
Between cyclists
and pedestrians
Between cyclists
and scooters
Between scooters
and pedestrians
Overall
Significance
Quick glance
0.5
(0.5
0.00***
(0.00***)
0.00***
(0.00***)
0.00***
(0.00***)
Small hand gesture
0.00***
(0.12)
0.00
(0.01**)
0.94
(0.78)
0.00***
(0.00***)
conversations
0.00***
(0.19)
0.63
(0.99)
0.22
(0.36)
0.00***
(0.2)
Reciprocal hand
gestures
Longer period of eye
contact
0.23
(0.88)
0.15
(0.56)
0.58
(0.16)
0.83
(0.18)
0.11
(0.17)
0.15
(0.07)
0.12
(0.13)
0.12
(0.08)
Observations
Total
observations: 900
Figure 14 P v* p < 0.05, ** p < 0.01, *** p < 0.001Lastly the age of the respondents seem to affect the prevalence of some of the interactions variables. As seen in figure fifteen below young people used less interactions than would be expected if age did not affect interactions and middle aged/older people used more. With young people using 278 compared to the expected 305 and old people using 97 compared to the expected 78. This was also the case for small hand gestures, although here the difference seems less large. With 75 observed interactions for young people compared to 86 expected interactions and old people using 29 compared to the 22 expected. This leads to the conclusion that there probably are significant differences between the prevalence of
interactions and the age of respondents. This difference does not seem to be there in the focused interactions were the observed and expected scores are much closer together.
Unfocused interactions Focused interactions
Type of interaction Observed quick glance Observed small hand gesture Observed ringing of bell Observed conversation Observed reciprocal hand gestures Observed longer period of eye contact Younger people 278 (305) 75 (86) 17 (13) 11 (8) 14( 15) 15 (19) Middle aged people 85 (77) 26 (22) 1 (3) 0 (2) 3 (4) 4 (5) Older people 97 (78) 29 (22) 2 (3) 1 (2) 7 (4) 9 (5)
Figure 15 osberved scores of interactions and expected chi2 score between brackets if the prevelance of interactions where perfectly independent of age
As expected by the difference between observed and expected scores younger people use significantly less unfocused interactions in the form of quick glances than older people. In figure sixteen the p values confirm this. With the only significant difference being between young people and old people. As can be seen in figure sixteen only the observation for the Shared space are used these differences only become greater. With small hand gestures and longer period of eye contact also becoming significantly more prevalent with older people compared to younger people.
