//img.uscri.be/pth/68455e4ccbf73e618ce517fa6d3a3b28c363d710
Cette publication ne fait pas partie de la bibliothèque YouScribe
Elle est disponible uniquement à l'achat (la librairie de YouScribe)
Achetez pour : 9,00 € Lire un extrait

Lecture en ligne (cet ouvrage ne se télécharge pas)

Réflexions sur la crise : éco-socialisme ou barbarie

De
110 pages
Il s'agit de réfléchir sur l'instauration d'un véritable ordre économique mondial nouveau, plus juste, reliant le projet du dépassement du capitalisme à celui d'une révolution écologique et démocratique, mettant fin à la destruction de l'environnement, au productivisme, à la pauvreté, aux inégalités, à la dictature de l'argent et au pillage du Tiers-Monde.
Voir plus Voir moins

Bulletin

de Méthodologie

Sociologique

(48 bd. Jourdan, F 75014 Paris), October 2009, N. 104 - p. 1

CONTENTS/SOMMAIRE
Karl M. van Meter Mathilde de Saint Léger German & French Contemporary Sociology Compared: Text Analysis of Congress Abstracts en cours Pessimisme et optimisme dans la mesure de la liaison d'un tableau de contingence: Réponse à Philippe Cibois ..32 Journée d'Etude du 21 août 2009 sur la méthodologie « Alceste» Arguments des interventions AAPOR 2009 Conference "Public Choices in Changing Times" Fall RC33 Newsletter

.5

Ongoing Research/Recherche Stéphane Champely Brice Lefèvre

Max Reinert

.39

Nick Moon John Goyder Centre

47 .54 .61 .63 68 73 79 .83 93 98

- 2009

Centres/Projects BookslLivres.. JournalslReviewslReports Articles Computers/OrdinateurslInternet New MeetingslNouvelles réunions Past MeetingslRéunions Calls/Appels passées

N. 104 OCTOBER 2009

BuLLelifl de MélllOdologie

Sociologique

(48 bd. Jourdan, F 75014 Paris), October 2009, N. 104

- p. 2

BMS - AIMS
The BMS is a peer reviewed trimestrial scientific journal published Methodology, 45 rue Linné, 75005 Paris), a non profit organization. correspondence aims@orange.fr should be sent to the BMS, CMH, 48 boulevard by the AIMS (International The BMSs parity number Association of Sociological is 0112 G 88707. All

Jourdan,

75014 Paris; tel/fax 33 1 40 51 85 19; email bms,

The publishers of the BMS are. Philippe Cibois (Université Paris), Lise Mounier (CMH-CNRS, Paris) and Marie-Ange Meter.

dc Vcrsailles-Saint Quentin), Karl M. van Meter (CMH-CNRS. Schiltz (CAMS-EHESS, Paris). The director is Karl M. van

The Scientific Committee of the BMS is composed of Duane F. Alwin (University of Michigan), Alain Degenne CNRS, Caen), Peter Ph. Mohler (ZUMA, Mannheim) and Henry Rouanett (Université Paris V). The BMS publishes twice a year the Newsiefler of Research Committee (RC33) "Logic and Methodology" Sociological Association. The BMS is abstracted by the principal institutions concerned with sociological

(LASMAS-

of the International methodology: the

lNlST (Institut de l'Information Scientifique et Technique) of the CNRS in Nancy; and Sociological Abstracts in San Diego, California, which classes the BMS among "core sociology journals". The BMS is a member of the Assocation des Revues Scientifiques et Culturelles (ARSe) <http://www.arse.be>. Recent articles of the BMS can be purchased as pdffiles from E13SCO. With Revues.org, the BMS is preparing to put all older issues on the Web as free, fully-searchable documents.

The BMS publishes in both English and French. The AIMS reserves all rights to translation, adaptation or reproduction in any form of all material puhlished by the BMS. The BMS also maintains an Internet listserv open to its subscribers and a free Internet mailing list for interested readers.

To submit an article for peer review and possible publication in thc BMS, send one hardcopy and one WORD copy by email. The article should includc title, author, contact infonnation (post, tcl, fax, email, weh), a short one-paragraph abstract with key words, footnotes (instead of endnotes), and references. Tables and graphics should be included as modifiable images. Further instructions the BMS. for authors is available at our web site <http://www.cmh.pro.ens.fr/bms/Welcome.htm>. or by contacting

-------------------BMS - AIMS
Le BMS est une revue scientifique trimestrielle à comité de lecteurs éditée par l'AIMS (Association luternationale de Méthodologie Sociologique, 45 rue Linné, 75005 Paris), une organisation sans but lucratif (loi 190 I). Le BMS a le numéro paritaire 0112 G 88707. Toute correspondance doit être envoyée au BMS, CMH, 48 boulevard Jourdan, 75014 Paris; tél/fax 33 1 4051 85 19; courrier électronique bms-aims@orange.fr Quentin), Karl M. van Meter Paris). Le responsable de la

Le comité de rédaction du BMS est composé de: Philippe Cibois (Université de Versailles-Saint (CMH-CNRS, Paris), Lise Mounier (CMH-CNRS, Paris) et Marie-Ange Schiltz (CAMS-EHESS, publication est Karl M. van Meter.

Le comité de conseil scientifique du BMS est compose de: Duane F. Alwin (University of Michigan), Alain Degenne (LASMAS-CNRS, Caen), Peter Ph. Mahler (ZUMA, Mannheim) et Henry Rouanett (Université Paris V). Le BMS publie deux fois par an la Newsletter du Comite de recherche (RC33) "Logique et Méthodologie" de l'Association Internationale de Sociologie. Le BIviS est analyse par les grands établissements qui s'occupent de la méthodologie sociologique: l'INlST (Institut de l'Information Scientifique et Technique) du CNRS à Nancy; et Sociological Abstracts à San Diego aux Etats-Unis, qui classe le BMS parmi les "journaux clefs de la sociologie" Le EMS est membre de l'Assocation des Revues fichiers Scientifiques et Culturelles (ARSe) <http://www.arse.be>. Des articles récents du BMS peuvent être achetés comme pdf à EBSCO. Avec Revues.org, le BMS est en train de metter gratuitement sur le Web tous les plus anciens numéros

en format reeherchable. Le BMS publie en français et aussi en anglais. L'AlMS se réserve tous droits de traduction, d'adaptation et de reproduction toute matiere publiée dans le BIviS. Le BMS gère sur Internet un listserv ouverte à ses abonnés et une liste de distribution gratuitc ouverte à tout lecteur intéressé. de

Pour soumettre un article au BMS, envoyez un exemplaire sur papier et une copie WaRD par email. L'article doit comprendre le titre, l'auteur, ses coordonnées (poste, tel, fax, email, web), un court résumé d'un paragraphe avec mots-clefs, des notes en bas de page, et des références. Les tableaux et graphiques doivent ètre en format image adaptable. Plus d'inlormation est disponible sur notre site web <hnp:l/www.cmh.pmens.frlbms/Welcome.htm>. ou en contactant le BMS.

----------------------------------------

Bulletin de Méthodologie

Sociologique

(48 bd. Jourdan, F 75014 Paris), October 2009, N. 104 - p. 3

EDITORIAL

This is the last issue of the BMS to be published in Paris by our non-profit Association Internationale de Méthodologie Sociologique (AIMS), or International Association of Sociological Methodology. The next issue in January 2010 will be published by Sage Publications in London and distributed directly from there. Although we are expanding the Scientific Committee, the Editorial Board remains the same. This issue of the BMS includes the Fall 2009 issue of the RC33 Newsletter, a research article in English, two Ongoing Research notes in French and one in English. The research article, "German & French Contemporary Sociology Compared: Text Analysis of Congress Abstracts", by Karl M. van Meter and Mathilde de Saint Léger, reviews results of the Calliope co-word analysis of all paper abstracts for the 2004, 2006 and 2009 congresses of the Association Française de Sociologie (AFS), and compares these results with those of a similar analysis of the 2004 Deutsche Gesellschaflfür Soziologie (DGS) congress. The first Ongoing Research note, "Pessimism and Optimism in the Measure of

Association for Contingency Tables - A Reply to Philippe Cibois", by Stéphane
Champely and Brice Lefèvre, examines Cibois' suggestion to use the generalised PEM (percentage of maximum deviation from independence) instead of Cramér's classical Y to study the strength of the relationship in a contingency table. The PKM (percentage of maximum chi-squared of Cibois, 1990) seems more directly comparable to Cramér's y2, and a new algorithm is proposed to optimise the calculation ofthis PKM (and of the generalised PEM). The second Ongoing Research note, "The 21 August 2009 Workshop on 'Alceste' Methodology - The Presentations' Arguments", by Max Reinert, reports on the Alceste Summer School in Carcassonne on 19-21 August, during which ten presentations were made and discussed by the participants. The third Ongoing Research note, "AAPOR 2009 Conference
-

Changing Times"', by Nick Moon and John Goyder, presents the sixty-fourth annual AAPOR Conference, which took place on 14-17 May 2009 in Hollywood, Florida, near Miami. With more than 400 papers presented and many important awards, it was major survey research meeting. ----------------------------------------

'Public Choices in

BlIlIelin de Mél/lOd%gie

Soci%g/lflle

(48 bd. Jourdan, F 75014 Paris), October 2009, N. 104 - p. 4

EDITORIAL

Ce numéro est le dernier du BMS qui sera publié à Paris par notre association 190l, l'Association Internationale de Méthodologie Sociologique (AIMS). Le prochain numéro de janvier 2010 sera publiée par Sage Publications à Londres et sera distribué directement de là-bas. D'autre part, nous élargissons le comité scientifique, mais le comité de rédaction reste la même. Ce numéro du BMS comprend le numéro d'automne 2009 numéro de la lettre d'information du RC33, un article de recherche en anglais, deux notes de Recherche en cours en français et une en anglais. L'article de recherche, « Les

sociologiesallemande et française contemporainescomparées - Analyse textuelle
des résumés de congrès », par Karl M. van Meter et Mathilde de Saint Léger, passe en revue les résultats de l'analyse par mots associés avec le logiciel Calliope de tous les résumés de présentations faites aux congrès 2004, 2006 et 2009 de l'Association Française de Sociologie (AFS), et compare ces résultats avec ceux d'une analyse similaire du congrès 2004 de la Deutsche Gesellschaft fiir Soziologie (DGS). La première note de Recherche en cours, « Pessimisme et optimisme dans la mesure de la liaison d'un tableau de contingence: Réponse à Philippe Cibois », par Stéphane Champely et Brice Lefèvre, examine la suggestion de Cibois concernant la mesure de la liaison globale d'un tableau de contingence en utilisant le PEM (pourcentage d'écart maximum) généralisé à la place du classique Y de Cramér. Mais le PKM (pourcentage du khi-carré maximum de Cibois, 1990) semble plus directement comparable au y2 de Cramér, et un nouvel algorithme optimisant le calcul du PKM (et du PEM généralisé) aboutit à des résultats finalement peu différents de ceux obtenus avec la statistique de Cramér La deuxième note de Recherche en cours, « Journée d'Etude du 21 août 2009 sur la méthodologie' Alceste' - Arguments des interventions », par Max Reinert, est un rapport sur cette journée au cours de laquelle dix exposés ont été présentés et discutés avec le participants. Le troisième Note de recherche en cours, « La conférence 2009 de l'AAPOR, 'Choix publics en temps de changement''', par Nick Moon et John Goyder, présente le soixante-quatrième conférence annuelle de l'AAPOR qui a eu lieu du 14 au 17 mai 2009, à Hollywood, en Floride, près de Miami. Avec plus de 400 présentations et de nombreux prix importants, c'était une réunion majeure de la recherche par enquête. ----------------------------------------

Blllletin de Méthodologie

Sociologiqlle

(48 bd. .lourdan, F 75014 Paris), October 2009, N. 104 - pp. 5-31

GERMAN & FRENCH CONTEMPORARY SOCIOLOGY COMPARED: TEXT ANALYSIS OF CONGRESS ABSTRACTS
by Karl M, van Meter (CMH, CNRS-ENS-EHESS Paris; karLvanmeter@ens,fr) Mathilde de Saint Léger (IPAM-CNRS Paris; mathilde,de-saint-Ieger@cnrs-didr)

Résumé - Les sociologies allemande et française contemporaines comparées - Analyse textuelle des résumés de congrès: Nous passons en revue les résultats de l'analyse par mots associés avec le logiciel Calliope de tous les résumés de présentations faites aux congrès 2004, 2006 et 2009 de l'Association Française de Sociologie (AFS) et nous comparons ces résultats avec ceux d'une analyse similaire du congrès 2004 de la Deulsche Gesellschaft für Soziologie (DGS). Sociologie allemande, Sociologie française, Analyse par mots associés, Calliope, Abstract: We review results of the Calliope co-word analysis of all paper abstracts for the 2004, 2006 and 2009 congresses of the Associalion Française de Sociologie (AFS) and compare these results with those of a similar analysis of the 2004 Deutsche Gesellschaftfür Soziologie (DGS) congress, German Sociology, French Sociology, Co-word Analysis, Calliope.

INTRODUCTION

Computer programs for co-occurrence of key word analysis of scientific textual data have been presented previously in the EMS (Callan, Courtial and Turner, 1991; De Saint Léger and Van Meter, 2009), In particular, the Calliope program has been presented and applied to the analysis of abstracts of the Association Française de Sociologie (AFS) congresses in Saint Denis-Villeteneuse in 2004 (De Saint Léger and Van Meter, 2005)1, in Bordeaux in 2006 (Van Meter and De Saint Léger, 2007), and in Paris in 2009 (De Saint Léger and Van Meter, 2009). But one of the major results of the comparison oftext analysis ofthese congress abstracts does not
] See also Demazière et al. (2006) which develops further these results and the use of Calliope, particularly in the fourth chapter (in French), "The Contribution of Scientometrics and the Associated-Words Method" (De Saint Léger and Van Meter, 2006: 1]9-146).

p. 6

- Bulletin

de Méthodologie

Sociologique

(48 bd. Jourdan, F 750]4 Paris), October 2009, N. 104

require a sophisticated multivariate method of analysis such as Calliope (which is described in some detail below) to see the relative weight of the key words Travail (Work) and Femme (Woman).2 At these three AFS congresses, we have seen the dominant and almost interchangeable roles played by both Femme and Travail (see respectively Figure 1 and Figure 3 below). Since neither Travail, nor Femme were never the central themes of any of the three AFS congresses, one can safely assume that they are a legitimate reflection of the structure of CUITentFrench sociology. Indeed, we have previously mentioned the fact that Travail was a far more structuring term than Femme (De Saint Léger and Van Meter, 2005: 50-51), and probably related to the historical influence of "Sociologie du Travail" in the development of French sociology. The term Femme and the "sociologie du genre" are more recent arrivals on the scene, and this is reflected in their position relative to Travail and "sociologie du travail".

THE CALLIOPE METHOD OF TEXT ANALYSIS Calliope is the Windows version of the Leximappe-Lexinet system of text analysis by co-occurrence ofkey words in units oftext. We have presented both LeximappeLexinet and Calliope in detail in previous publications, but for those unfamiliar with the method it can be easily described as a hierarchically ascending classification method based on the co-occurrence of key words selected by the researcher in the units of texts to be analyzed. The classes or clusters are constructed using CO-OCCUTI'enceskey words as the measure of similarity between of the key words. Thus, key words that appear often together in the different units of text - here, in abstracts of scientific presentations made at a conference - will have a high index of similarity and will be placed together in the same class. The classes thus constructed are then used as the basis for the calculation of "inties" (ties between key words within the same class) and "out-ties" (ties by key words outside of their class and with key words in other classes). The "in-ties" are used to calculate an axis of "density" and the "out-ties" to calculate an axis of "centrality". With density as a horizontal axis, and centrality as a vertical axis, a two-dimensional "Strategic Diagram" is constructed on which it is possible to place the different classes according their density and centrality. The Strategic Diagram, by construction, therefore has a dense and central first quadrant, which we call the "mainstream". The dense but not central second quadrant is called "ivory tower". The not dense and not central third quadrant is sometimes called "chaos", or simply "unstructured". The central but not dense fourth quadrant is called "bandwagon". See Callon, Courtial and Turner, 1991 This means that "mainstream" classes (in the first quadrant) will have relatively numerous "in-ties" (co-occurrences of the class' key words together in the text
2

The wieght of of each word is calculated according to its co-occurences, frequency in the data set oftexts, on the other (De Saint Léger, ] 997).

on one hand, and its

6

Bulletin

de Méthodologie

Sociologique

(48 bd. Jourdan, F 75014 Paris). October 2009, N. 104 - p. 7

units) and relatively numerous "out-ties" (co-occurrences in text units of the class' key words with other key words from other classes). "Bandwagon" classes have key words with numerous "out-ties" but relatively fewer "in-ties". And, of course, "ivory tower" classes have numerous "in-ties" but relatively few "out-ties".

ASSOCIATION FRANCAISE DE SOCIOLOGIE

(AFS) 2004 CONGRESS

The first congress of the French Association of Sociology (AFS) in February 2004 included over 1,000 different presentations and 3,000 participants. Since it was difficult or impossible to attend all sessions or talk with all authors, the idea soon came up in the "Méthodes" Thematic Network (RT20) that we should do a "mapping" of the congress with textual data analysis methods to represent the structure and diversity of themes covered during the meeting (De Saint Léger and Van Meter, 2005). In talking with Philip Cibois, one of the founders of the section "Méthodes", this idea evolved and became clearer. We agreed to map the abstracts of presentations as they were published in the conference program book and distributed to all participants.3 The comparison and confrontation of results of various methods of textual analysis would be necessary for any such research project organized by the "Méthodes" section. The methods chosen were: Philip Cibois' Trideux factor analysis of correspondence (Cibois, 1984, 1990, 1995), the analysis of cooccurrence of key words by Calliope, Max Reinert's Alceste hierarchically descending classification of text units and key words (Reinert, 1987, 1990), and Francis Chateauraynaud's Prospero text-based analysis of controversies (Chateauraynaud, 2003; Chateauraynaud, Bernard and Van Meter, 2003). This cooperation resulted in the publication of the BMS, n. 85, January 2005, with the results of the analysis of the 2004 AFS abstracts by these four methods. These results and the articles published in the BMS later served as the basis for the book Analyses textuelles en sociologie - Logiciels, méthodes, usages [Textual analysis in sociology - Software, methods, uses] (Demazière et al., 2006), Here, we will concentrate on the results produced by Calliope and, in particular, the Strategic Diagram of all AFS 2004 abstracts (Figure 1). It includes 40 clusters, but their distribution throughout the quadrants is not uniform with quadrants 1 and 3 together containing 25 of the 40 clusters or 62%, The quadrant 3 clusters are of low centrality and density and are thematic clusters which are either emerging (they could become more structured and consistent), or are older in the sense that they are losing ground in the mapped domain. Conversely, quadrant 1 contains clusters that are structuring the domain and are themselves structured. Given the distribution in Figure 1, one is confronted with a rather dynamic domain and a plethora of ideas, and not a partitioned and aging field. The cluster Femme is by far the most central (centrality 1,134), with in descending order, Emploi (Employment) (centrality = 793), and then Sociologie and
3 This data set remains available tree of charge on the AFS Web site: http://www.afs-socio.fr/ 7

p. 8 . Blllletin

de Méthodologie

Sociologiqlle

(48 bd. Jourdan, F 75014 Paris), October 2009, N. 104

Cellule Familliale (Family Cell) (centrality = 753). However, internai links or "inties" of the Femme cluster are rather weak, the majority of them don't even reach the 150 threshold (see Figure 2 below) for a maximum of I ,000. Thus, the density of the cluster is rather low, yet it includes over 590 abstracts out of the total of 1,036. One wonders how these 590 abstracts are organized among themselves if the internal ties of the Femme cluster are so weak.
Figure 1: AFS 2004 Strategic Diagram of All Abstracts

--'-'-wnOS;lTi41

To answer that question and analyze the content of the cluster Femme, we took the sub-population or sub-corpus of 590 abstracts of the cluster and we remapped them with Calliope, having previously removed the term Femme to allow the other key words to organize new clusters within the sub-population (De Saint Léger and Van Meter, 2005). The new Strategic Diagram generated by this analysis appears in Figure 3. One should note that the form and structure of the new Strategic Diagram are rather similar to Figure 1, the Strategic Diagram for the entire data set of AFS 2004 abstracts, with a similar distribution of clusters concentrated in quadrants I and 3. The cluster Travail stands out became it has become the backbone of the structure with high centrality and high density. Its contents include key words most 8

Bulletin

de Mêl1lOdologie Sociologique

(48 bd. Jourdan, F 75014 Paris), October 2009, N. 104. p. 9

Figure 2: AFS 2004 Femme Cluster

Calliope

Viewe,

characteristic of Travail and not Femme (see Figure 4 below). The other keys words previously associated with Femme have beell distributed into four subgroups illFignre 3; Couple, Espace Mas(:ulîn (Male Domain), Cellule Familliale (Family Cell) and Domain de la Sexualité (Sexuality Domain) (see figures below). They are all situated in quadrant l, of course, but with centrality and density lower than the cluster Travail. In summary, although Femme was the most central cluster and included many of the abstracts of the AFS 2004 data set, it was the key word and cluster Travail that were both central and dense, and were structuring the entire corpus of abstracts. Indeed, the similarities between Figure l and Figure 3 show that theentire corpus and the Femme sub.corpus can almost be substituted one for the other.

9

p. JO. Bulletin

de Méthodologie

Sociologique(48

bd. Jourdan, F 75014 Paris), October 2009, N. 104

Figure 3: AFS 2004 -Sub-population

Femme Reanalyzed

Figure 4: AFS 2004 - Travail (Work) population Femme
~fW.i~r.1~~(,","1ITI!!!YAlIJ'"

Cluster

in the Reanalyzed

Sub-

""""''11_,

10

Bulletill de Méthodologie

Sociologique

(48 bd. Jourdan, F 75014 Paris), October 2009, N. 104 - p. 1 J

ASSOCIATION

FRANCAISE DE SOCIOLOGIE

(AFS) 2006 CONGRESS

One of our major questions in analyzing the 2006 AFS abstracts was to see if the complex intertwined relationship involving Femme and Travail continued to playa major role in structuring this new corpus. The answer is "yes", but in an even more complex and intertwined relationship that is almost the mirror image of the 2004 AFS abstracts (Van Meter and De Saint Léger, 2007). See Figure 5 which is the Strategic Diagram of the entire corpus. Figure 5: AFS 2006 Strategic Diagram of All Abstracts