The contribution of football clubs to the international projection of cities: a causal analysis based on German and English cases

Abstract The study analyses the extent to which the presence of a football club with an international projection increases the visibility and global projection of the city that it belongs to. It specifically focuses on knowledge of a selection of English and German cities. The analysis is based on probabilistic methodology and uses data from surveys completed by Catalan university students. Our findings demonstrate that cities with famous football clubs are better known than others from the same country and similar urban hierarchy. Moreover, the best-known cities are also the ones where football has the greatest weight in the overall knowledge of the city. Furthermore, the study also demonstrates that the inclusion of the name of the city in that of the football club becomes a key factor for determining the intensity of the previously mentioned effects.


Introduction
Issues related to sports competitions form part of the everyday life of many citizens and are present in both the mass media and informal debate and also through active participation in sporting events. As a result, football clubs currently have an important level of social visibility. At present, some clubs have become important brands and have economic values that extend far beyond the purely sporting arena to include intangible aspects. This explains why Forbes (2017) magazine valued the net worth of individual clubs like Manchester United, FC Barcelona and Real Madrid to be each in excess of $3,000 million.
The attention given to these clubs implies an important level of visibility in the mass media and a level of global projection that reaches beyond their sporting activity and often also encompasses their host city (Edensor and Millington 2008). As a result, sports clubs may become ambassadors for their respective cities (Ginesta and de San Eugenio 2014). This positioning of a city within a category of spaces that is well-known to the whole population tends to bring with it a series of other important economic, social and cultural advantages (Glaeser 2011).
The development of city brands has been possible thanks to collaboration between public entities and private agents (Kavaratzis 2004), and the inclusion of sports within this strategy has helped to strengthen the resulting synergies. Hence, Olins (2005, 167) points out that sports can be used as powerful elements of urban marketing. Engaging in sport coincides with marketing objectives in that: (1) it establishes an elite group with representatives from the city; (2) it seeks critical audiences; (3) it creates a central idea of a place-specific brand, visualized through a widely known element and (4) it coordinates and integrates the messages of complementary agents. The importance of football as a brand therefore forms part of the commodification of the image of the city (Edensor and Millington 2008;Nauright 2000;Whitson and Macintosh 1996); football can help cities to reposition their image and to improve the development of their economic, social and cultural objectives (Lloyd and Peel 2008).
The general impact of sporting activity on the image of a city has already been widely studied. However, the accent has previously been placed on mega-events (such as world championships and the Olympic Games, etc.). In contrast, our study focused on the repercussions that the continuous activity of sports clubs can have on the international projection of their host cities.
The main objective of this study was to determine the extent to which the presence of an internationally recognized football club increases the visibility and global projection of the city that it belongs to. More specifically, the study calibrated the contribution of different football clubs to the knowledge of their host cities.
The study used five English and five German cities as case studies, and was based in data obtained from surveys to Catalan university students. We applied a probabilistic causal approach developed by Young, Fernandez Young, and Wu (2005). This method allows determining the influence exercised by a concrete variable in the production a given event. This methodology had previously been used by: Fernandez Young and Young (2008), within the context of tourist activity associated with films; Young et al. (2010), for tourism linked to fairs and markets; Saladié et al. (2014) and Anton Clavé et al. (2015), to determine how the choice of a sun and beach tourist destination is influenced by the presence of a low-cost airline; and Saladié, Gutiérrez, and Anton Clavé (2016), to quantify the same influence in the case of providing high speed rail services. The work was undertaken based on the following research hypotheses: 1. The presence of an internationally known football club implies a greater level of global exposure for its host city compared to other cities occupying a similar position in the urban hierarchy in the same country. 2. There is a causal relationship between the two variables: the greater the international fame of the football club, the greater its impact on knowledge of its host city. 3. The inclusion of the name of the host city in the name of the football club is key to determining the level of impact the football club has on the level of global knowledge of its host city. As a result, football clubs that do not incorporate the name of their host city have a lesser degree of impact on global knowledge of their host city (both in terms of the number of people who know the city and the weight that football has on that level of knowledge).
The paper contains six sections and is organized as explained below. Section 2 presents the reference context within which the research was carried out. Section 3 introduces and justifies the selection of case studies and the data and approach used. Section 4 presents and discusses the results obtained. Section 5 presents our conclusions and the implications of our study.

The role of football clubs in the promotion of cities
The creation and management of the image of a city is not an easy task. It depends, to a large extent, on the capacity of local policy-makers and numerous other actors to communicate and transmit perceptions of what it, and its inhabitants, are like to a potential public (Santos and Buzinde 2007). At the same time, the image that this potential public already has of the city in question will also have an impact on, and further reinforce, the resulting image (Paül i Agustí 2014). This complex relationship means that, at times, the images that different territories project are almost inevitably stereotyped; we are not capable of capturing all of the diversity in the world and must therefore fall back on generalizations. These results in the creation of global images of regions and their inhabitants which, in turn, produce an illusionary sensation of complete understanding and of order and stability for the citizen: a recomforting perception of having everything under control (Hottola 2012).
Within the framework of these global images, the individual identification of a given territory, which could be a city, is complex. The number of places that the whole of the population is capable of recognizing is extremely limited. The textual images of a region tend to be associated with just a few toponyms. A good example of this can be seen in the case of China. Pan and Lii (2011) showed how only a limited number of toponyms, including Xi'an, Beijing and the Great Wall, were well-known to a large part of the general public. Once these places had been named, the examples cited were far more diverse. This led to much greater sampling variability, but at the same time, there was a tremendous increase in the lagoons in people's spatial knowledge. A fact which helped people to form a yearned for mental image of the country but which, at the same time, relegated the majority of its cities to oblivion.
To a large extent, it could be said that communication is fundamental for being able to fix a new city in the collective imagination and/or to modify a pre-existing image. If a city is not known a priori, or does not have at its disposal any sufficiently important element of diffusion, it will lack the basic elements required to reproduce its symbology and begin the process of diffusion. This situation translates into an increase in the interest that different cities have in positioning themselves within the collective imagination. This explains why actions aimed at strengthening or providing a given city or region with a strong and recognizable image have increased, and not only in quantity, but also in complexity (Paül i Agustí 2013). From old strategies based solely on publicity, we have now passed on to much more ambitious ones that mobilize agents from a wide range of sectors. Amongst these agents, we find activities based on leisure and sport (Wojciech and Koenigstorfer 2016).
In the present urban context, local governments have transformed sporting activities into an element of the highest order, associated with the energy and dynamism of a particular location. Barcelona provides a good example of this. Sport, and especially FC Barcelona, has been placed as one of the eight most important elements for developing the city's tourism (López Palomeque 2015). The tourist-appeal of sport is at the same level as the city's infrastructure, cultural offer and commercial capacity.
The desire of local governments to position their cities in the collective imagination has led them to promote singular elements which help differentiate their city (Richards and Palmer 2010). Thus, football teams have become a major element of the urban promotion: it offers a strategic element that lends itself well to the goal of producing a singular city image (Edensor and Millington 2008).
Spaces associated with a previously known image and those that are related to people's own experiences (whether lived or imagined) offer more possibilities of being remembered by the population (Diaz-Soria 2015). The personal character and passion inherent to relationships associated with football would tend to further strengthen this argument. This led Gartner (1994) to observe that it is images created 'autonomously' , or in other words, those generated by agents who are not directly controlled by the city-destination (such as the press, personal relations and opinions found on the internet), which are the most effective. Again, football has an advantage here, as it already enjoys an important presence in the mass media.
Even so, to date, relatively few studies have focused on the relationship between football and the image of the city (Smith 2005, 218). The majority of the research undertaken has focused on three specific areas: (1) the positioning of the city as a venue hosting football matches (Kurtzman 2005 The present study focuses on an issue that has so far received rather less attention from academic sources: it studies the impact that football clubs have on the cities in which they are located. This is an approach which has only been adopted by a limited number of authors. Comparable examples would include studies referring to Olympique de Marseilles (Bromberger 2016), Borussia Mönchengladbach (Hamm 2015) and the classic reference to Glasgow Celtic and Glasgow Rangers (Murray 1994). Our study also presents the novelty of comparing results for various teams from two different national leagues: those of Germany and England.

Case studies
We selected ten cities in Germany and England as case studies. These are two territories with relevant demographic volumes and urban networks. In both cases, we refer to football leagues which have a tremendous global projection (the Bundesliga and the Premier League) and a large number of clubs of international prestige that regularly take part in the UEFA Champions League.
We selected five case studies from each country (Table 1 and Figure 1). The aim was to avoid selecting megacities (such as London or Berlin) which would have been well-known to a wide audience regardless of the presence of their football clubs. We therefore used the hierarchization of Rozenblat and Cicille (2003), as updated by Halbert et al. (2012). To be more precise, we selected cities included in the fifth and sixth levels of the hierarchy of the European urban network ( Table 1). The fifth level is defined as including cities that have a specific function at the European level. In contrast, cities in the higher hierarchies are defined as playing richer roles and having more complex networks. The sixth level includes cities which, without having a European dimension, exhibit an important degree of demographic dynamism. These are cities that are not state capitals, and which are not relevant tourist destinations in their own right, in which football clubs could be regarded as constituting an important element of the urban marketing.
The information relating to the urban network was complemented with data obtained from UEFA's all-time ranking by club (2015). To respond to the hypotheses formulated, we selected three case study typologies: • Typology A: cities with a football club (or clubs) that uses the name of the city and which has either played in the Champions League in the last few years and/or has achieved relevant results in its (or their) respective national leagues: Bremen, Dortmund, Leeds and Manchester.   • Typology B: cities in which the main football club does not use the name of the city and has either played in the Champions League in recent years and/or has obtained relevant results in its national league: Birmingham and Gelsenkirchen. • Typology C: cities that, at present, do not have football teams playing at the highest competitive level, independently of whether or not they use the name of the city: Bristol, Dresden, Freiburg and Southampton.
The objective was to establish various types of relations between clubs and cities which would allow us to quantify the different types of behaviour observed in relations between the two.
It is necessary to highlight two peculiarities of this classification. The last time that Leeds United participated in the Champions League was in the 2000/2001 season; this club currently plays in the second tier of the English league. As a result, this club could also have been assigned to Typology C. In the case of Birmingham, the city's main club, in terms of its history and success (Aston Villa), does not use the name of the city (Typology B). However, the city has a second club (Birmingham City) which has spent three seasons in the Premier League during the last ten years (and the rest of the time in the Championship: the English league's second tier), which in this case would correspond to Typology C.

Data
The study is based on the answers to a questionnaire by 273 students studying at universities in Catalonia, Spain. It is habitual to work with university students on this type of study (Dupré 2006;Guerrero Tapia 2007;Paül i Agustí 2016). Even so, the results obtained should be taken as indicative and exploratory; they represent a general trend in what is a changing group that can be treated as a homogenous unit based on a common characteristic (Gueben-Venière 2011, 4). Thus, the common characteristic was that the students surveyed were studying degree courses in Geography and Tourism at the Universitat de Lleida (UdL) and the Universitat Rovira i Virgili (URV).
The surveys were carried out during the months of December 2015 and January 2016. We ensured that the surveys did not coincide with weeks in which European football competitions were played in order to prevent this from influencing the results obtained. Table 2 shows the distribution of the 273 students (150 women and 123 men) who participated in the survey (confidence level: 95%; margin of error: 5%). The sample was also consistent with the distribution of students by university (with a greater presence of students from the URV), degree course (with a greater presence of those studying tourism) and gender (with a greater presence of women). The combination of the gender and university studies variable showed that the great majority of female students were studying tourism and that the male students were equally distributed between the two degree courses. On the other hand, the majority of the geography students were male. The average age of those who completed the survey was 21.5 years old.
To carry out the study, we showed those surveyed a list of ten cities (Birmingham, Bremen, Bristol, Dortmund, Dresden, Freiburg, Gelsenkirchen, Leeds, Manchester and Southampton) which were organized in alphabetical order, but without any form of geographical reference. Once they had read the names of the cities, we asked the students whether they knew the cities (binary answer: yes/no) and if they did, we then asked them to give one reason why they knew them (open answer). The students then answered another two questions which were only related to those cities that they had claimed to know. The answers given to these last two questions provided the scores for the causal analysis, which was the main objective of this study.

Methodology: causal chain approach
The causal analysis applied was a probabilistic approach developed by Young, Fernandez Young, and Wu (2005). It is based on the premise that when someone is exposed to certain information about a product or service, this exposure may have an effect on their decision to consume this product or service. The authors call this the causal chain. Even so, it must be underlined that a single cause will not necessarily be either necessary or sufficient to guarantee a given objective (Anton Clavé et al. 2015). In our case, the presence of a football club (F) may be one of the reasons why a city (C) is known. This will not, however, in itself guarantee that the city in question will become known. There may be other reasons that make knowing of the existence of the football club unnecessary for knowing about a given city. Even so, the presence of the football club increases the probability of people knowing its host city, as it adds another associated element to it. Thus, P(C/F), in Equation (1), is the probability of a person knowing a city due to the presence of a football team. In contrast, in the same equation, P(C/~F) is the probability of the city being known independently of the existence of its football club (due to knowledge attributable to other sources). The difference between these two probabilities gives us the increase in the probability of someone knowing a city that is attributable to the existence of its football club.
The two causal questions are therefore as follows: A) On a scale from 0 to 10, to what extent is it true that you know 'city X' due to the existence of a football club which is based in that city? -where 10 indicates that this is totally true and 0 indicates that it is not true at all. B) On a scale from 0 to 10, to what extent is it true that you would have known 'city X' , independently of the existence of a football club which is based in that city? -where 10 indicates that this is totally true and 0 indicates that it is not true at all.
The first of these two causal questions is designed to obtain the probability of knowledge of the city being due to the existence of the football club (ki in Equation (2)). The second question refers to the probability that those surveyed would have known about the city regardless of the existence of its football club (1 − c i ). Thus, ci is the probability that the city would not have been known without the existence of its football club.
The surveys gave us answers to the two causal questions in the form of scores ranging from 0 to 10. The two scores obtained from Equation (2) (k i and c i ) both evaluated the effect of the existence of the football club on knowledge of the city, albeit in different ways. The net effect was expressed with reference to their product (Δp i ), as indicated by Fernandez Young and Young (2008).
At one extreme of the range of replies were those surveyed who declared that the local football club had not had any influence on the fact that they knew the city and that they would have known about it city even if the football club had not existed (Q1 = 0 and Q2 = 10). The resulting scores for ki and ci would therefore be 0. At the opposite extreme, were those answering the survey who stated that the fact that they knew the city was only and exclusively due to the existence of the football club and that they would not have known the city if the football club had not existed (Q1 = 10 and Q2 = 0). Therefore, the scores for k i and c i would be 10. In the first group, the existence of the football club would not influence knowledge of the city, while in the second, only the existence of the football club would justify knowledge of the city. As a result, Δp i (Equation (2)) was the percentage, or fraction, of knowing about the city of each one of those surveyed which could be attributed to the existence of the football club. To calculate Δpi, the scores for k i and c i were converted to a scale ranging from 0 to 1.
This allowed us to measure the contribution made by the whole sample, or by selected segments of it, in relation to the sum of the individual fractions (Δp in Equation (3)). This is a more accurate way to make the calculation than to classify the respondents based on a binary type (yes/no) response, in which the individual fractions would be lost.

Knowledge of the cities and the different reasons for this
The research focuses, first of all, on the level of knowledge of the different cities revealed by students. Following Berthoin Antal and Friedman (2017), we have considered that getting to know a city means mentally differentiating distinct places and regions and seeing their location (in our case, in one State). The results revealed that only ten of the students surveyed (3.7%) knew all of them. On the other hand, only four students (1.5%) said that they did not know any of the cities highlighted. Table 3 shows the frequency of the possible results for the whole sample, taking into consideration the degree that the students were studying for (Geography or Tourism) and their gender. The highest frequency result (16.5%) corresponded to students who indicated that they knew five of the cities highlighted, although these were not necessarily the same cities. The frequency then progressively fell for both knowing more cities and fewer cities. For the whole sample, the average number of known cities was 5.2. This level of knowledge increased in the case of students studying Geography (6.2), and especially among male students (6.5). On the other hand, the average number of cities known to students studying for a degree in Tourism was 4.8, and 4.1 for female students. Here, the first relevant point to highlight was the limited knowledge that the students had of the different European cities. Paragraph: use this for the first paragraph in a section, or to continue after an extract. Table 4 shows the level of knowledge of the cities analyzed. Over the whole sample, the only city with a very high level of visibility was Manchester (>80%). Other well-known cities (60.1-80%) were Birmingham, Bristol and Dortmund. Bremen and Southampton were quite well-known (40.1-60%); Dresden, Freiburg and Leeds were not very well-known (20.1-40%); and Gelsenkirchen was very little known (≤20%). Table 5 shows the frequency of replies relating to the reason for indicating knowledge of the different cities. The open replies were aggregated in four different categories in order to make it possible to compare the results for the different cities: • city known because of its football club(s) • city known because the student(s) had visited it • city which friends or relatives of those surveyed had previously visited or lived in. • other reasons. A series of very heterogeneous replies were received, but these were repeated for all of the cities. These replies referred to such questions as the country, general culture, television, internet, mass media, books and topics studied in class. In the same way, for some cities there were some very specific replies. The most outstanding was for Bremen, the fifth best known city. 46.9% of those surveyed related it to its football team: Werder Bremen. However, 15.6% related it to 'the musicians of Bremen' , one of the tales by the Brothers Grimm. Other links that stood out for a single city were the association between Manchester and the Industrial Revolution, Dresden and the  (150) Male (123) 10 3.7% (10) 1.5% (3) 9.2% (7) 0.7% (1) 7.3% (9)  9 8.4% (23) 6.1% (12) 14.5% (11) 2.0% (3) 16.3% (20)  8 8.8% (24) 7.6% (15) 11.8% (9) 3.3% (5) 15.4% (19)  7 10.3% (28) 10.2% (20) 10.5% (8) 8.0% (12) 13.0% (16)  6 11.7% (32) 12.2% (24) 10.5% (8) 9.3% (14) 14.6% (18)  5 16.5% (45) 15.7% (31) 18.4% (14) 18.0% (27) 14.6% (18)  4 13.2% (36) 14.7% (29) 9.2% (7) 19.3% (29) 5.7% (7)  3 11.0% (30) 11.7% (23) 9.2% (7) 15.3% (23) 5.7% (7)   bombing that it suffered in the Second World War, Bristol and its port, and Southampton as the port from which the Titanic sailed on its maiden voyage.
In Table 5, it can be observed that football was an important element for people recognizing the different cities. Even so, this recognition differed enormously in relation to the importance of the team. Manchester and Dortmund were the cities that were most closely associated with football. More than two-thirds of the students who said that they knew these cities (268 and 166, respectively) indicated that this knowledge was due to football or because they associated the city with a specific football team: Manchester United and Manchester City, while Borussia Dortmund was the second most important team in the Bundesliga after Bayern Munich. In contrast, Gelsenkirchen was the least known city. It was identified by only 47 students, with just over half of these associating the city with football. This is the city where Schalke 04 is based. In recent years, this club has participated in various editions of the UEFA Champions League and for three seasons Raúl González, a well-known player and ex captain of Real Madrid and the Spanish national team, played for the club. This was a relevant issue as the students surveyed were Catalan.
In contrast, the cities that those surveyed least associated with football were Dresden, Bristol and Birmingham. The German city was the third least known, with only 78 positive responses and with only 6.4% of these relating it to football. It should be added that Dynamo Dresden, a historic football club from the former German Democratic Republic, was in the German second division at the time of the survey. In the case of Bristol, the main team is Bristol City. Its greatest achievement was finishing second in the English league, but that was back in 1901. During the decade prior to the survey, this team had only played in the lower tiers of the English league. After Manchester, Birmingham was the second bestknown city on the list, although fewer than a quarter of the people surveyed said that they knew of the city because of its association with football. As in the case of Gelsenkirchen, the city's most important football club does not use the name of its host city. Aston Villa was playing in the Championship (the second tier of English football) in the 2015-2016 season. Even so, historically speaking, it has been one of England's most successful clubs and indeed won the European Cup (current Champions League) in the 1981-1982 season. The city's second most important team is Birmingham City, whose history is much more modest.
The cities with the highest percentages for being well-known for 'other reasons' than football were Birmingham, Bristol and -most clearly of all -Dresden (82.1%). In other words, these were the three cities for which football had the lowest relative influence on them being known. At the other extreme were Leeds, Dortmund and Manchester (21.3%),  (57) with the second and third of these cities being the ones for which football had the greatest relative influence upon their fame. Finally, Table 6 shows the frequency with which football was cited as the main reason for cities being well-known broken down by the gender of those surveyed and the degree that they were studying for. Football was the answer given by more than half of those surveyed for seven of the ten cities in the case of males and for two (Dortmund and Manchester) in the case of females. Moreover, the table also shows in brackets the amount of answers for each case. Dortmund, Manchester and Bremen are the only cases that achieve at least 10 answers of a woman -within the 150 surveyed -highlighting that football is the main reason. Significant differences between male and female answers were observed in the cases of Birmingham, Bremen, Freiburg, Gelsenkirchen, Leeds and Southampton. In fact, a greater reference of football as main cause for the knowledge of cities is reported by men in all cases.
Along these lines, we should point out the relationship highlighted by several studies about football and masculine gender. In contexts such as the English one, there have been wider studies of the ties between women and the world of football. These studies have pointed to the link between female fans and football being built at the level of social interaction. In this case, female fans of mass interest male sports would tend to find limitations to their interactions. Moreover, some males have traditionally questioned the identification of females with football clubs (Woodhouse and Williams 1999). Following this same line, some other studies have underlined how the lack of a favourable environment and appropriate social references have made it difficult to establish close ties between women and football (Crawford 2004, 47). This would tend to lead them into establishing a different set of relations and identification with the world of football.

Causal analysis results: the incidence of football clubs on the level of knowledge of cities
Knowing a city can be related to various different causes. It is also probable that a given city would have been equally well-known even if its football club had not existed. In the same way, the football club could also have contributed to making the city well-known amongst those who said that they knew it. This could have been because either they, their friends or their family had visited it, or for a number of other reasons. Taking both of these premises into account, we sought to establish the increase in the probability of knowing each of our ten cities due to football. The result was based on the replies given to the two causal questions by students who said that they knew the city in question.  Table 7 shows the distribution of the extreme scores for the two causal questions that were asked with respect to each of the ten cities analyzed. The relative weight of the students who knew the city and gave an answer of 10 to the first question -in other words, those who were completely certain that they knew about the city because of its football club -ranged from 40.6%, in the case of Dortmund, to 5.3%, in that of Dresden. In contrast, the students who gave 0 as their answer -which was equivalent to saying that their knowledge of the football club was not at all related to their knowledge of the city -ranged from 73.3%, in the case of Dresden, to 4.5%, in that of Manchester. The distribution of extreme scores in answer to the second causal question showed that Dresden was the city which students were most certain (Q2 = 10) that they would have known irrespective of whether it had had a football club (49.3%). Gelsenkirchen and Dortmund were the cities to which the greatest percentage of students gave a score of 0 (27.7% and 26.3%. respectively). In this case, this means that it was completely certain that if their football clubs not existed the students surveyed would not have known these cities. Taking into account the combination of the extreme scores for each of the two questions, we concluded that the cities with the highest percentages of students whose knowledge of them was almost totally attributable to football were Southampton (24.6%) and Gelsenkirchen (23.4%). In contrast, the lowest scores corresponded to Dresden (2.7%) and Bristol (6.1%). These were awarded by students whose knowledge of the city was due only and exclusively to the existence of its football club (Q1 = 10, then k i = 1); and who stated that they would not have known the city at all if its football club had not existed (Q2 = 0; then c i = 1). In these cases, the increase in the probability of knowing the city due to the existence of the football club was 100% (Δp i = 1).

Answers to causal questions
At the other extreme were the students who stated that the football club had not had any influence at all on their knowledge of the city (Q1 = 0; then k i = 0) and who said that they would still have known about the city even if its football club had not existed (Q2 = 10; then c i = 0). The increase in the probability of knowing the city attributable to its football club was therefore 0. The cities with the highest scores for this scenario were Dresden (49.3%) and Bristol (43.1%), while those with the lowest were Dortmund (1.9%) and Manchester (2.6%). At this point, it is important to remember that these two pairs of cities (Dresden-Bristol and Dortmund-Manchester) were the ones for which football as the reason for knowing the city respectively received the lowest and highest scores.
Even so, it must be noted that the majority of the students surveyed introduced subtle nuances into the answers that they gave to the two causal questions based both on the greatest and least influence that its football club had on their knowledge of the city (Q1) and on the greatest and least knowledge that they would have had of the city, regardless of its football club (Q2). Figure 2 shows the distribution of the students who said that they knew Manchester and Dortmund (the two cities that were best known due to football). The graphic  Figure 2. Distribution of the students surveyed who stated that they knew the cities based on the increase in the probability of them knowing a city that was attributable to its football club; results for Manchester (above) and Dortmund (below).  Figure 3. Distribution of the students surveyed who said that they knew the cities based on the increase in the probability of them knowing the city attributable to its football club; results for Gelsenkirchen (above) and Dresden (below).
representation fractionally weights the increase in the probability of knowing these cities attributable to football for each of the people surveyed who said that they knew them. Figure 3, on the other hand, shows this same distribution for Gelsenkirchen (the least known city: 17.2%) and Dresden (the city which was least known on account of football: 6.4%).
If football had not had any influence at all on knowledge of the city, the individual score would have been 0. This was the score assigned when students answered 0 to question 1 and gave a score of other than 10 in response to question 2. It was also the score for those who gave a score of 10 in response to question 2 and a score of other than 0 in answer to question 1. In all the other cases, the bars show the increase in the probability of the city being known attributable to its football club according to each person surveyed. The minimum score of 0.01 corresponded to an increase of 1%, while the maximum score of 1 corresponded to 100%. In this case, knowledge of the city exclusively attributable to the existence of its football club would be absolutely certain.

Contrasts in causal analysis results between cities
From the individual results obtained by combining the two causal questions, we were able to quantify the direct influence of the football club on knowledge of the city (k), the probability of this knowledge being the same whether the football club had existed or not (1 − c), and the increase in the probability of knowing the city attributable to the existence of the football club; results for the whole sample and for each of the ten cities (Table 8).
Manchester and Dortmund had k scores of greater than 0.7. Gelsenkirchen was in third position, with a score of 0.574. Similar scores, all of which were greater than 0.5, were registered for Bremen, Freiburg and Southampton. Although these last three teams currently play in their countries' top divisions, Freiburg is a club that tends to go up and down between the German first and second divisions. It is significant to note that, in contrast to Gelsenkirchen, the main football club from each of these three cities bears the name of its city.
A lower degree of influence was found in cities where the local football team played in a lower division of the national league: Dresden and Bristol. This was, however, observed in cities whose clubs enjoyed only limited projection. Finally, the city of Birmingham was a special case. Although it had a team playing in the Premier League when the survey was carried out, its k score was the third lowest (0.399). This result could largely be explained by the fact that the city's main club (Aston Villa) does not bear the name of the city; furthermore, unlike Schalke 04, Aston Villa has only had a rather testimonial presence in European competitions in recent years. It has sporadically participated in the Europa League, but that is widely regarded a second tier competition.
The results presented a lesser degree of dispersion with respect to the existing knowledge of the city, independently of the existence of the football club. As expected, this score was highest for the cities of Birmingham, Bristol and, especially, Dresden (0.735), while it was only 0.4 in the case of Gelsenkirchen. If we compare these scores (1 − c) with the percentage of people surveyed who said that they knew the cities, but not because of football, it is possible to identify three different groups: (1) very low 1 − c scores; (2) very high 1 − c scores; and (3) very similar 1 − c scores. The first group includes the cities of Dresden, Bristol and Birmingham, while second includes Manchester and Dortmund. In the former cases, some of the students who indicated that they knew the city also said that they knew the football club, but they did not select that option when they justified their knowledge of the city. On the other hand, the cities whose football clubs receive the most media attention (Manchester United, Manchester City and Borussia Dortmund) were also the ones that had highest 1 − c scores for justifications not linked to football. It was shown there were also other relevant elements of knowledge attached to other questions and the results derived from the second causal question allowed us to globally quantify them. As a result, the increase in the probability of knowing the city (Δp) created by the existence of the football club was very similar for six of the ten cities. Their scores ranged from that of Manchester (0.348) to that of Southampton (0.391). Four of these cities were those initially assigned to 'Typology A' (cities whose football clubs have recently played in the UEFA Champions League): Bremen, Dortmund, Leeds and Manchester. The other two were cities corresponding to 'Typology C' (cities whose football clubs have not recently played in the UEFA Champions League), which play in the highest tier of their respective national leagues: Freiburg and Southampton. These are cities which were known to different degrees by the students surveyed. However, they share the common characteristic of having had clubs with a certain relevance in the Premier League and Bundesliga in recent decades.
In contrast, the smallest increase in the probability of the city being known associated with the existence of its football club corresponded to the cities of Bristol (0.142) and Dresden (0.08). Both were initially assigned to 'Typology C' and were -at the time of the survey -playing in the second tier of their respective national leagues. The students surveyed had a good knowledge of the English city but a much more limited knowledge of the German one. The nexus of union was their football club. Dynamo Dresden had last played in the Bundesliga in the 1994-1995 season, whereas Bristol City's most recent participation in the highest tier of the English football league dated back to the 1979-1980 season. Along these lines, it is important to remember the youth of those surveyed (average age 21.5); this no doubt influenced the fact that the most recent sporting achievements of the different clubs had the greatest impact on the results.
Birmingham and Gelsenkirchen constituted special cases ('Typology B': football clubs that do not use the name of their city). Those surveyed had a good knowledge of the English city and the increase in the probability of them knowing it as a result of the existence of its football team (0.217) was greater than for Dresden, but smaller than for the first mentioned group of six cities corresponding to Typology A and Typology C. The main club from the city of Birmingham (Aston Villa) is a historic member in the top division in England, but this club does not use the name of its city. A similar situation is found with the club based in the city of Gelsenkirchen; Schalke 04 has uninterruptedly participated in the Bundesliga, and sporadically in the UEFA Champions League, and won the UEFA Cup in 1996. However, the knowledge that the students had of this city was very limited. However, the increase in Table 8. Direct influence of the football club on knowledge of the city (k), the probability of that knowledge being the same without the presence of the football club (1 − c) and the increase in the probability of knowing the city attributable to the football club (Δp the probability of them knowing this city attributable to the existence of its football club was the greatest of all the ten cities studied (0.442). The knowledge of this city would therefore have been clearly lower if its football club had not been so successful, in both the Bundesliga and the UEFA Champions League, in recent years. Having said that, knowledge of this city would have been expected to have been even greater if its football club had borne the name of the city. This more limited knowledge of the city would also have been noted, albeit to a lesser or greater degree, in the other cases if the local football club had not existed. The smallest variation in this respect corresponded to Dresden. The limited initial knowledge of this city amongst those surveyed did not change greatly because the football club was not an important element for identifying this city. At the other extreme, the English city of Manchester, exhibited the fourth smallest increase in the probability of knowledge of the city being closely associated with the existence of its football club(s). Manchester was well known by 98.2% of those surveyed. Without its football teams, it would have continued to be the best-known city, but this fame would probably have been significantly reduced. The two teams from Manchester, and especially Manchester United, played an important role in the knowledge of the city.
Finally, Table 9 shows the increase in the probability of knowing the city thanks to its football club (Δp) broken down by the degree studied and the gender of those surveyed. In line with previous results, the influence of football was greater amongst men than women. This difference was observed, to a greater or lesser extent, for all of the cities studied except Manchester; this was probably because Manchester's football clubs are much better known than the others. The results presented in the table confirm the deep differences in the effect of football clubs in knowledge of cities between men and women. The notable presence of males studying Geography and females studying Tourism explain the differences in results between those studying these two degrees. These findings allows us to highlight that, with the exception of the really well-known football clubs (Manchester), the positive contribution of football clubs on the knowledge level of cities is clearly concentrated on men.

Conclusions
The results of the study show the importance that top-level football clubs have for the projection of cities. In each case, we used a Causal Chain Approach to give a weight to the increase in the probability of a city being known due to the existence of its local football club(s). The first research hypothesis was therefore confirmed: the presence of an internationally famous, top-level, football club gave a city a greater level of global visibility compared with other cities from the same country and urban hierarchy. The level of this influence varied according to each club's sporting success in the UEFA Champions League and the global projection of its football club (see Borussia Dortmund and Manchester United and Manchester City). Even so, for this relationship to exist, it is necessary for there to be a clear association between the name of the city and the name of the football club. Without a shared name, the image of the city tends not to be associated with sport and the benefits that this projection offers the city will tend to be fewer. The second hypothesis, related to the causal relationship between the two variables, was also confirmed. We can therefore affirm that the existence of an internationally known football club not only increases the visibility of the city, but also increases knowledge of the city directly related to football. This was especially relevant in the cities that were less wellknown for other reasons. Along these lines, we could add that football has become an element of popular urban discourse and helps to enrich the number of attractions for which a city may be known.
From what has been previously stated, it would also be possible to deduce the veracity of the third hypothesis. The inclusion of the city in the name of a football club seems to be a key factor for determining the level of impact that the football club has on the knowledge of the city that it belongs to. As a result, football clubs that have not incorporated the name of their city tend to have less impact on the overall visibility of this city (both in terms of the number of people who know it and on the influence of football on this level of knowledge). In such cases (as those of Aston Villa and Schalke 04), the football club may be well-known, but it is not necessarily associated with any specific city.
Previous studies had underlined the significance of certain mega-events (Olympic Games, UEFA European Nations' Cup and FIFA World Cup) and their impact on the visibility and global projection of the cities that have hosted them. The results of our study show that the presence of professional football clubs can also contribute to a greater knowledge of a given city. They can also help to enrich the image of cities by incorporating new attributes, such as sporting connections, into the collective images of cities that might otherwise have poorly defined profiles. In this sense, football clubs can serve as a good urban marketing tool and make an important contribution to generating positive dynamics in a city. It has also been shown that image of a football club can have a far-reaching influence. The visibility of Leeds United was high amongst those interviewed, despite the fact that this club has not participated in any form of elite competition during the last decade and that its main international successes date back to before 2001.
This situation is important as it shows us the potential persistence of the image associated with football. Its persistence would contrast with mega-events limited effects in time (Roche 2017). The football club's continuously and long-lasting presence in people's consciousness could contribute to generate repercussions even more important for a town's visibility than mega events, which take place only occasionally. An affirmation that should be confirmed with more extensive studies.
Finally, based on the results of this work, it has been possible to identify a number of aspects that could serve as the object of future research. One key question, in this respect, would be to analyze the hierarchization of the different factors that contribute to the visibility provided by a football club: its players, titles, shirt and sponsors, the name of its stadium and historic events, etc. Specifically identifying these influences would make it possible to identify the weight attributable to each component and to analyze its role in the projection of the city. For instance, some studies have highlighted the role of superstars on city branding (Currid-Halkett and Scott 2013). Thus, the role of football stars would be an interesting topic to be addressed. However, most of them are concentrated in the most well-known football teams. Another element of interest could be the contribution of football clubs to the visibility not only of their own cities, but even of the regions or states in which they are located.