Cette thèse est composée de trois articles qui posent chacun une question sur les discriminations. Le premier article interroge si les habitants de la Seine-Saint-Denis sont statistiquement discriminés sur une violence supposée ; où la pratique d’un sport de combat serait un signal de violence. Le second article regarde si les accents marseillais sont discriminés sur le marché du travail. Le dernier article se demande si les employeurs pourraient utiliser le patronyme et l’adresse des candidats à l’embauche pour faire des corrélations sur leur accent et, s’ils infèrent qu’ils ont l’accent des grands ensembles, de les discriminer sur ce point. La méthode du « testing » est utilisée pour répondre à ces trois questions. Les deux premiers testings amènent à répondre par la négative aux deux premières questions. Concernant la troisième question, même si l’on ne peut pas y répondre, le troisième testing prouve que l’accent des grands ensembles est discriminé. De plus, il démontre que la discrimination sur le patronyme peut mettre dans l’ombre celle sur l’adresse. De façon générale, ce travail permet de mettre en évidence que les employeurs ne prennent pas en compte le sport dans leur pratique d’embauche et que lors de l’élaboration du design d’un testing on doit distinguer les signaux forts des signaux faibles. Il montre aussi que les testings peuvent être utilisés jusqu’au moment où les employeurs contactent les candidats par téléphone et qu’ils engagent des stéréotypes différents selon l’accent qu’ils entendent. Enfin, il propose aux décideurs publics de chercher à d’abord diminuer la discrimination basée sur le patronyme s’ils veulent diminuer la discrimination totale. / This thesis is composed of three articles. Each of them asks one question about discriminations. The first article interrogates if Seine-Saint-Denis inhabitants are statistically discriminated on a supposed violence ; where to do a fighting sport would be a violence signal. The second article investigates if Marseille accents are discriminated in the labor market. The third article asks if employers may use the name and the address as proxies to determine job candidates’ accent and, if they infer he holds a poor area accent, to discriminate on the basis of this accent. I use the correspondence study method to reply to these three questions. The first two correspondence studies lead to conclude to the negative to the first two questions. Concerning the third question, even if we cannot reply to it, the third correspondence study proves that the poor area accent is discriminated. Moreover, it shows that the name discrimination may overcomes the address one.General conclusions are that this thesis indicates employers do not take into account the practiced sport to hire and that to elaborate the design of a correspondence study implies to consider weak and strong signals. It also shows that correspondence studies can be used up to the moment where employers contact job applicants by phone and that they have different stereotypes in mind according to the accent they hear. Finally, it proposes to public policy makers to first decrease the name discrimination if they want to significantly impact the total discrimination.

潛在群體分析及對應分析是在探討多個類別變數間關係常用的兩種分析方法，因其運用之領域不同，解說之方式不一，以致過去常被認為是兩種不相關的分析方法，然而實際卻非如此。本研究之目的係就雙變數及多變數的情況，針對這兩種分析方法間之關係作詳細的探討，並說明其對等及不對等的時機。 / Latent class analysis and correspondence analysis are two well-known methods that can be used to study the relationship between categorical variables. Since the two were developed and applied in two different fields in the past, they were never thought to be related. In this study, we examine the relationship between the two in more details. We further point out the situations where they are equivalent, and where they are not.

<p>Enkäter används flitigt idag. Spector (1994) menade att endast en liten del av den förklarade variansen i enkäter utgörs av det som önskas mäta. Bryggan mellan respondenten och enkäten är språket. Enligt Whorf-Safir hypotesen styrs våra tankar av språket. Semin och Fiedler (1988) menade att ordklasserna har olika abstraktionsnivåer. Desto högre abstraktionsnivå ju mer rör den sig bort från kontexten resulterande i stereotyptänkande (Gilbert & Malone, 1995). Denna studie undersökte hur olika ordklasser påverkar svar. Huvudstudien var utformad som ett personlighetstest med påståenden. Resultatet uppvisade en skillnad mellan svaren då påståendena innehöll substantiv eller adjektiv. Respondenterna tenderade att instämma mindre då ordet var ett substantiv, något som förklarades med correspondence bias. Resultatet bekräftade tidigare studier. Replikering med engelsk enkät efterlystes.</p>

We describe a method for modeling object classes (such as faces) using 2D example images and an algorithm for matching a model to a novel image. The object class models are "learned'' from example images that we call prototypes. In addition to the images, the pixelwise correspondences between a reference prototype and each of the other prototypes must also be provided. Thus a model consists of a linear combination of prototypical shapes and textures. A stochastic gradient descent algorithm is used to match a model to a novel image by minimizing the error between the model and the novel image. Example models are shown as well as example matches to novel images. The robustness of the matching algorithm is also evaluated. The technique can be used for a number of applications including the computation of correspondence between novel images of a certain known class, object recognition, image synthesis and image compression.

The aim of this dissertation is to study segregation within the Swedish school system, with regard to the reformation during the last decades of the 20th century which researchers have described as a system change. After an initial research survey two questions are identified: one which concerns the process of human becoming, and the other concerning segregation, within and between social milieus in a school system. The first is answered from a relationistic philosophical perspective, which leads to the study object the educative milieu. In turn the study object leads to the development of a research approach to investigate such settings. The research approach consists in the construction of epistemic objects (in short: epiobjects) – in this case two school classes whose pupils have actively made the choice of these classes – understood as monistic entities constituted of internal relational structures. The construction of the epiobjects identifies the families involved in the classes as informants and decides on a number of research themes (for example description of the school, the teachers, leisure activities) for the data collection. The disposition of the study consists of tree main steps: determining relevant symbols, positioning and description. Determining relevant symbols means to decide what is assigned meaning in respect of the research themes. This was achieved through an interview study. Positioning means to decide how the symbols and bodies are distributed within the epiobjects in order to represent their relational structures. This was done with questionnaire data that was organised with aid of correspondence analysis. In the description the epiobjects can be understood as educative milieus, by focusing on their educative tensions and doxa. Educative tensions are differences between positions in the epiobjects relational structure, while doxa are areas of agreement. The analysis shows that the epiobjects are obviously different in terms of educative tensions and doxa. The main difference between the epiobjects concern the families’ motives for the choice they have made: families in one class maintained they chose the school because of its subject specialisation, while families in the other claimed that they sought a school characterised by orderly behaviour and a better learning of subject knowledge and skills. Associated with these differences are dissimilarities concerning, for instance, political orientation, education level, income, democracy, freedom of choice, as well as regarding inclination to attend different kinds of cultural events and institutions. The results can be seen as an illustration of how a system of freedom of choice leads to segregation, by orienting people with dissimilar preferences to different alternatives while, at the same time, these preferences concerning school choice correspond with differences in other areas of life.

Enkäter används flitigt idag. Spector (1994) menade att endast en liten del av den förklarade variansen i enkäter utgörs av det som önskas mäta. Bryggan mellan respondenten och enkäten är språket. Enligt Whorf-Safir hypotesen styrs våra tankar av språket. Semin och Fiedler (1988) menade att ordklasserna har olika abstraktionsnivåer. Desto högre abstraktionsnivå ju mer rör den sig bort från kontexten resulterande i stereotyptänkande (Gilbert &amp; Malone, 1995). Denna studie undersökte hur olika ordklasser påverkar svar. Huvudstudien var utformad som ett personlighetstest med påståenden. Resultatet uppvisade en skillnad mellan svaren då påståendena innehöll substantiv eller adjektiv. Respondenterna tenderade att instämma mindre då ordet var ett substantiv, något som förklarades med correspondence bias. Resultatet bekräftade tidigare studier. Replikering med engelsk enkät efterlystes.

Nel 1969, il poeta, scrittore, giornalista e pittore Enzo Benedetto, calabrese di nascita, romano d'adozione, la cui adesione al Movimento Futurista risale al 1923, fondava nella capitale la rivista «Futurismo-oggi», assumendone la direzione fino al 1993, anno della sua morte e della cessazione delle pubblicazioni. Lo studio si concentra sulla ricca e selezionata corrispondenza epistolare, relativa al periodo 1969-1992, tra il Direttore e numerosi collaboratori del periodico laziale che, nella seconda metà del Novecento, si contraddistinse quale organo di difesa, di propaganda e di valorizzazione dell'avanguardia marinettiana. Al corpus documentario seguono i profili bio-bibliografici dei corrispondenti e l'indice dei nomi. Si presenta infine lo spoglio integrale di «Futurismo-oggi» uniformato alle norme del progetto di ricerca Iride 900. / The poet, author, journalist and painter Enzo Benedetto, was born in Calabria with Rome as his adopted city. Since 1923, he belonged to the Futurist Movement. In 1969, he founded, in Rome, the review «Futurismo-oggi» providing its direction until 1993 the year of his death. The research concentrates on the copious and selected correspondence concerning the period 1969-1992, between the Director and several collaborators of the roman periodical. This review has represented in the second half of the twentieth century a defence, propaganda and exploitation of the Marinetti's avant-garde. Biographic and bibliographic outlines of the correspondents and the list of the names follow the index of the documents. At the end the whole filing of «Futurismo-oggi» is of the equivalent standard of the research-plan Iride 900.

Il presente contributo si propone di ricostruire il sodalizio Fra valentino bompiani e libero bigiaretti, sulla base del carteggio inedito, desunto dall'archivio personale dell'editore e dall'archivio della casa editrice. Nella personlaità dei due protagonisti, e nelle ragioni insite nella rispettiva opera intellettuale, trova adeguata valorizzazione lo scambio umano e professionale intercorso fra l'autore e il suo editore, a partire dai primissimi anni sessanta, in un periodo di importanti cambiamenti dal punto di vista dell'assetto aziendale, che culminerà nel 1973 con la vendita del marchio. Si deve all'indiscusso "protagonismo" dell'editore la peculiarietà del legame con lo scrittore marchigiano (all'amico Valentino questi affidò la ristampa di molte sue opere passate, nonché l'edizione di quelle future), il quale aggiunge tasselli importanti alla ricostruzione di numerosi aspetti: in particolare, trapelano dalle lettere, fra gioie e delusioni, umori e confessioni che bene motivano l'evoluzione creativa e interiore di Bigiaretti nella fase più matura della propria esperienza narrativa, con esiti nuovi soprattutto dal punto di vista espressivo. I numerosi riferimenti, nel carteggio, alla scrittura e alla poetica Bigiarettiana - in un continuo confronto con Bompiani, fido confidente e affascinato lettore - trovano rispondenza nelle stesure autografe dei romanzi in oggetto, comprovando, al contempo, il faticoso labora limae dell'autore. / The aim of this work is to rebuild the fellowship between Valentino Bompiani and Libero Bigiaretti on the basis of the unpublished correspondence found in publisher's private archives and publishing house's ones. In the two protagonists' personality, and in the reasons implied in their intellectual work, there is a fair appreciation of the human and professional relationship passed between the author and his publisher starting from the early 70's, a time of important changes in the company structure, which ends with the sell of the brand name in 1973. It's due to the publisher's undisputed desire to be the centre of attention the peculiarity of the relationship with the author form the Marche (to his friend Valentino he entrusted the reprint of his past works and also the print of the future ones), a relationship that adds important particulars to the reconstruction of many aspects; the letters show especially, among joys and disillusions, moods and confessions that explain, in a clear way, Bigiaretti's creative and inner development in the more mature phase oh his narrative experience, that led to new results above all from the expressive point of view. The large number of references, in the exchange of letters, to Bigiaretti's writing and poetics – in a continuous comparison with Bompiani, faithful confidant and charmed reader – find a correspondence in the drafts (in Bigiaretti's own hand) of the analyzed novels, proving, at the same time, the author's strenuous labour limae.

The purpose of this dissertation was to examine the institutional history of Clark Wissler's professional career as an anthropologist and to determine his influence upon American social science in this context. By focusing on specific historical contexts in which Wissler affected social science research in America, the study attempted to show the extent of Wissler's influence and impact on the development of social science. As well, the study considered and offered an explanation of how Wissler became a relatively obscure figure in the history of American anthropology after a period of considerable impact on the discipline. Primary data for this study were several pieces of correspondence and personal papers in the collection of "Wissler Papers" at the Department of Anthropology, Ball State University, Muncie, Indiana.This study revealed that Clark Wissler occupied a unique position among American social scientists. That position Was unique in terms of Wissler's activities in a strictly institutional context as "Curator" of the American Museum of Natural History's Department of Anthropology; but more so, the position at the museum (one of the regional centers of American anthropology until the Second World War) led directly and indirectly to Wissler's influence upon social science research through ancillary positions with research foundations and institutes during the formative years of modern social science. What the study indicated, in this context, was that personal relationships often influenced ties between individuals in an institutional framework. Also, in the development of anthropology as an academic discipline in America, the more extensive that one's institutional network was -- in terms of personal and/or institutional ties, the more impact one had on ideological constructs and research trends.Furthermore, the study indicated that the extent and duration (from one generation of students to another) of a figure's impact on a discipline was dependent upon three factors -- politics, polemics, and progeny -- which were postulated as critical determinants of influence. That is, the study suggested and posited that influential figures in American anthropology were determined in a three-fold context: those one patronizes and is patronized by (politics); those trend-setters and organization officials that one agrees with and/or is thought of in association with (polemics); and those one proselytizes and converts to one's frame of reference, and thusly, who become disciples (progeny). All three contexts are in terms of personal relations that develop into institutional structures and functions, and thereby, determine one's influence and stature in an academic discipline.Thus, the study concluded that: 1) more than "ideas" are involved in the history of a social science discipline, namely anthropology; 2) Wissler, with an extensive institutional network but virtually no "progeny," was very influential among social scientists during his professional career, but he became a forgotten figure within twenty years of his death; 3) influence, in terms of historical "facts," must be determined in a. situational context that does not remove personalities and concrete personal relations from a holistic view of a specific cultural milieu.Extensive appendices to the dissertation provide primary data for further study in the history of anthropology, as well as support for contentions in the dissertation. As such, the dissertation, in itself, serves as a basis for further research.

The purpose of the study was to identify problems associated with unaccredited private schools and home instruction programs and possible solutions to problems as perceived by State Education Officials in the Great Lakes States. The States were Illinois, Indiana, Michigan, Minnesota, New York, Ohio, Pennsylvania and Wisconsin.Eleven State Education Officials having responsibility for unaccredited private schools and home instruction programs were interviewed by telephone to collect the data. An interview form was designed to elicit non-directed responses to major problems confronting the states, methods of handling problems, possible solutions to problems, and practices in the next five years.Major problems existed with state regulations regarding equivalent instruction, teacher qualifications, courses of study, and definition of a school. Lack of criteria for interpreting the terms and assuring compliance with requirements were reported. Local school superintendents were confronted with similar problems regarding state regulations. State Education Officials identified problems with lack of certified or qualified teachers in fundamentalist schools and home instruction programs, difficulty obtaining data from pastors and parents, and extent of state authority to regulate private schools. Also, student loss of services, materials, and diplomas was a concern.Solutions to the problems as perceived by State Education Officials included changes in statutes from compulsory school attendance to compulsory education with minimum standards and achievement testing, clarification of terms in state statutes, and separate state regulations for private education. Other recommendations were a separate accrediting agency, a national co-op agency, and communication and cooperation between public and private educators.State Education Officials projected no substantial changes in regulations or practices in the next five years. However, expanding interest and increasing enrollments in private schools were expected.