La thèse s'est donnée pour objectif d'élaborer une approche innovante de la corruption, permettant de saisir les réalités complexes du phénomène et d'expliquer sa persistance dans le monde émergent. Suite à une analyse critique de l'orthodoxie économique de la corruption - la compréhension actuelle la plus répandue comprenant un mix d'approches théoriques et culminant avec le modèle de bonne gouvernance -, nous soulignons la faiblesse d'attention accordée aux facteurs structurels. Dans notre recherche nous avons effectué une enquête de terrain (méthode Delphi) et une analyse statistique exploratoire multidimensionnelle qui ont permit des nouvelles interprétations du phénomène observé. En nous appuyant sur les enseignements des écoles hétérodoxes contemporaines (unifiées au sein de l'économie politique institutionnaliste décrit par A. Caillé), construisant une approche positive et compréhensive basée sur une démarche holindividualiste, nous proposons une analyse structuro-institutionnaliste de la corruption. Avec un renouveau paradigmatique sur la rationalité et la coordination des agents économiques, nous avons reconstruit le cadre interprétatif de la corruption. Dans ce but, nous avons élaboré le concept de la corruption institutionnalisée, qui décrit les situations où la corruption n'est plus une déviance, mais, au contraire, est devenue une règle de comportement. Nous avons étudié les mécanismes de la corruption institutionnalisée à travers la mise en évidence du rôle des attitudes culturelles, des principes inhérents aux réseaux sociaux et des caractéristiques néo-patrimoniales des États contemporains. Ainsi, nous avons expliqué les origines de la corruption, qui renvoient à son historicité et à l'encastrement (culturel, social, étatique), les mécanismes de fonctionnement reposant sur des coordinations hybrides, ainsi que les impacts et liens ambigus de la corruption institutionnalisée avec la croissance. Notre analyse dynamique de la corruption se penche sur les questions de transformations des systèmes néo-patrimoniaux à travers la compréhension d'évolutions institutionnelles, où la path dependency et les actions des clans politico-économiques expliquent la résistance et la reproduction des systèmes corrompus. Nous proposons des scénarii de sortie du clientélisme des États néo-patrimoniaux et mettons en avant les enjeux et les leviers d'une évolution graduelle endogène. Nous avons attiré l'attention sur le rôle du clientélisme dans la construction démocratique (un vecteur paradoxal), où les réseaux de réciprocité rendent la démocratisation possible. Nous avons mis en lumière deux moments clés dans une évolution vertueuse : la déconcentration (ou dispersion) des pouvoirs à travers l'autonomisation de certains réseaux par rapport aux grands clans politico-économiques, et la diversification des ressources de nouvelles unités, rendue possible notamment à travers le processus d'insertion dans l'économie mondiale. La thèse aboutit à une série de propositions concrètes pour l'amélioration de la pertinence des programmes anti-corruption. Nous reconsidérons le rôle de l'extérieur et mettons en évidence la supériorité des programmes bottom-up dans le combat contre la corruption, celui-ci devant reposer sur les évolutions graduelles des structures sociétales. La portée de notre approche dépasse le monde émergent et est applicable à différents systèmes économiques. En effet, la corruption institutionnalisée n'est pas spécifique aux pays émergents et ne peut être donc réduite à une dynamique de développement. / This thesis develops an innovative approach of corruption in order to capture the complex reality of the phenomenon and explain its persistence in the emerging world. After a critical analysis of the economic orthodoxy of corruption, i.e. the most common understanding of corruption based on a patchwork of theoretical approaches culminating with the model of good governance, we emphasise the weakness of attention on structural factors. In our research we conducted expert's opinion survey (Delphi method) and multidimensional exploratory statistical analysis (factoral analysis) which led to new interpretations of the observed phenomenon. Relying on lessons of contemporary heterodox schools (unified within the institutionalist political economy described by A.Caillé), building a positive and comprehensive approach based on holindividualism, we propose a structuro-institutionalist analysis of corruption. With a renewed paradigm on rationality and coordination of economic agents we rebuilt the interpretative framework of corruption. In this regard, we have developed the concept of institutionalised corruption that describes situations where corruption is no longer a deviance but a rule of behaviour. We studied the mechanisms of institutionalised corruption by highlighting the role of cultural attitudes, of inherent principles in social networks and neo-patrimonial characteristics of contemporary states. In this way, we explain the origins of corruption referring to its historicity and embededness (cultural, social, institutionnal), the operating mechanisms based on hybrid coordination, as well as the effects and ambiguous ties between institutionalised corruption and economic growth. Our dynamic analysis of corruption focuses on transformation patterns in neo-patrimonial systems through the understanding of institutional evolutions, where the path dependency and actions of politico-economic clans explain resistance and reproduction of corrupt systems. We suggest scenarios to get out of clientelism in neo-patrimonial states, where we put forward issues and levers of gradual and endogenous evolutions. In our analysis we emphasise the role of clientelism in democratic construction (a paradoxal vector), where reciprocity networks make democratisation possible. We highlighted two key moments in a virtuous evolution: the decentralisation (or dispersion) of powers among clientelist networks and the diversification of autonomous networks resources, which are notably possible throught integration in global economy. The thesis leads to a set of concrete proposals for improving the relevance of anti-corruption programmes. We reconsider the role of international integration and highlight the superiority of the bottom-up programmes in the fight against corruption, which must be based on the gradual evolution of societal structures. The significance of our analysis exceeds the emerging world and is applicable to different economic systems. Indeed, institutionalised corruption is not specific to emerging or developing countries and therefore can not be reduced to the dynamic of development.

The purpose of this study was to discover what impact, if any, personal social networks have on an individual’s snowboarding skill progression, and whether any differences emerge based on gender, class, race, or age. Interviews with 10 demographically diverse snowboarders in Washington State, USA, were conducted and analyzed. The study revealed that personal social networks are highly important to snowboarding skill progression, with nine progression benefits noted. It was also found that the type of relationship was important, with the roles of gatekeepers, “godfathers,” and crews as the most critical for progression. Lastly, respondents indicated a variety of methods to access social networks, including social media, events, organizations, industry employment and through existing social networks. To further analyze the findings, an intersectional feminist reading of Castells networking theory was applied, with investigations into gender, race, class, and age patterns. The analysis revealed two overlapping values systems, one based on snowboarding ability and commitment, and one based on alignment of demographics with those who are most valued in the snowboarding world--mainly young, white, middle class men. The research closes with potential solution ideas to improve equitability and inclusion, which can be applied from the grassroots level to large-scale implementation.

Traditional investment practices have been revamped with more and more modern methodsto fit entrepreneurial settings. The evolution of equity crowdfunding has created an alternative to venture capital,bank loans and business angels. Due its novelty, recent studieshave onlyfocusedon its mechanisms and financial regulations. We draw from crowdfunding concept and social theories to develop an understandingonthe connection between social media networkand equity crowdfunding success.Utilizingdata from LinkedIn and a leading equity crowdfunding platform in the Nordics, we explore the impact of social media network to funding campaigns’ success. We collected data on every company on the platform, and we collected data on the key members’ LinkedIn connections. The results propose that it is possible to predict certain types of success in equity crowdfunding based on the number of LinkedIn connections. This study contributes to existing literature by providing better understanding on another social network, LinkedIn, which is often connected to project pages on crowdfunding platforms. We contribute even for improved insight in decision making processes behind the investments; which is in high interest of entrepreneurs, investors and platforms

This thesis is about entrepreneurial leadership and how entrepreneurs are affected by social media. Our aim with this study is to investigate how social media affects entrepreneurial leadership. One of the main questions we are trying to figure out is, can social media be used as a tool for entrepreneurial leadership and if that is the case, how? To be able to answer the aim, the thesis takes a qualitative researched approach, we conducted individual face-to-face semi-structured interviews with entrepreneurs. The data from the interviews will be analysed using grounded theory coding with the approach of finding 1st order concepts to look for patterns in the data.  The 1st order concepts led us to the 2nd ordered themes, which became three aggregated dimensions; cast expander, leadership enhancer and new social context.   The result gave us a deeper understanding of entrepreneurial leadership and how the entrepreneurial leader perceives social media as a leadership tool. By combining our three aggregate dimensions, cast expander, leadership enhancer and new social context we were able to build a model of how these three dimensions intersect. What makes this model useful is that it gives us some insight into how social media affects entrepreneurial leadership

In this thesis I study how networks are formed and I analyse the strategies that well-connected individuals adopt in public good games on a network. In chapter one I study an artefactual field experiment in rural India which tests whether farmers can create efficient networks in a repeated link formation game, and whether group categorisation increases the frequency of in-group links and reduces network efficiency. I find that the efficiency of the networks formed in the experiment is significantly lower than the efficiency which could be achieved under selfish, rational play. When information about group membership is disclosed, in-group links are chosen more frequently, while the efficiency of network structure is not significantly affected. Using a job-referral network experiment in an urban area of Ethiopia, I investigate in chapter two whether individuals create new links with the least connected players in the network. In a first treatment, competition for job-referrals makes it in the player's interest to link with the least connected partners. In this treatment, links to the least connected players are significantly more likely than links to better connected individuals. In a second treatment, connections only affect the welfare of the new partner. Choosing the least connected player minimises inequality and maximises aggregate efficiency. This may motivate other-regarding players. In this treatment, however, links to least connected partners are not significantly more likely than links to other players. In chapter three I explore the characteristics that individuals value in the people they approach for advice. Using cross-sectional data on cocoa farmers in Ghanaian villages and a matched lottery experiment, I find an association between the difference in the aversion to risk of two farmers and the probability that one farmer is interested in the advice of the other farmer. In chapter four I study a one-shot public good game in rural India between farmers connected by a star network. Contributions by the centre of the star have a larger impact on aggregate payoffs than contributions by the spoke players. I use the strategy method to study whether the centre of the star contributes more than the average of the spokes. In selected sessions, I disclose participants' expectations about the choices of the centre of star. I find that the centre player contributes just as much as the average of the spokes, and that he is influenced by the expectations that other players hold about his decisions.

How do informal groups emerge in adolescent communities? What distinguishes a group from just a set of students? Who will end up together in a group and who will be left out? Why are there more groups in some classrooms and fewer in others? What determines whether these groups overlap in their members or they are completely segregated, perhaps antagonistic? While a huge body of research in sociology and social psychology focuses on these questions, an integrated approach that is able to answer all of them is yet to be developed. Without realizing that these five issues are interrelated, we cannot hope to understand how groups influence individuals and how they shape our communities. This thesis proposes an integrative theory of informal group formation in communities. Based on the tradition of Social Network Analysis, it develops a framework in which interpersonal relations and reputations are formed through a process called informal status production. Groups emerge from this micro-process by the alignment of positive relations and agreement in peer-perceptions between actors. The main micro-mechanisms predicted by the theory are tested on a unique longitudinal network dataset from school classrooms. To this end, a new empirical procedure was developed, by which a few aggregated networks can be constructed from tens of relational items. This allows the analysis of rich network data with several relational dimensions. The empirical studies of multiplex network dynamics confirm that there are strong interdependencies between friendships and perceptions. Students who agree about their peers tend to become friends, but more so when they hold a minority opinion in the class. This contributes to group formation. Friends also influence each other's perceptions, but we manage to show that the presence of groups around them interferes with this process by moderating the influence of individual peers.

This thesis focuses on the dynamic interplay between race and social ties. Even though in sociological studies, race is usually treated as a cause of social segregation, we argue that this is a two-way process. Our approach distinguishes between racial self-identifications and racial perceptions. In the first part of the thesis, we focus on the joint effects of these aspects on the prevalence and emergence of social ties in secondary school communities. The second part investigates social effects on racial perceptions. For the analyses, we take a social networks approach, estimating exponential random graph and stochastic actor-oriented models. First, we take a look at the state of racial segregation in friendships and negative ties within communities, and we investigate the dynamic processes that have led to the described state. We also take endogenous network mechanisms into account. We provide evidence that given an initial state of segregation, reciprocity and clustering can maintain the relative infrequency of cross-race friendships in the group, even without (additional) same-race preference. Further, we find that negative ties describe interracial segregation better than friendships: majority students tend to dislike their minority peers, but no such tendencies were found for friendships. Second, we find that minority students tend to overperceive their friends' similarity to themselves in terms of race. Moreover, there is evidence for social influence: classmates tend to accept each other's, especially their friends', opinions about their peer's race. Altogether, both empirical parts of the thesis suggest a hierarchical relationship between the majority and the minority groups: besides majority students' tendency to exclude their minority peers, those who try befriending majorities - but get rejected by them - are more likely to be perceived as minorities. There are also indications of some minority students showing outgroup preference, while others seem to compete against the majority group. This can contribute to the observed emergence of enmity between minority students.

To offer a deeper understanding of how Peruvian firms internationalize this paper aims to throw in an acknowledged theoretical gap in international business research by giving a contribution towards the internationalization of Multilatinas based on a network perspective. The purpose is to investigate what drives Peruvian Multilatina's to internationalize, what market they select, how they enter these markets, and how different network relationships can influence their internationalization. Research question: What does the internationalization process look like for a Peruvian Multilatina? Method: Includes a qualitative research method with the use of semi-structured interviews. A multiple case study was utilized to gain comparative insights into how Peruvian Multilatinas internationalized through their networks relationships. For the analysis, the theoretical framework consisted of: internationalization drivers, market selection, choice entry mode, and importance of networks. Conclusions: The results conclude that Multilatinas do not follow a unique pattern of internationalization, but follow an accelerated expansion strategy based on the context of their political and economic environment. The Multilatinas studied started off by exporting to markets that had similar characteristics to their own. Once they had adequate experience and knowledge of operating there, they followed up with contractual and investment entry modes. The Multilatinas relied on different network relationships in order to internationalize. The most important being the business relationship and institutional relationship in order to gain insidership.

Nous considérons les Réseaux Sociaux Mobiles Tolérant le Délai (DTMSN), constitués de nœuds sans-fil avec une connectivité intermittente, et groupés en communautés sociales. Cette thèse traite de l’analyse et de la conception de stratégies de transfert de l’information dans les DTMSN. Elle est principalement dédiée à l’étude de codage réseau inter-session (ISNC) dans ce but. Le codage réseau est une généralisation du routage et ISNC est connu comme un problème d’optimisation difficile en général, spécifiquement parce qu’il peut vite devenir nuisible si non conçu avec soin. Le premier chapitre répond théoriquement au problème d’optimisation du routage (sans ISNC) dans les DTMSN. Nous généralisons les résultants existants pour les topologies homogènes. Le deuxième chapitre conçoit et modélise un contrôle de ISNC par paire, qui englobe conjointement le contrôle du routage et du codage, avec une contrainte d’énergie. Pour s’attaquer de façon heuristique à l’optimisation de ce contrôle, le troisième chapitre présente une étude expérimentale visant à identifier quand ISNC est bénéfique ou nuisible, en fonction du nombre maximum de copies par paquet, de la charge du réseau, de la taille de buffer des nœuds relais et de la gestion de buffer. Le quatrième chapitre présente la conception de critères décentralisés de codage, pour déclencher en ligne le mélange de sessions si ISNC peut être bénéfique. Nous testons ces critères sur des topologies simples et sur des traces réelles, en expliquant les limites de notre approche. / We consider Delay Tolerant Mobile Social Networks (DTMSN), made of wireless nodes with intermittent connections and clustered into social communities. This thesis deals with the analysis and design of information transfer strategies in DTMSN. It is mostly dedicated to investigate the use of Inter-Session Network Coding (ISNC) towards this goal. Network coding is a generalization of routing and ISNC is known as a difficult optimization problem in general, specifically because it can easily get detrimental, compared to no coding, if not designed carefully. The first part of this thesis addresses theoretically the optimization problem of the (non-ISNC) routing policy in DTMSNs. We generalize the existing results for homogeneous topologies. The second part of the thesis designs and models a parameterized pairwise ISNC control policy that encompasses both routing and coding controls with an energy constraint. In order to tackle heuristically the optimization problem, the third chapter presents an experimental study of pairwise ISNC to investigate when it can be beneficial or detrimental. We examine the impact on ISNC performance of a number of parameters, such as the constraint on the maximum number of copies per packet, the network load, the buffer size of the relay nodes and the buffer management policies. The fourth chapter addresses the design of decentralized coding criteria allowing to trigger online session mixing if ISNC may be beneficial. We test these coding criteria on both toy topologies and real-world traces, pointing out and explaining the limits of our approach.

O presente trabalho propõe tipos de comunidades virtuais para redes sociais compostas de páginas pessoais para fotografias, que permitem que outras pessoas comentem-nas (fotologs) observadas no sistema Fotolog.com. A definição de comunidade virtual parte de uma análise dos conceitos denominados utópicos (TÖNIES, 1947 e 1995; DURKHEIM, 1978; BUBER, 1987) e dos conceitos denominados contemporâneos (WEBER, 1987; WELLMAN, 2001 e 2002; MAFFESOLI, 1996 e 2000) e da compreensão de seus elementos em comum: interação social, laços sociais e capital social. Essa definição é levada em conta conjuntamente com a visão de comunidades em redes: a estrutura de clusters (muitos nós ou nós muito conectados na rede). A partir dessas idéias e da observação de campo dos fotologs, é proposta a tipologia de comunidades virtuais. Os tipos propostos constituem-se em comunidades emergentes, ou seja, aquelas formadas através da interação social mútua (PRIMO, 1998 e 2003), compostas por laços relacionais (BREIGER, 1974), capital social diversificado (BERTOLINI E BRAVO, 2004) e trocas comunicacionais; comunidades associativas, ou seja, aquelas formadas através da interação social reativa (PRIMO, 1998 e 2003), compostas por laços mais associativos e capital social diversificado (BERTOLINI E BRAVO, 2004); e comunidades híbridas, que são aquelas que possuem características de ambos os tipos anteriores. Essa tipologia é, então, analisada a partir de sua estrutura, dinâmica e composição, a partir das ferramentas da abordagem de redes sociais, onde demonstra-se que há diferenças entre os três tipos e que cada um possui elementos diferentes associados. / The following work proposes a typology for virtual communities in social networks based on personal webpages for pictures which can allow other users to comment them (fotologs), hosted at Fotolog.com. The concept of virtual community adopted in this thesis comes from an analysis of utopian concepts (TÖNIES, 1947 e 1995; DURKHEIM, 1978; BUBER, 1987) and contemporary concepts (WEBER, 1987; WELLMAN, 2001 e 2002; MAFFESOLI, 1996 e 2000) of community and from the similarities and common elements from the two views: social interaction, social ties and social capital. Our concept is also based in the network idea of community: the cluster structure (several nodes or densely connected nodes). From these elements and field observation, we propose types of virtual communities: emergent communities, which are based on mutual social interaction (PRIMO, 1998 e 2003) and composed of relational ties (BREIGER, 1974), diversified social capital (BERTOLINI E BRAVO, 2004) and communication exchanges; associative communities which are based on reactive social interaction (PRIMO, 1998 e 2003), associative ties and also diversified social capital (BERTOLINI E BRAVO, 2004); and hybrid communities, which are those with characteristics both from the emergent and the associative types. This typology is then analyzed on each structure type, composition and dynamics, both from social network analysis tools and empirical observation, where we show that they have similarities and dissimilarities.