BMS English Abstracts 1993-2003 (40-79)
N. 40, September 1993
Le PEM, Pourcentage de l’Ecart Maximum: un indice de liaison entre modalités d’un tableau de contingence
The PEM, Percentage of Maximum Deviation: An Index of Tie between Modalities of a Contingency Table. We define an index of tie between modalities of a contingency table, the PEM, percentage of maximum deviation. Two concrete examples are given. The PEM allows us to construct profiles or a set of modalities of survey responses which are attracted to a modality. The PEM also permits us to construct styles of behavior, modalities which have a set of similar PEMs in a cross-tabulation of descriptive variables. PEM, Profiles, Styles, Tri-deux.
A method for measuring network effects in scientific cooperation
A method for quantitative assessment of the evolution of scientific networks with a light-and-fast mailed questionnaire is presented, through its use in the evaluation of the SCIENCE/STIMULATION programmes of the European Community Framework programme. The methodology is based on the behavioral description of relationships between labs sharing an EC contract, before and after the programme. Networks are described by a “mean value” of relationships between pairs within the network. Quantitative indicators, obtained through monovariate and multivariate statistical methods, allow comparing the situation before and after the programme, and therefore yield clues for programme impact assessment. An attempt to visualize network evolution with multivariate analysis is presented.
Mathematical Modelling and Computer Simulation of Social Processes: Problems and a New Solution1
Klaus G. Troitzsch
The paper discusses the state of the art in mathematical modelling and computer simulation of social processes. Its first section is devoted to the question why so little use has been made so far of these powerful tools. In the second section it describes the steps taken in any scientific modelling activity – the identification of some part of reality as a “real system”, the detection and/or reconstruction of the laws governing the part of reality to be modelled, the combination of our notions of these laws into a more or less formal model, and (in the case of computer simulation) the running of a computer language version of the model. Models – and computer simulation as well – may be classified for their dynamicity, stochasticity, linearity, for the level (micro vs. macro) and for the structure of their state and time spaces (discrete vs. continuous). In a third section the paper discusses uses and problems of some well known simulation approaches used in the social sciences – DYNAMO and microanalytical simulation models – and mentions some recent achievements in the development of microsimulation tools, especially (in the fourth section) the micro and multilevel modelling and simulation software MIMOSE developed in the last years by the author’s working group. The last section of the paper is devoted to general problems of computer simulation in the social sciences as compared to mathematical approaches: wherever inferences from multiple hypotheses cannot be drawn by logical or mathematical deduction, simulation at least allows us to draw inferences from fixed initial conditions and combinations of parameters, yet it is no substitute for any logical and mathematical analysis. Whoever makes use of simulation to contribute to the solution of socially or politically relevant questions should be aware and make his audience aware that simulation is never more than the solution of a formal model for a given parameter vector and a given set of initial conditions – which both have to be justified -, and that stochastic simulation is even less: one single realization of a stochastic process. Simulation tools should not only make this awareness possible, they should promote and – even better – enforce it. Mathematical modelling of social processes, computer simulation, microanalytical simulation, multilevel modelling, formal models, chaotic behaviour.
N. 41, December 1993
Mode Effects in Survey Research: A Comparison of Mail, Telephone, and Face to Face Surveys
Edith D. de Leeuw
In survey data four potential sources of error can jeopardize the results: the respondents, the interviewers, the questions, and the data collection method. In the last two decades, a shift has occurred in the way survey data are collected; telephone surveys and to a lesser degree mail surveys are now more extensively used. This has stimulated empirical research on the influence of the data collection method on data quality. A review of the empirical literature suggest small differences between mail surveys on the one hand and interview surveys (both telephone and face to face on the other. A controlled field experiment in the Netherlands confirms these results. It is somewhat harder to have people answer questions in a mail survey as indicated by a higher percentage of item missing data, but when questions were answered the data quality is better, as indicated by more self-disclosure on sensitive topics and more consistent answers. No data collection method was superior on all criteria. Furthermore, two substantive structural equation models (one on loneliness and one on general well-being) were compared across the three data collection methods. The different data collection methods turned out to produce significantly different covariance matrices. Subsequent analyses showed that structural models also differed: the same structure was found in each case, but the relative importance of some estimated parameters varied considerably across data collection methods. Mode Comparison, Data Collection, Data Quality, Measurement Error, Face To Face Survey, Telephone Survey, Mail Survey, Interview.
Sociology as a Profession in Europe
The first part of this article concerns Germany and the professionalization of sociology, its development as an academic discipline, the concept of professionalization, elements hindering professionalization and possibilities of professionalization in Germany. The next section gives an overview of how sociology is taught in France, Great Britain, Netherlands and Italy, and job opportunities for sociologists as an element of professionalization in these same countries. This section concludes by remarking that no ideal model is in sight, however sociologists are beginning to analyze their position. The last section concerns sociology in the old Länder of the Federal Republic of Germany, the historical situation of sociology in Germany, teaching and training in sociology, and the job market for sociologists. Sociology, Sociologists, Professionalization, Germany, France, Great Britain, Netherlands, Italy.
Researching THE MARKET FOR SOCIOLOGISTS AND SOCIOLOGISTS’ CAREERS: OVERVIEW AND INITIAL RESULTS
Karl M. van Meter
In preparation for the International Sociological Association’s Research Council conference in Onati, Spain, on 7-11 April 1992 on “Contemporary Uses of Sociological Research” a methodic study of research, documentation and data of sociologists’ careers and the market for sociologists was done. Major data bases were searched and analyzed. Professional organizations were also contacted as were governments and international organizations. An analysis and overview of this material is presented. Sociologists, Sociologists’ Careers, Sociologists’ Training, Job Market for Sociologists, Data Sources on Sociologists.
N. 42, March 1994
Karl M. van Meter
Sociology progresses through the joint advancement of both theory and methodology. The diversity of its theoretical approaches is mirrored by the diversity of its methodologies. There has been a disappearance or loss of pertinence of the distinction between “quantitative” and “qualitative” methodologies. Perhaps a more constructive distinction is that between “ascending” and “descending” methodologies, noting however that these are complementary and may be combined in a single research project. All methodologies have a “non universal” character even though each has its particular domain of competence and pertinence. Multi-method analysis is valuable in obtaining stable results and opening up communications between subdisciplines. A general sociological research procedure is presented in four “classic” points and to which two often unexplicited basic steps are added: the initial transformation and the final transformation of information during research procedure. Ascending and Descending Methodology, General Research Procedure, Multi-Method Analysis, Qualitative and Quantitative Methodology.
From Paradigms to Eclecticism: Thematic Profiles of German Language Core Sociology Journals 1984-1991
Heinrich Best, Renate Ohly
The bibliographic database SOLIS (Social Sciences Literature Information System) of the Social Sciences Information Centre in Bonn was used as an empirical basis for an investigation of the changing thematic and paradigmatic structure of sociology in German speaking countries. The articles published in six leading professional journals from 1984 to 1991 constitute the universe of observation. A content analysis was performed using highly aggregate bibliographical descriptors as the classification scheme. The resulting data are analyzed by means of correspondence analysis. As a general trend a shift from “paradigms” towards “eclecticism” can be detected. While in the beginning the journals and subjects addressed can be located within a thematic space formed by the two main axes, macrosociology vs. microsociology and theory/methodology vs. practical application, in the end this clear cut thematic structure wanes in favour of looser relationships between subjects and journals. Bibliometric Research, Correspondence Analysis, Paradigmatic Change, Scientific Development.
PRACTICAL ISSUES IN COLLECTING LIFE-TIME WORK HISTORIES IN SURVEYS
Pamela Campanelli, Roger Thomas
This research addressed a critical methodological problem raised by the proposal to conduct a survey of `Working Lives’ in the UK. The problem was: how to elicit a detailed and dated record of episodes constituting respondents’ working lives with the maximum degree of completeness and accuracy, but within the constraints of a standardized, large scale, quantitative interview survey. After a discussion of practical issues, this paper describes a small-scale empirical study using respondents from the general population. Respondents were first interviewed using a newly developed hybrid quantitative questionnaire. They then participated in an in-depth interview drawing from both cognitive and qualitative traditions designed to maximize the completeness and accuracy of the final career record. A comparison of the two approaches suggested that the quantitative survey had missed 25 percent of the required labour force events and that certain types of events were much more likely than others to be missed. General topics discussed include the use of calendar aids, the use of an advance letter, the presence of other people during the interview, the availability of external `validation’ documents, the impact of the length of the interview on data quality, and the implications and concerns of collecting other types of life histories at the same time. Whereas the focus of this study was on recall of jobs and other labour force events and episodes, we believe that the findings and interpretation are of relevance also to the recall of other types of life event. Life Histories, Work Histories, Surveys, Data Collection, United Kingdom.
N. 43, June 1994
Applying sociology: conversation analysis in the study of human-(simulated) computer interaction
In recent years much research has been done to try to design computer-based speech systems which will be sensitive to the requirements of human users. However, this entails a ‘chicken and egg’ type of problem: how can system designers know how people will react to computers, and what their requirements will be, prior to the development of an experimental system, and how can an experimental system be developed prior to an understanding of users’ behaviour and requirements? This paper discusses two methodological responses to this dilemma adopted by researchers in the Department of Sociology at the University of Surrey, England. First, a Wizard of Oz (WOZ) simulation technique is described. This involves an accomplice (the ‘wizard’), whose voice is electronically disguised, pretending to be a talking computer; experimental subjects are then led to believe that they are interacting with a piece of existing technology. These exchanges are recorded, and then studied to yield information about human users’ requirements and communicative competencies. There then follows a discussion of the appropriateness of a qualitative, conversation analytic approach in the study of the data generated by the Surrey simulation experiments. The analytic sections focus on two communicative strategies used by the subjects to identify and deal with possible trouble sources in the exchange with the ‘system’. The analysis examines the organisation of these strategies, and discusses some of their interactional benefits. Conversation Analysis, Discourse Analysis, Human Computer Interaction, Simulation Studies, Requirements Capture.
L’application de la lexicométrie dans une perspective sociologique
This article shows some possible uses of lexical analysis through the analysis of a sociological corpus concerning expectations and roles of parents of North African origin in their children’s education. This corpus will be automaticaly processed in order to show the importance of computer analysis and test whether available software allows easy corpus analysis. Lexicographic Analysis, Symatic Analysis, Discourse Analysis, Content Analysis, Family, Education.
RECENT TRENDS IN VALIDITY THEORIES OF MEASUREMENT
K. Walter Schwager
Since 1954 the dominant approach to measurement adequacy in U.S. psychology, education, and sociology has been the theory of validity. At first three separate types of validity were identified: criterion, content, and construct validation. More recently important changes have taken place in validity theory, including the rise of a unified approach to validation, where all types are integrated into construct validation; some disillusionment with the ideas of construct validation; and a growing influence of Donald Campbell’s approaches to measurement. Causal modelling has become accepted as another approach to validity testing. This paper reviews these modifications, and briefly assesses them from a metatheoretical point of view. Validity, Content Validity, Criterion Validity, Construct Validity, Face Validity, Causal Modelling.
Une Petite Pre-Enquete: The Challenge of Social Network Research in France
Alden S. Klovdahl, Alexis Ferrand, Lise Mounier
A small exploratory study was carried out in two areas of France (Paris and Reims) to test the feasibility of obtaining nominative data (names and addresses) needed to reconstruct the large social networks that connect individuals in modern society. In carrying out the study two kinds of ‘resistance’ were encountered: firstly, a reluctance to participate in surveys in general; and, secondly, a hesitation among those who did participate to provide nominative information. An evaluation of the exploratory study leads to suggestions for future research on social networks in France. Social Networks, France, Surveys, Data Collection Problems.
N. 44, September 1994
Cultural Differences in Organizational Communication: A Semantic Network Analysis
Ha-Yong Jang, George A. Barnett
This paper examines the impact of national culture on organizational culture by analyzing messages directed to external audiences. The structural equivalence of Japanese and American corporations with stock offerings in the United States was examined through semantic network analysis. 35 Fortune 500 companies were matched by their businesses. The full texts of the chief operating officers’ letters from the annual reports for 1992 were analyzed by first determining the most frequently used words in all 35 letters. Then, the frequency of each word for each company was determined. A companies by word matrix was created, which was pre-multiplied by its transpose creating a 35 X 35 companies sociomatrix based on the coocurrence of the words. The analysis revealed two distinct groups, one composed of the Japanese companies and another made up of the American. The Japanese clustered very tightly, while the American was fairly loose. The American companies discussed financial information and the structure of organization, while the Japanese describe organizational operations. A discriminant analysis revealed that the two groups could be perfectly differentiated by the texts. The companies’ business were not reflected in the messages, only their national culture. National Culture, Organizational Culture, Japan, U.S.A., Semantic Network Analysis, Coocurrence of Words.
Computer assisted data collection, data quality and costs: A taxonomy and annotated bibliography
Edith D. de Leeuw
Through published calls for contributions and data base searches, the author has surveyed literature on computer assisted data collection and data quality to compile this annotated bibliography in which each reference is coded according to a taxonomy. Bibliography, Key Words, Taxonomy, Computer Assisted Data Collection, CATI, CAPI, CASI, DBM, EMS, Telepanel, Costs, Interviewer Training, Interviewer Effect, Response Rate, Data Quality.
INTENSITE DE LIAISON ET MASSE D’INFORMATION DES TABLEAUX DE CONTINGENCE: DEUX PROBLEMES DE MESURE EN ANALYSE DES DONNEES.
Link Strength and Volume of Information in Contingency Tables – Two Problems of Measure in Data Analysis. This article presents on-going work concerning the unhealthy use of certain statistical measures. It first of all proposes a tool for measuring the internal correlation within a large dimensional contingency table in order to be able to compare other similar table in this framework by using a coefficient whose values are between 0 and 1. The article then studies the information contained in a table which would correspond to the situation of statistical independence between variables during correspondence analysis. Using the hypothesis this quantity of information should be taken into account before a correspondence analysis, the article then proposed a tool for measuring that volume of information that sociologists usually do not treat as such. Contingency Tables, Statistical Independence, Measure of Information, Correspondence Analyse.
Liaison et information dans les tableaux de contingence: Commentaire de l’article de Frédéric Michel
Links and Information in Contingency Tables – Commentary on Frédéric Michel’s Article. Frédéric Michel’s article (“Link Strength and Volume of Information in Contingency Tables – Two Problems of Measure in Data Analysis”) published in this issue, proposes solutions to a certain number of problems. It would be helpful for research on these problems to situate the author’s solutions in the framework of preceding research on the question. Data Analysis, Contingency Tables, Measures of Information, Link Strength.
N. 45, December 1994
The social structure of sociology in spain
Josep A. Rodriguez
The goal of this paper is to draw the social structure of academic sociology in Spain applying network analysis techniques to the study of the social organization of sociology and of its cultural product through the relations between areas of research and university departments. Specifically I use MDS to study the institutional field of sociological knowledge and academic sociology as well as the social organization of the sociological production. I use two sets of relational data: one built upon the exhaustive description of the 42 areas of sociological work in Spain prepared for the World Congress of Sociology in Madrid 1990, and the other built upon the analysis of the articles published during the past five years by the leading Spanish journal of sociology (Revista Espanola de Sociologia). The analysis is contextualized within the process of construction of Sociology as an academic specialty during the past decades on one side, and the characteristics of current sociological production on the other. My objective is not only to show the system of relations and influence between universities and areas of research and their impact on current sociological knowledge and work, but also to open the door to the understanding of Sociology as a social construction attempting to address old and new social phenomena. Network Analysis, Social Construction of Sociology, Spanish Sociology.
A qualitative protocol for Studying Technological change in the labor process
James R. Zetka, Jr., John P. Walsh
The article specifies, illustrates, and defends the logic of a qualitative approach for studying the workplace outcomes of technological change. The first stage involves the use of intensive case-study methods to discover the normative orders that mediate shop floor reactions to technological innovation. The second stage involves the use of archival analyses to test the hypotheses developed in the first stage and to generalize them to the industry over time. The third stage involves cross-case comparison to generate general theoretical statements. This approach is illustrated with reference to Zetka’s (1992a; 1992b) study of technological innovation in auto body building and to Walsh’s (1989; 1991) study of retail meatcutting. Qualitative Methodology, Historical Methodology, Labor Process, Technological Change, Workplace Relations, Automobile Industry, Meatcutters, Retail Food Industry.
The Use of Weberian Ideal-Type Methodology in Qualitative Data Interpretation: An Outline for Ideal-Type Analysis
This article gives a short overview of what ideal-type analysis is, how it is grounded in Weber’s works and how it may be used in qualitative data interpretation. It has four main parts. First, Weber’s conception of ideal-types is reconstructed from his writings. Second, Weber’s ideas are reinterpreted such that a three-step methodology emerges, which is, third, depicted in more detail. The three steps of ideal-type analysis are outlined which is an adaptation as well as an application of Weber’s thought while fitting the needs of interpretive data analysis. Fourth, from one of two longitudinal studies the author has conducted, evidence is given to clarify how systematic qualitative data interpretation may use ideal-type methodology and yield insightful findings. Max Weber, Ideal-Type Methodology and Analysis, Longitudinal Analysis, Qualitative Data Analysis.
LA DEMANDE DE LOGEMENT: L’EXEMPLE DE SAINT-PRIEST (RHoNE)
Requests for Housing – The Case of Saint Priest in the Rhone. East of the city of Lyons, the town of Saint Priest is a perfect example of a suburb. Its public housing constitutes a third of the city’s buildings. The major objective of the research on which this article is based is to describe the different types of requests for housing sent to the public housing authority, the OPHLM, between 1987 and 1991. By presenting a systematic analysis of the great diversity of those requesting housing, this study’s intention is also to question the management of public housing policy concerning its evolution and the new problems it has to face. This work is, in a larger perspective, part of the research program that the “Groupe de Recherche sur la Socialisation,” a sociology research laboratory of the CNRS and the University Lumière-Lyon 2, has been pursuing for some time now on the relationship between space and socialization. Requests for Public Housing, Socio-Economic Profiles.
N. 46, March 1995
COGNITIVE MAPPING OF AIDS RESEARCH 1980-1990: STRATEGIC DIAGRAMS, EVOLUTION OF THE DISCIPLINE AND DATA BASE NAVIGATION TOOLS1
Karl M. van Meter, William A. Turner, Jean-Baptiste Bizard
Using the LEXINET and LEXIMAPPE computer programs for cognitive mapping and scientometric analysis, two of the authors have previously published a strategic diagram based on the analysis of the 290 last research articles (all of 1989 and part of 1990) on AIDS in the Sociological Abstracts data base covering 1980 to 1990 (K. M. van Meter and W. A. Turner, Current Sociology, 1992, 40, 3, 123-134). Here we present a re-analysis of the same data and two other strategic diagrams concerning this AIDS research data base, respectively covering the periods 1980-1986 and 1987-1988. These strategic diagrams, each with its four quadrants corresponding to “Main Stream”, “Ivory Tower”, “Up-and-Coming” and “Bandwagon” research, permit us to describe the birth and development of a scientific discipline, analyse the evolution of its structure and content over time, and set forth certain hypotheses concerning scientific innovation and itineraries or trajectories of researchers in a new discipline. AIDS, Scientometry, Scientific Innovation, Structure over Time.
Analyse latticielle d’un réseau de proximité
Ameziane Cherfouh, Vincent Duquenne
Lattice Analysis of a Proximity Network. Two hundred interviews were done in an industrial neighborhood of Kigali, Rwanda (Central Africa), by Sylvia Servaes (Cologne) before the current bloody civil war. The interviews describe: personal possessions, quality of vie in the neighborhood, and reasons for remaining in the neighborhood or leaving it. Lattice analysis and the computer program GLAD are used to study associations between respondents and possessions, and implications between conjunctions of possessions concerning facilities, the house and living rooms. Facilities include source of water and lighting. The house involves a description of the living quarters and construction materials employed. Living rooms include the description of possessions that are found there. Quality of life in the neighborhood is analyzed in two way: through the reasons inhabitants give for wanting to stay at Camp Zaire, or on the contrary, their reasons for wanting to leave. In the last part, added details on wanting to stay or leave are given and furnish a description of inhabitants who are not satisfied in general with their situation but all the same want to stay where they are. This analysis reveals the hierarchy and status of possessions. Reasons for wanting to leave the neighborhood show more implications and dependencies than reasons for staying. Reasons for being unsatisfied are rather clearly hierarchical and are nonetheless compatible with wanting to stay. An interpretation of these descriptive results can help (re)formulate hypotheses and test anthropological models of neighborhood relations. Lattice Analysis, Proximity Networks, Galois Lattices, Basis for Implication, Extension/Intension Duality.
COVARIANCE STRUCTURE MODELING IN WINDOWS: A MULTITRAIT-MULTIMETHOD ANALYSIS USING AMOS, EQS, AND LISREL
Joop J. Hox
This article examines the three major packages that allow covariance *sanalysis (CSA) under Windows. The main point of comparison is the analysis of a 5×5 multitrait-multimethod matrix. This shows that the availability of different fit functions and bootstrap procedures in modern CSA programs allow a detailed analysis of problematic data. Covariance Structure Analysis, Structural Equation Modeling, Causal Modeling, Bootstrap, Fit Functions, Multitrait-Multimethod.
FEMINIST METHODOLOGY – IS IT FACT OR FICTION?
This paper debates the concept of a feminist methodology in the context of methodological pluralism and absolutist/superiority claims of knowledge. It briefly outlines the definitional confusions surrounding the use of the term and the risks of tautology. It argues against exclusiveness, imperialism and certainty of knowledge and proposes that, as feminism’s greatest contributions lay in opposing these qualities, any notion of prescriptive models of research paradigms must be resisted, i.e., a feminist methodology per se is theoretically unthinkable and intellectually undesirable. Gender and Methodology, Feminism, Methodolatry, Interview, Research Bias, Methodology and Epistemology.
N. 47, June 1995
QUALITY AND QUANTITY, AGENCY AND STRUCTURE, MECHANISM AND CONTEXT, DONS AND CONS
The author reviews and breathes life into the ‘debate’ on qualitative and quantitative methods in sociological research by turning upside down the notion that the terms ‘quantitative’ and ‘qualitative’ provide a basic organising framework for research. Instead, he argues that these labels speak of an essentially technical distinction which should be understood as the distant outcome of more fundamental concerns about the nature of social being and sociological explanation. The article begins with some considerations concerning the history and context of the qualitative and quantitative ‘debate’, followed by observations about the cross-purposes underlying this early debate. The author then establishes the key current question about the precise mechanics of fusing the two modes of inquiry. The main body of the paper establishes a framework for the juxtaposition of the two traditions, under the title of the ‘realist synthesis’ with the establishment of some proposals on research design. The author then tackles the issue of how to escape the design mind set which asks ‘is it a survey?’ or ‘is it an ethnography?’ Finally, he focuses on data collection and the old, supposed opposition between ‘structured questionnaires’ versus ‘unstructured interviewing’. To illustrate and substantiate with some real examples in evaluation research (the investigation of social programmes to discover whether they ‘work’ to the desired policy ends), examples of ‘criminal justice’ evaluation are presented from the author’s own investigations of the effect of educational programmes on the rehabilitation of prisoners. Qualitative Methods, Quantitative Methods, Realist Synthesis, Evaluation Research, Prisons, Rehabilitation.
Un exemple d’articulation de méthodes d’analyse qualitatives et quantitatives sur des entretiens semi-directifs: les représentations du handicaP
Alain Giami, Jean-Louis Korpès, Chantal Lavigne, Régine Scelles
Linking Qualitative and Quantitative Methods of Analysis Applied to Semi-Structured Interviews: Representations of Handicap. This article presents an exploratory process in methodological research to link the use of computer data processing with a qualitative analysis of categories in a sample of semi-structured interviews. After presenting the theoretical basis for the analytical grid, and the problems encountered in its application to qualitative material, the authors explain their use of a data base which allows processing the categories with the help of a data segmentation and classification software (AC2). This analysis was carried out secondarily on the categories established by the researchers in order to produce a more precise representation of the way the different variables were organised, by pinpointing their discriminating power. This automatic method is used as a complementary tool for systematising results obtained with the qualitative method and to support interpretations based on exhaustive consideration of these categories. Qualitative Methods, Quantitative Methods, Semi-Structured Interviews, Computer Methods, Representations of Handicap.
METHODOLOGIE DE LA DEMARCHE DE RECHERCHE EN SOCIOLOGIE Didactique du projet de recherche
Methodology of Sociological Research Procedure: Didactics of a Research Project. The objective of this article is to present sociological research procedure, and, more concretely, to show that sociological investigation – even in its most technical aspects – is inseparable from theoretical considerations. The problems a researcher encounters in trying to link theoretical considerations with empirical operationalization are not really treated in sociological literature. The author presents and analyzes the three components of a problematic: the frame, the approach, the prospective. The research frame – and through it, the research object – has less sociological value the closer it is to common sense knowledge. An approach can by atemporal or chronological. A perspective is an individual researcher’s theoretical position in which he expresses his scientific institutional identity. The author applies these ideas to an example of the study of social mobility. Sociological Methodology, Sociological Literature , Research Problematic, Social Mobility.
Using computer simulation to study social phenomena
Computer simulation has enjoyed a recent revival as a methodological tool in the social sciences. In sociology, advances in hardware and, particularly in software, have allowed the construction of much more interesting simulation models than before. This paper reviews two strands of recent work on computer simulation: dynamic micro-simulation and simulation based on distributed artificial intelligence, and suggests some general methodological principles for simulation research. Simulation shares some of the same methodological difficulties as other types of modelling, including statistical modelling, but also offers new opportunities. Computer Simulation, Computer Programs, Sociological Methodology, Modelling.
N. 48, September 1995
SURVEY NONRESPONSE, MEASUREMENT ERROR, AND DATA QUALITY; AN INTRODUCTION
Johannes van der Zouwen, Edith D. de Leeuw
In this introduction, the authors show the relationship of the five papers selected for this special issue on survey nonresponse and measurement error, with three important indicators of data quality in surveys, i.e., representativeness, completeness and accuracy. Consequences for survey practice and methodological research are sketched. An overview of related (research) efforts are presented. Nonresponse, Measurement Errors, Data Quality.
RESPONSE DEVELOPMENTS AND THE FIELDWORK STRATEGY
Cees Maas, Wim de Heer
Nonresponse in household surveys is a matter of great concern in many countries. However, response is not decreasing for all surveys and not in all countries. Response rates can be affected for a great deal by design factors and factors related to the fieldwork strategy. Comparison of response results and survey characteristics across countries supports this idea. Depending on the survey design very high response rates can be obtained if the fieldwork strategy is a well controlled one. Response, Fieldwork Strategy, Survey Design.
SURVEY PARTICIPATION AS REASONED ACTION; A BEHAVIORAL PARADIGM FOR SURVEY NONRESPONSE?
Joop Hox, Edith de Leeuw, Harrie Vorst
Survey nonresponse is a threat to the inferential value of the survey method. Nonresponse is by now a recognized problem in the USA and in Western Europe. To adequately fight the nonresponse problem extensive knowledge about respondents and nonrespondents is necessary. In this article we describe a method to successfully gather information of both respondents and nonrespondents. This method was developed and tested by Cialdini in the USA; we show that this method can be successfully utilized in the Netherlands and that Cialdini’s results have cross cultural validity. Furthermore we show that the decision to respond or refuse to a request for survey participation can be partially explained by the theory of reasoned action. Theory of Reasoned Action, Nonresponse, Response, Survey, Experiment, Cialdini, Cross-Cultural Replication.
THE PROFILE OF THE DIFFICULT-TO-INTERVIEW RESPONDENT
In this article we investigate the assumption that all respondents are motivated to cooperate and that they have sufficient cognitive and communicative skills to execute the respondent’s role adequately. A typology of respondent’s motivation and ability is constructed with a latent class analysis. The classes of this typology can be characterized according to how easy or how difficult they are to interview. It is possible to relate the typology of respondents with age and education. The typology is also related to concrete aspects of response behaviour: “use of the don’t know response category”, “inconsistent answers” and the “use of a restricted response style”. The difficult-to-interview respondents have a low educational level, are older, use the DK category frequently, and tend to give inconsistent answers. Survey Data Quality, Respondent Motivation and Ability, “Don’t Know” Response, Inconsistent Answers, Response Style, Interviewer Reports.
Computer-Assisted Questioning: The New Survey Methods in the Perception of the Respondents
After providing a literature review about existing mode comparison studies regarding respondents’ reactions and data quality in the field of computer-assisted questioning methods, we will present results from a study in Germany (total N=152). This study compared Computer-Assisted Personal Interviewing (CAPI) and Computerized Self-Administered Questionnaires (CSAQ) with traditional Paper and Pen Interviewing (PAPI). The focus was on acceptability and resulting data quality aspects of the new technologies. In conclusion, there is some evidence to the assumption of better data quality in computer-assisted questioning methods. However, only a few well designed methodological studies on CAPI and CSAQ have been conducted and further research in this area is recommended. Data Collection, Computer-Assisted Interviewing, CAPI, CSAQ, Data Quality.
Mental Construal Processes and the Emergence of Context Effects in Attitude Measurement
Norbert Schwarz, Herbert Bless
We present a theoretical model of the cognitive processes underlying context effects in attitude measurement. The model predicts (a) the conditions under which context effects are likely to emerge; (b) their direction (i.e., assimilation or contrast); (c) their size; (d) their generalization across related items; and (f) their dependency on the mode of data collection used. Experimental research bearing on these predictions is reviewed and implications for questionnaire construction are discussed. Context Effects, Attitude Measurement, Response Effects, Assimilation, Contrast, Judgment, Mental Construal.
Levels of Protection: Confidentiality in Network Research
Alden S. Klovdahl
A desire to protect human subjects from harm has stimulated discussion of appropriate safeguards in various areas of scientific endeavor. Ironically, procedures for protecting the confidentiality of information provided by participants in social network research have received little attention in the literature. Here the protection of data obtained in social network research is discussed and it is suggested that simple, easily implemented procedures provide adequate protection in most cases. Privacy, Data Protection, Network Analysis.
N. 49, December 1995
Modeling Conflict and Exchange in Collective Decision Making
Frans N. Stokman
Two dynamic models of collective decision making are introduced and illustrated with a simple example. A more extensive presentation and application concerning the European Community can be found in Bueno de Mesquita and Stokman (1994). The two dynamic models reflect two alternative views of collective decision making and politics. The first, represented in the expected utility model, conceives of collective decision making as conflict resolution, a non-cooperative game, fundamentally different from exchange relations in economics. The second, represented in the exchange model of Stokman and Van Oosten (1994), does not see fundamental differences between economic exchanges and political decision making. Dynamic Models, Collective Decision Making, Conflict Resolution, Economic Exchange Relations.
Se souvenir de son passé professionnel: Appel à la mémoire dans les enquêtes rétrospectives et construction sociale des données
Thomas Couppié, Didier Demazière
Remembering One’s Work History – Memory Recall in Retrospective Surveys and Social Construction of Data. This article analyzes the role of memory in the production of retrospective data concerning work history trajectories. It is based on the comparison of two surveys carried out in 1989 and 1993 on the same sample populations of youths and concerning, in part, the three and a half years following the end of their education. The article examines the nature of the collected information and reveals two essential dimensions in the social construction of recall of career developments: the subjective production and reinterpretation of one’s own trajectory for the survey; and the translation of biographical events into official categories which are increasingly vague (unemployment, work, inactivity). The article highlights the discrepancies between the answers to the two surveys, shows that the discrepancies cannot be reduced to a simple question of data reliability, and reveals the social processes governing entry into a professional career trajectory. Retrospective Surveys, Memory, Entering Professional Careers.
The Reliability of Recall Data: A Literature Review
This paper reviews the literature which claims to test the reliability and validity of recall data primarily on individuals’ life and work histories. It is important for both survey designers and analysts to understand the constraints on collecting reliable recall data and the levels of the quality of any data collected. The references are mainly from academic social survey researchers, some government social survey research and psychological studies in the USA and Britain. The main aims of the paper are to summarise what we know about the reliability of various sorts of recall data and provide guidance, in the light of this summary, to survey designers and survey analysts. In order to fulfill these aims, the methods which have been used to test validity and reliability of recall data are reviewed, evaluated and ranked. In practice it is very difficult to test the validity of recall data, whereas there are a number of methods of testing for reliability. Reliability, Recall Data, Retrospective Data, Data Quality.
Statistique, Imagerie et Sciences Cognitives
Monique Le Guen
Statistics, Imagery and Cognitive Sciences. For two decades, the teaching and application of statistics have been dominated largely by the calculatory and logical procedures of mathematics. But mathematics is not accessible to all, while statistical “thought” should be shared by the largest possible number of people. How can one facilitate the learning of statistical thought? “I have ideas because I have images”, said Euler. Cognitive sciences are trying to show that Euler was right. The role of imagery in cognitive processes and learning strategies developed during the learning process are currently the subject of research and results are now being made available outside the restricted specialized research community. Today, research in cognitive sciences can furnish conceptual aids for designing programs of learning intelligently assisted by computers (LIAC). Statistics, Scientific Visualization, Imagery, Cognitive Sciences, Pedagogy, Learning Intelligently Assisted by Computers (LIAC).
n. 50, Marche 1996
MODELES D’EQUATIONS STRUCTURALES ET SENS DE LA CAUSALITE DANS LES ETUDES LONGITUDINALES: UNE APPLICATION AU BIEN-ETRE SUBJECTIF(1)
Paul Dickes, Jean-Luc Kop, Jocelyne Tournois
Structural Equation Models and Direction of Causality with Longitudinal Data – An Application to Subjective Well-Being. When two variables are correlated, the researcher is often confronted with the question of the direction of causality. This substantive question is, however, difficult to answer, especially when use of an experimental design is impossible. In such situations, longitudinal data furnish some invaluable information: the temporal order of the variables. Unfortunately, the temporal order is not sufficient to answer questions of causality. The well-known technique of cross-lagged correlations has been criticized because of its unrealistic assumptions. The major purpose of this paper is to illustrate the use of structural equations models to help answer questions regarding the direction of causality in longitudinal data. After a brief presentation of structural equations models, focalizing on the LISREL model, we stress their advantages over more traditional approaches. An empirical illustration is presented which uses the data on subjective well-being published by Headey et al. (1991). We study the direction of causality between marriage satisfaction and general satisfaction. The different models tested seem to infirm the conclusions of Headey et al. Causality, LISREL, Longitudinal Data, Structural Equation Models, Subjective Well-Being.
THE SENSITIVITY OF HIERARCHICAL CLUSTERING SOLUTIONS TO IRRELEVANT VARIABLES
Istvan Hajnal, Geert Loosveldt
Cluster analysis is often used as an exploratory tool. We study the sensitivity of clustering solutions to irrelevant variables. After a brief review of the literature, a simulation study is presented in which a noise variable is added to a set of variables with an a priori cluster structure. This simulation study shows that most clustering techniques are sensitive to irrelevant variables, but that Ward’s method was only mildly affected. Even without the noise, the performance of the single linkage method was troublesome. We conclude that classical cluster analysis should only be used when one has a good idea of which variables to use in the analysis. Otherwise, one should use a weighting scheme or a dimension reduction technique that preserves cluster structure. Unfortunately, these methods are not yet widely available to sociologists. Cluster Analysis, Noise, Sensitivity, Simulations.
Using SAS to Convert Ego-Centred Networks to Whole Networks
Caroline Haythornthwaite, Barry Wellman
This paper presents a technique for using SAS software to convert ego-centered network data into whole network data. Until now, these two types of networks have largely been analyzed separately, using different techniques. Ego Networks, Whole Networks, SAS.
Are university departments research organisations? A paradox of academic life
Tensions being experienced within the higher education system arise from challenges to the principle of departmental autonomy. One of the challenges currently facing social science departments is how best to mobilise themselves as effective units of productive research. This article addresses one issue arising from this change in the funding climate: do academic departments outside the sciences regard themselves as research organisations, or are they simply a collection of individual scholars doing individual research? Following a national British survey, five modes of research organisation may be distinguished in the social sciences, which stand in different relationships to the departmental structure of universities. Social Sciences, Research, Teaching, Universities, Departments.
n. 51, June 1996
INTRODUCTION – MULTILEVEL ANALYSIS
Pieter van den Eeden, Joop J. Hox
The authors of this article, who are also the guest editors for this issue on multilevel analysis, give an overview and brief history of multilevel analysis and present the following four research articles. Multilevel Analysis – Overview, Methods, History.
PRIMACY OF MULTILEVEL ANALYSIS WITH RESPECT TO HIERARCHICALLY ORGANIZED DATA
Cora J.M. Maas
Nested data are increasingly analyzed by multilevel-analyzing-programs, specially developed for this purpose. The advantages of such programs for analysis compared with the techniques previously used are set out in this article, both in theory and together with an empirical example. Hierarchically Organized Data, Multilevel Analysis, Multilevel Computer Programs, Comparison of Methods.
MODELLING INTERGENERATIONAL TRANSMISSION IN LONGITUDINAL BIRTH COHORTS USING MULTILEVEL METHODS
Richard D. Wiggins, Chris J. Wale
This contribution presents a multilevel analysis of intergenerational processes. The methodological issues of standardization and selection effects are considered. The results show, that age standardization does not work by itself, but age must be introduced in the models as well. Multilevel Analysis, Longitudinal Cohorts, Effects of Age.
CONTEXTUAL EFFECTS OF CLASS MOBILITY ON VOTING BEHAVIOUR IN 16 WESTERN COUNTRIES: 1956-1990
The author tests several hypotheses about the contextual effects of class mobility on the voting behaviour of intergenerationally stable class members. These hypotheses pertain to the effects of levels of outflow from and inflow mobility to a specific social class (in a certain nation and a specific year). The hypotheses are tested by multi-level modelling, and analyzing data on 20,619 respondents from 113 surveys held in sixteen industrialized democratic countries in the period 1956-1990. Despite these efforts, the analyses show no significant contextual effect of either the level of intergenerational inflow or the level of outflow social mobility in a country on the voting behaviour of intergenerationally immobile persons. In the discussion section possible explanations for these negative results are discussed. Multilevel Analyse, Voting Behaviour, Social Mobility, Context Effects.
THE EFFECTS OF INTERVIEWER AND RESPONDENT CHARACTERISTICS ON ANSWER BEHAVIOUR IN SURVEY RESEARCH: A MULTILEVEL APPROACH
Pieter van den Eeden1, Johannes H. Smit, Dorly J.H. Deeg, Aartjan T.F. Beekman
Until recently, the study of interviewer effects has focused on establishing direct effects of interviewer characteristics on respondent response. Recently, an alternative approach has been developed which emphasizes the conditioning influence of the interviewer characteristic on the respondent’s answering process. The objective of this paper is to illustrate this alternative approach with empirical evidence, using the random coefficient hierarchical regression model. This model’s structure is basically as follows. First, the answering process is described at the level of the respondent. Subsequently, respondent specific parameters are related to interviewer specific variables. This structure allows inclusion of the coefficient resulting from the intra-interviewer regression in the regression equation at the interviewer level (inter-interviewer regression model). Thus, the variance to be explained is split up in a respondent part (level 1) and an interviewer part (level 2). This two-level model is applied to data collected in the Longitudinal Aging Study Amsterdam (LASA; 2838 respondents within 43 interviewers). The dependent variable is a scale indicating well-being (Center for Epidemiologic Studies Depression Scale); background variables on respondent level are age, sex, and self-perceived health. Interviewer variables are age, education, personality traits and social skills. Questionnaire Surveys, Response Process, Interviewer Influence, Multilevel Analyse.
n. 52, September 1996
THE ANALYSIS OF POVERTY SPELLS
Eric Schulte Nordholt
The Analysis of Poverty Spells. In this paper data of the longitudinal database of the Dutch Socio-economic Panel survey (SEP) are used to analyse poverty spells. Net household incomes are standardized by means of the Statistics Netherlands equivalence scale. The half median method (a relative poverty measure) is used to determine which persons should be considered poor. Cox’s proportional hazards model is used to test the dependence of the poverty spell on a number of demographic, geographic and labour variables. Equivalence Scales, Longitudinal Data, Poverty, Proportional Hazards Model.
LES INCOHERENCES DANS LE RENSEIGNEMENT DU TEMPS POUR LE SECTEUR LINGE-COUTURE-TRICOT DE L’ENQUETE “MODES DE VIE”1
Incoherencies in the Time Data for Linen-Sowing-Knitting Activities in a “Life Styles” Survey. In INSEE’s “Modes de Vie (1988-1989)” life styles survey, whose objective is to quantify household activities, the comparison of two estimations of time spent on different activities for a non-negligible number of households, showed the data was incoherent. This article, which limits itself to the study of knitting and needle-work activities, based on the data files of the linen-sowing-knitting sector of the survey, gives an explanation of the existence of these incoherencies. Survey Incoherence and Errors, Effect of Imposing a Problematique, Memory, Quantification, Data Consistency.
COMPARATIVE ANALYSIS USING LARGE SCALE NATIONAL DATA SOURCES OF WOMEN’S EMPLOYMENT
Shirley Dex, Heather Laurie
This paper considers two main questions. First, how do the various theoretical frameworks for carrying out comparative research help us to gain a better understanding of the role women are playing in the labour markets of advanced industrial economies and are some better suited than others for this task? Secondly, we ask whether research practice can lead to better theorising on comparative research. The paper discusses the usefulness of various theoretical frameworks through the experience of using a number of cross-national data sets including a European Science Foundation network of Household Panel Studies and a number of two-country cross-sectional comparative studies. The paper argues that there may be things which can be learnt from engaging in the process of doing comparative research which are relevant to theorising about such research, and which are to some extent independent of the theoretical approach taken up at the outset. Cross-National Research, Cross-Cultural Research, Comparability, Women’s Labour Market Behaviour.
SUR LE TOURNANT METHODOLOGIQUE DE PIERRE BOURDIEU ET LE DEVELOPPEMENT DE LA METHODOLOGIE QUALITATIVE EN SOCIOLOGIE1
On the Methodological Evolution of Pierre Bourdieu and the Development of Qualitative Methodology in Sociology. This article concerns the “provoked and accompanied auto-analysis” presented by Pierre Bourdieu in La Misère du Monde. This latter work constitutes a true methodological experience defining, in a new manner, the epistemological break, the status attributed to common sense knowledge, the representativity and the singularity of sociology, as well as the construction of explanations as found in sociological literature. The criticism inspired by this method opens up a debate on qualitative methodology in sociology. Qualitative Methodology, La Misère du Monde, Common Sense Knowledge, Sociological Explanation.
n. 53, December 1996
FRAME ANALYSIS OF OPEN INTERVIEWS ON INTERETHNIC RELATIONS
Harry van den Berg
In recent literature on interethnic relations, the dynamics and flexibility of ethnic identities get much attention. The concept “situational ethnicity”, characteristic for this approach, implies that ethnic identities are context dependent: Selfconcepts and attitudes towards other ethnic groups are not conceptualized as stable and enduring dispositions but as frames which may vary with: (i) the domain of social interaction with members of other ethnic groups, and (ii) the specific conversational context in which these repertoires or frames are verbalised. As a consequence, it is assumed that the belonging to a specific ethnic group is not always an important element in the frames of references to be used for the perception and evaluation of other people and/or to be verbalised in conversations. The same applies to the way different ethnic groups are distinguished and defined. The paper will focus on the question how to get valid information on the context dependency of ethnic identities. After a discussion of pro’s and con’s of different research methods for this research goal, the paper will mainly focus on possibilities and pitfalls of open interviewing. Special attention will be paid to problems in interpreting interview protocols in theoretical meaningful terms. A main difficulty in this respect is due to the fact that the empirical variety in answering patterns is not only dependent on the type of social interaction which respondents have in mind but also on characteristics of the interview situation and the research design. A frame-analytic model is proposed on behalf of the analysis of frames as articulated in answering patterns and frames respondents use in defining the interaction with the interviewer. Special attention is paid to the possible relations between both types of frames. The research case in question concerns open interviews with pupils of a secondary school. Frame analysis of these interviews shows the role of frames concerning general aims of the research in the answering behaviour of respondents. In analyzing the ways these frames are constructed, the possible effects of the research design are discussed. The results of the study gives rise to a critical view on the method of triangulation: Methodological virtues of triangulation are often overestimated and it seems worthwhile to pay more attention to the methodological risks of triangulation. Open Interviews, Ethnicity, Frame Analysis, Triangulation.
THE COMPATIBILITY OF STRUCTURAL EQUATION MODELLING AND CLUSTER ANALYSIS: AN EXAMPLE1
Timothy L. Seifert, Jeff Bulcock
A structural equation model was constructed to examine the effects of mathematic achievement and feelings about school on feelings towards self. Results indicated a good fit to the data. A cluster analysis of the two exogenous variables, math achievement and feelings towards school, resulted in the identification of seven subgroups. The model from the population did not fit the data within each cluster suggesting that different processes were at work within each subgroup. Parameter estimation within subgroups suggested indeed that the contributions of math and feelings towards school differed in each group and that different processes were being utilized. The conclusion was that cluster analysis may help clarify and refine structural equation models. Structural Equation Models, Cluster Analysis, Comparison of Methods, School Performance.
SOME REMARKS ON THE INTERPRETATION AND POSSIBLE USES OF THE “SOCIAL CAPITAL” CONCEPT WITH SPECIAL REGARD TO THE HUNGARIAN CASE1
The author examines the various definitions of the concept of capital and then tries to see what the concept of social capital implies. He proposes three different definitions which are analyzed and applied to existing survey data in Hungary. Capital, Social Capital, Survey Data, Hungary.
PANEL COMPARABILITY ET LA BASE DE DONNEES PACO
Panel Comparability and the PACO Data Base. The CEPS/INSTEAD has been coordinating the European Commission-financed “Panel Comparability (PACO)” project which involves national longitudinal surveys carried out in Luxembourg, France, Germany, Great Britain, and the United States. The authors address the numerous methodological problems involved in setting up such a data base which permits the comparative study of panel survey data from different countries. Panel Data, Comparative Panel Studies, Data Compatibility.
n. 54, Marche 1997
THE ISCO-88 INTERNATIONAL STANDARD CLASSIFICATION OF OCCUPATIONS IN CROSS-NATIONAL SURVEY RESEARCh1
Cross-national research on social stratification has to be based on measures of social status that are comparable and interculturally valid. As most status measures are based on information about occupations, an internationally comparable taxonomy of occupations is of central importance. An especially useful classification in this respect is ISCO-88, the 1988 version of the International Standard Classification of Occupations. First, the conceptual framework of this classification is outlined and some basic differences between ISCO-88 and its predecessor, ISCO-68, are illuminated. Secondly, three versions, which depart from the original classification in one way or another, are presented. Finally, Ganzeboom’s and Treiman’s status scales based on the new classification of occupations are introduced and the performance of the new scales relative to the old ones is assessed empirically. Occupation, Classification, ISCO, Status, Comparative Research.
IS CURRENT VALIDATION VALID? SOME INTRIGUING PROBLEMS WITH TRADITIONAL VALIDATION DESIGNS FOR GENERAL (VALUE-) SURVEYS1
In most methodology textbooks, it is argued that both content and criterion-related validity are very limited, if not impossible validation strategies for the social sciences. Construct validity, based on the ideas of Cronbach and Meehl, is proposed as an acceptable alternative. The first part of this article elaborates on the relation between causal models as a means of validation (the nomological network) and social theory. The second part of the article evaluates different aspects of the nomological network. In the third part, triangulation, as a way of validating survey-empirical research, is discussed. We argue towards the conclusion that construct validity remains limited as well. We contend that, to a large extent, the superiority of construct validity over other kinds of validation is overestimated. Validity, Construct Validity, Survey Methodology, Measurement Theory.
METHODES ET PRATIQUES FORMALISEES D’ANALYSE DE CONTENU ET DE DISCOURS DANS LA RECHERCHE SOCIOLOGIQUE FRANCAISE CONTEMPORAINE. ETAT DES LIEUX ET ESSAI DE CLASSIFICATION
Formalized Methods and Procedures in Content and Discourse Analysis in Contemporary French Sociological Research – State of the Arts and a Classification Attempt. An initial awareness is needed of the debates regarding the choice of research approaches in sociology and the diversity and specificity of methods currently being used in the domain of textual data analysis in France. In general the influence of the French socio-linguistic tradition looms large, including, on the one hand, the older works of Michel Pécheux on the “discursive formations” and his A.A.D. (Analyse Automatique du Discours, 1969), and on the other hand, two main perspectives of the “Ecole Française d’Analyse du/de Discours” – which refer to the “speech act” concept and to the problematics of enunciation, and emphasizes the processes and “sociodiscursive practices” between socially-located speakers. Such theoretical conceptions and specific requirements lead to build on methodologies different from the classic, theme-based content analysis, though not yet translated into an operational software. Then the main software developments currently having an impact (at least potential) on practices of computer-aided sociological analysis of textual data, in France, are classified : from the lexicometric using procedures of “French Data Analysis” (‘Analyse Factorielle des Correspondances’ of Benzecri, and so on…), to a set of “expert-systems” working on specific theoretical frameworks, through more classical methods of content analysis and coding-sorting-retrieving socio-semantic procedures, eventually with various statistical methods. Content Analysis Methods, Discourse Analysis Methods, French Sociological Research.
THE QUALITY OF THE QUANTITY: INFORMATION TECHNOLOGY AND THE EVALUATION OF DATA*
Robert C. Yamashita, Howard Besser, Troy Duster, Thomas Piazza, Michael Hout
One of the problems with survey data is the need to impute an interpretive frame for the data. Generally this frame is developed ad hoc (in the design of the instrument) and modified post hoc (in explaining findings, apparent contractions, and variances from ad hoc expectations). While statistical methods provide measures of data consistency and validity, they cannot interpret the understandings of respondents. This paper offers an outline for using information technologies to aid researchers in the interpretation of respondent data. One method of re-capturing the interpretive context of subjects is to analyze the natural “talk” that necessarily occurs when subjects respond to survey questions. The new information technologies provide mechanisms for doing this. For example, with surveys that use Computer Assisted Telephone Interview systems, with modifications to the underlying computer technologies and the background database, voice and text records of respondents can be captured and stored. These records can then be used in conjunction with the statistical record. While there are physical limits (i.e. disk space; time, etc.) to such exercises, the basic information derived from text and audio sources can provide important contextual material for the evaluation of numeric data. Simple technological processes can be used to help gain a proxy understanding of the interpretive context (especially in conjunction with statistical evaluation). The data can also be used to further refine the survey instrument and evaluate the performance of the interviewers. A two stage development trajectory is offered. The first consists of a relatively simple modifications to the existing technology using text records, and the second looks to more expensive and complicated development of the core technologies. Other possible applications of these technologies are then discussed, and the significance of the impact of these efforts are reviewed. Textual, Conversational & Numerical Data, Information Technologies, Qualitative & Quantitative Analysis.
n. 55, June 1997
TESTING SURVEY QUESTIONS: NEW DIRECTIONS IN COGNITIVE INTERVIEWING
The four papers which follow in this issue were originally presented as part of an invited session on Cognitive Methods in Questionnaire Design which I convened as part of the Fourth International Conference on Social Science Methodology, July 1-5, University of Essex, UK. This note provides an introduction to the topic and a context for the papers which follow. Questionnaire Design, Questionnaire Testing, Cognitive Interview.
PERSPECTIVES ON PRETESTING: “COGNITION” IN THE COGNITIVE INTERVIEW?1
Eleanor R. Gerber, Tracy R. Wellens
The aim of this paper is to discuss the cognitive think-aloud interview in historical perspective and to outline its uses and limitations in the context of questionnaire pretesting and development. This qualitative technique is traced from its early predecessors in psychology and its original uses in survey methodology. Such practices as retrospective probing, definitional probing, and think-aloud “practice” questions, have been added to the technique, and differ from study to study. The technique has also been adapted to different respondent populations, and to “quasi-sampling” schemes. Self-administered questionnaires also require adaptations of the original, entirely oral technique. In addition, the technique is being used by interviewers with varying degrees of training. The implications of these variations in the conduct of the interview are explored. The definitions of “cognition” and “cognitive processes” which are implicit in these practices are also discussed. We examine how these definitions have expanded to encompass such things as language usage, language interpretation, and background knowledge. The usefulness and necessity of theory-driven analysis are also discussed. Cognition, Survey Methodology, Cognitive Interviewing.
EVALUATION OF COGNITIVE INTERVIEWING TECHNIQUES: DO THE RESULTS GENERALIZE TO THE FIELD?
Gordon B. Willis, Susan Schechter
Cognitive interviewing techniques have been widely used to pretest survey questionnaires. However, to date there has been little work to evaluate the usefulness of these techniques. This paper reviews critical issues relating to the demonstration of validity of cognitive interviewing methods, and proposes that researchers ascertain the degree to which the findings of this testing are applicable to field survey situations. To demonstrate this approach, we examined five survey questions that had been tested in a cognitive laboratory, and based on the results of this testing, made predictions concerning the manner in which modifications to these questions would be reflected in the data obtained under conditions of field administration. In three split-ballot survey experiments that administered alternate question versions, it was found that predictions based on cognitive testing were supported. The results provide support that there is significant carry-over between cognitive laboratory and field environments. Survey Pretesting, Cognitive Techniques, Evaluation.
METHODOLOGICAL ISSUES SURROUNDING THE APPLICATION OF COGNITIVE PSYCHOLOGY IN SURVEY RESEARCH
In the rush to apply cognitive psychology, and its methods, in survey research over the last decade, not enough attention has been given to scientific principles. This paper provides a framework for correcting this problem by focusing on methods to improve both the validity and the reliability of data gathered using cognitive psychology. These methods include better experimental designs and measurement techniques. Experimental designs which facilitate comparisons between alternative cognitive procedures and different experimenters are presented. Another feature of these designs is that they will require more care in the development of the experimental protocol. As for measurement, techniques for making the use of qualitative data more systematic will be discussed, and methods for constructing ordinal and interval indicators will be offered. Experimental Design, Survey Methods, Questionnaire Design, Measurement Error.
COMPUTER-ASSISTED QUALITATIVE INTERVIEWING: A METHOD FOR COGNITIVE PRETESTING OF COMPUTERIZED QUESTIONNAIRES*
At Statistics Netherlands, almost all household surveys are computer-assisted. Within this setting, in the Questionnaire Design Resource Centre of Statistics Netherlands a computer-assisted method for pretesting computerized questionnaires in a cognitive laboratory was designed: Computer-Assisted Qualitative Interviewing (CAQI). A CAQI protocol is expressed by instruction screens and probes built around the questions that are to be tested. Our experience is that CAQI is a workable method, resulting in good qualitative information about how questions are processed and how answers are obtained by respondents. Pretesting Questionnaires, Cognitive Interviews, Computer-Assisted Interviewing.
n. 56, September 1997
INTERACTION CHARACTERISTICS IN SOME QUESTION-WORDING EXPERIMENTS
Interaction characteristics in some question-wording experiments. Data from interviewer-respondent interaction analysis are seldom used in other types of research about data quality. Using data from an interaction analysis that was carried out on taped interviews in an experimental question-wording research, it was possible to validate some of the explanations about certain effects produced by the way the questions are worded. Question-wording experiments about the range of the response scale, the use of the explicit “no opinion” filter, and the use of the verbs “forbid” or “allow” are used. Question-Wording Experiments, Interviewer-Respondent Interaction, Interaction Analysis.
REPRESENTATION AND CONFRONTATION OF THREE TYPES OF LONGITUDINAL NETWORK DATA FROM THE SAME DATA BASE ON SOCIOLOGICAL AIDS RESEARCH
Karl M. van Meter, William A. Turner
Our previously published work on the analysis of sociological AIDS research between 1980 and 1990, based on entries in Sociological Abstracts, has shown that the evolution of research themes, authors, and journals in this domain are quite distinct from each other. Indeed, they each form a separate set of longitudinal network data generated from the same data base. The research themes seem to divide up into solid, less flexible themes which change little with time, and those that are more flexible, even “adaptive” and change over time. The set of authors also seems to have this aspect of division between those that move in a clearly designated direction and those that change over time. The third type of longitudinal network data is journals in which this research is published. In an expanding field such as sociological AIDS research between 1980 and 1990, some older journals try to “adapt” and compete with new thematic journals which appear. Other journals try to dominate a specific aspect of the field. Each of these three types of data generates its own problems of analysis and results. Moreover, there is the additional problem of confronting and synthesizing these different results which are all based on the same original data set. AIDS, Network Analysis, Longitudinal Analysis, Journals, Authors, Themes.
THE CHALLENGE OF (LONGITUDINAL) INTRA-ORGANIZATIONAL SOCIAL NETWORK RESEARCH (IN THE NETHERLANDS)1
Gerhard G. van de Bunt
Some problems are discussed that a researcher encounters when he or she decides to study intra-organizational social networks (over time). The paper is based on personal experiences of a longitudinal network research in a medium sized general hospital in the Netherlands. The main conclusion is that the building of trust is by far the most important determinant of success (in terms of response rate). This takes time, energy, and patience, but is worth the effort. Finally, the respondents’ reactions afterwards are summarized, and the subject of anonymity is discussed. Intra-organizational Social Networks, Longitudinal Analysis, Data Collection Problems. Anonymity.
SOCIAL SCIENCE MICROSIMULATION
G. Nigel Gilbert, Klaus G. Troitzsch
Below are three papers based on presentations and discussions at the Dagstuhl Seminar on Social Science Microsimulation: “A Challenge to Computer Science”, Schloß Dagstuhl, May 1-5, 1995, that have since then been published in K. G. Troitzsch, U. Mueller, G. N. Gilbert and J.E. Doran (editors), Social Science Microsimulation, 1996, Springer Verlag. The three papers are “Simulation as a Research Strategy” and “Environments and Languages to Support Social Simulation: Summary of an Informal Discussion”, by G. Nigel Gilbert, and “Computer Simulation and Social Sciences: On the Future of a Difficult Relation — Summary of an Informal Discussion” by Klaus G. Troitzsch. Simulation Research, Computer Simulation Programs, Simulation and Social Theory.
n. 57, December 1997
APPLICATIONS OF COMPUTER SOFTWARE IN THE SOCIOLOGICAL ANALYSIS OF QUALITATIVE DATA
Nigel G. Fielding, Raymond M. Lee
This article profiles the use of computer-assisted qualitative data analysis (CAQDAS) in sociology. It reviews the development of CAQDAS, relates software types to different approaches to the sociological analysis of qualitative data, examines the place of CAQDAS in the research environment, and comments on future developments in software for qualitative data analysis. The article is illustrated by examples of working procedures in particular software packages, and is informed by a fieldwork-based study of the experiences of qualitative researchers with CAQDAS. The article concludes with a ‘resources list’ showing details of a number of CAQDAS packages. Qualitative Methods, Computer Software, Qualitative Data Analysis, Content Analysis, Discourse Analysis, Methodological Innovation.
APPORTS DE LOGICIELS D’ANALYSE DE DONNEES TEXTUELLES (ADT) DANS LES PROCEDURES D’ANALYSE DE CONTENU D’ENTRETIENS SEMI-DIRECTIFS DE RECHERCHE: ALCESTE ET HYPERBASE
Two Software Packages for Computer-Assisted Textual Data Analysis in Research Involving Semi-Directive Interviews: Alceste and Hyperbase. This article discusses the contribution of two textual data analysis computer programs, Alceste and Hyperbase, to producing results that can be communicated, verified and contested by other researchers. The two programs perform a uniform and extensive analysis of any corpus and quickly identify the points of resemblance and difference between the discourse of various individuals or groups of individuals. The programs’ limits are specified, but the programs provide clear support in the description of a corpus and the interpretation of discourse recorded during research interviews. Textual Data Analysis, Semi-Directive Interviews, Alceste, Hyperbase.
THE RELATION BETWEEN QUANTITATIVE AND QUALITATIVE SOCIAL RESEARCH1
Starting from a schematic characterisation of the history of the debate about qualitative and quantitative social research methods, a six point argument is presented. The six steps involve drawing attention to (1) the multiple paradigms available to social researchers, each characterising explanation and data in essentially opposed ways; (2) the variety of research procedures adopted to collect or construct social data; (3) the key point that paradigms do not entail procedures nor vice versa; (4) the similarities between qualitative and quantitative data in that both are extracted or constructed from the richness of lived experience; (5) the point that although the nature of data must be appropriate to the paradigm framing the research nevertheless the procedures adopted to generate data of this form are in principle unbounded; and (6) how confusions arise when it is assumed that particular paradigms and particular procedures necessarily fall into logically coherent and mutually exclusive sets. Parallels are drawn between the six points and recent arguments by Hammersley and Smaling. The conclusion is that acceptance of the six-point argument allows researchers to be more imaginative in their selection of methods. Qualitative Methods, Quantitative Methods.
`NEW AGE’ TRAVELLERS, URBAN SLUM DWELLERS, ABORIGINES AND DRUG USERS: EXPERIENCES OF COLLECTING SENSITIVE DATA FROM MARGINALISED COMMUNITIES
Zoe Matthews, Richard Velleman
This paper draws together common themes from the experience of data collection in two current studies. These research projects have aimed to increase understanding of behaviour and attitudes within marginalised communities and have all dealt with sensitive issues; one with sexual health (of ‘new age’ travellers in south-west Britain) and the other with the experience of families within which drug and alcohol problem exist (comparing aborigines in Australia, the urban poor of Mexico City and problem drug and alcohol users in South West Britain). The hostility of some community members to the research process has focused attention on the most appropriate methods of data collection, and the issue of using insider researchers has been raised both as a strength of study designs, and also as a problem area. The paper describes the methods of data collection for these studies, including sample selection, instruments, and composition of research team. A comparison of the experiences of the respective research workers is also presented, in the context of the cultural environments of the three study populations. Although the settings for these studies are geographically diverse, the problems of data collection experienced have shown some methodological common ground which may be valuable for other researchers working with excluded populations whose culture may be antithetical to being researched. New Age Travellers, Slum, Urban Poor, Aborigine, Drug, Alcohol, Sexual Health, Data Collection, Marginalised Community, Exclusion, Anthropological Approach.
n. 58, Marcj 1998
AN EVALUATION OF SOME CLUSTERING METHODS FOR MIXED MODE VARIABLE DATA SETS*
Istvan Hajnal, Geert Loosveldt
Clustering, when objects are measured on a mix of nominal, ordinal or numerical variables, has always been a problem in classification research. This article briefly reviews some clustering methods for dealing with mixed mode variable data sets. These methods include the coefficient of Gower, the coefficient of Kaufman and Rousseeuw, the combined resemblance matrix approach of Romesburg, the method of Everitt and Merette, and finally Groupals. Some of the methods were compared in a simulation study. We used a conditional Gaussian mixture approach to generate artificial mixed mode variable data sets with a known clustering structure. We conclude that in our simulation set up that Gower’s coefficient performed less well than Groupals and the use of binary variables. Clustering Methods, Mixed Mode Data, Comparison by Simulation.
ROBUST STRATEGIES FOR APPLICATION OF CLUSTER ANALYSES
In this article, cluster analysis is used to identify and describe homogeneous social and economic clusters using geographical regions in Alaska as an example of a transitional society. Classifications of underlying social and institutional contexts are identified and compared using several clustering methods. The analyses are performed on a data set with empirical characteristics of twenty-three census places (equivalent to counties) in Alaska in order to evaluate their underlying structures. The author then examines cluster results for substantive validity and empirical distinctiveness. Cluster Analysis, Comparison of Methods, Social and Economic Homogeneity, Alaska, Transitional Society.
MEASURING COLLECTIVE IDENTITIES OF CANADIANS: A NESTED CONCENTRICITY MODEL*
The paper examines the measurement by structured survey questions of collective identities based on territory. It is a measurement issue that has frequently been recognized in the literature as difficult. Data come from a 1994 mailed questionnaire administered in a region within the Province of Ontario, Canada. The survey has over 2000 cases with a 71% response rate. Identification with the entire globe, with Canadians, with Ontarians, and with residents of the local region was double measured within separate matrix format closed questions. The design allows valid measurement to be separated from method effect within each question. A LISREL approach is used to estimate correlation between latent variables for each of the four identities. This leads to the nested concentricity model, within which identities “closest” to each other (e.g., Canadian-Ontarian vs. Canadians-local region) are the most strongly related. Collective Identities, Canada, Latent Variables, LISREL.
DETECTING ABERRANT RESPONSE PATTERNS IN MULTI-ITEM SCALES: SOME NONPARAMETRIC INDICES AND A COMPUTER PROGRAM
Edith D. de Leeuw, Joop J. Hox
The psychometric literature contains many indices that are aimed at detecting aberrant c.q. deviant response patterns. This paper discusses three well-known nonparametric indices, gives an example of an application, and describes a computer program that calculates these indices. Person Fit Research, Appropriateness Research, Respondent Error.
n. 59, June 1998
ILO UNEMPLOYMENT AND REGISTERED UNEMPLOYMENT: A CASE STUDY
ILO (International Labour Office) unemployment in Britain is less responsive to changes in the level of employment than the number of registered unemployed. Keynesian-type registered unemployment (i.e., the Count of Claimants) focuses on male “bread-winners”, but the ILO seeking-work criterion used in the Labour Force Survey (LFS) captures entry to unemployment as well as volition to employment. The ILO/LFS Unemployment Series in Britain identifies the top layer of a reserve army of labour of changing composition but relatively constant size. Reconciliation of ILO and registered unemployment statistics is necessary for understanding of the dynamics of the labour market in Britain. The positive and negative interdependencies between ILO unemployment and employment may not be limited to Britain. ILO Unemployment, Unemployment Statistics, Reserve Army of Labour, Labour Force Survey, Longitudinal Studies, Social Security System, Discouraged Workers.
APPLICATION DE LA THEORIE DE REPONSE AUX ITEMS A L’ANALYSE D’ECHELLES D’ATTITUDE
Application of Item Response Theory to the Analysis of Attitude Scales. The article presents Item Response Theory (IRT) and illustrates its application to the analysis of two measures of work-related attitudes — value given to social aspects of the job and satisfaction with extrinsic aspects of the job. Results of IRT analysis show that extrinsic satisfaction cannot be measured in the same way among contract professionals as opposed to regular employees. Using information provided by IRT analysis, it appeared that satisfaction with salary cannot be included in a scale measuring extrinsic satisfaction because it does not contribute to the measure in the same way for the two groups of employees; furthermore, satisfaction with social benefits may be included in the scale only if discrimination and difficulty parameters are allowed to vary across groups. Analyses of variance using theta scores estimated in this way give statistical conclusions that are notably different form those obtained using traditional additive scores and permit refinement of interpretation. Item Response Theory, Attitude Scales, Work-Related Attitudes, Contract Professionals, Regular Employees.
THE USE OF LOCAL NETWORKS IN A STUDY OF HEROIN USERS: ASSESSING AVERAGE LOCAL NETWORKS
Ingegerd Jansson, Marinus Spreen
In a study of a population of heroin users in a Dutch city, the local network approach was used to collect and analyse information about the social environment of the heroin users. Instead of applying a snowball sampling procedure to contact respondents by successive waves, the focus in this study was on collecting relational data with a one-wave snowball sampling procedure. The population was divided into two strata: heroin users attending, respectively, not attending, regular drug assistance sessions. From the group attending regular drug assistance sessions, a random sample was drawn; for the other unknown group, a non-random sample. Several aspects of representation reliablity are discussed when assessing in the social environment of the respondents, including the proportion of users visiting the regular drug assistance, of users visiting other drug assistance sessions, of users visiting both kinds of assistance sessions, and of users visiting no assistance sessions at all. Heroin Users, Relational Data, Snowball Sampling, Drug Assistance Programs.
CENTRE – COURSE SYLLABUS – EMPIRICAL KNOWLEDGE IN SOCIOLOGY: STUDYING RESEARCH PRINCIPLES AND APPROACHES, COGNITIVE OPPORTUNITIES OF METHODS AND SPECIFICITY OF OBTAINED RESULTS
This course was developed at the St. Petersburg State Academy of Culture during 1990-1998, based on empirical and methodological research over more than two decades. It is used for teaching students in 2-3 and 5 year undergraduate programs, and some topics are offered for post-graduate diplomas. The syllabus is devoted to communicating empirical knowledge in sociology. The present course involves teaching through a number of comparative contexts: in this case each of the components “holds” the other components together and, in addition, clarifies their sense. A student should constantly compare them and, in the end, will understand the specificity of each research dimension and meaning of the final empirical results in sociology. Within the present program, there are at least four “axes” that can be considered independently or in relation to each other. In particular, an important aspect concerning the existing three levels of empirical substantiation in sociology is presented here. The second system derives from understanding the structure of an empirical result in any research. The third one is connected with the research types. The fourth is about the sequence of research stages. Other important explanations are also given. Social Reality, Empirical Knowledge, Research and Results, Investigative Collision, Investigative Approach, Types of Research, Standardized Survey, Research Based on Texts, Classification of Texts, Primary Data, Methods of Data Actualization, Unit Selection, Data Analysis.
N. 60, October 1998
CONTACT AND STEREOTYPING IN A VOLUNTARY ASSOCIATION
Bonnie H. Erickson, T. A. Nosanchuk
Contact and Stereotyping in a Voluntary Association: Contact between social groups reduces prejudice and stereotyping – sometimes, depending on the nature of the contact and the setting. Voluntary association settings have been little explored, though they should often meet the contact hypothesis scope conditions. We analyze a large association in which the conditions can be checked unusually thoroughly and the conditions are met. We also use unusually refined measures of contact (derived from network analysis) and of age and gender stereotyping (derived from work on belief systems) and detailed hypotheses (derived from social cognition theory). No form of contact reduces stereotyping. Greater involvement in the subculture actually increases stereotyping. We argue that attention is a function of rank, so that our respondents notice the inequality of a tiny elite and ignore the equality of the far larger majority. These show results that the cognitive approach to social stereotyping should be combined with work on social structure and its implications for contact, inequality and attention. Stereotypes, Contact, Inequality, Social Networks.
UNIFIED BHPS WORK-LIFE HISTORIES: COMBINING MULTIPLE SOURCES INTO A USER-FRIENDLY FORMAT
Longitudinal data is often difficult to use, and continuous histories collected in a panel are a particularly unfriendly case. This paper reports an exercise to re-organise the British Household Panel Study’s work-life history data into a format more convenient for analysis. The British Household Panel Study collects extensive labour market history information from its respondents, both during the panel period and retrospectively from labour market entry. That this information is of necessity stored in multiple locations, and of varying levels of detail, has made use somewhat inconvenient. This paper describes an exercise to bring the labour market information together in a more convenient format. It also considers some of the problems of retrospective and panel longitudinal data, and discusses issues of recall error and measurement error. The data files described are available through the UK Data Archive. Longitudinal Data, Work-Life Histories, Recall Bias, Measurement Error.
THEORY-DRIVEN INTERVIEWING: FROM THEORY INTO PRACTICE
Niall Hamilton-Smith, Matt Hopkins
This paper considers two approaches to utilising Pawson’s theory-driven interview model as an attempt to overcome the traditional schism between quantitative and qualitative interview techniques. Both pieces of research are criminologically based though methodologically different. The first employs Pawson’s model in the development of mid-range theoretical propositions relating to the ’embedding’ of distinct policy initiative (phase II, Safer Cities). The second seeks to test a theory about the reasons behind contextual variations in acts of abuse and violence against business premises. Through these examples, the paper attempts to develop a critical account of the operationalisation of this model in social science research. Interviewing, Epistemology, Theory-Driven Interviewing, Pawson.
STRATEGIE DISCURSIVE DE PROFESSIONNELS DE LA REPRESENTATION
The Discursive Strategies of Professional Union Representatives. This study examines arguments concerning the replacement of labor union activists by professional labor union employees, a modification supposedly associated with the current crisis of labor unions. The job position of labor union organizer is analyzed through its historical development. An empirical research in the field with two labor federations was carried out to reconstitute the stages of a union career. The central part of the recherche consists of the production of interviews with union representatives. This social-historical approach reveals the fragile and unfinished character of the professionalization of labor union management. Careers, Identity, Trade, Professionalization, Full-Time Unionists, Union Members, Union Representatives.
N. 61, January 1999
AN INTUITIVE INTRODUCTION TO BLOCKMODELING WITH EXAMPLES
This paper presents an intuition-based formulation for blockmodeling. It starts with an informal discussion of roles, relations and equivalence and outlines the idea of a blockmodel as a model of role structures. The idea of a social network is then considered and these two ideas are joined to suggest the notion that blockmodeling ideas can be used to partition a social network. This is followed by a discussion of the direct approach to establishing blockmodels. Two empirical examples are considered in some detail. One natural generalization of blockmodeling is considered and illustrated. Blockmodeling, Role Structures, Partitioning Social Networks.
CLASSIFICATION, CROSS-CLASSIFICATION ANALYSIS AND GENERALIZED BLOCKMODELING
Karl M. van Meter
Blockmodeling has been closely associated with social network analysis and often works with symmetric individuals-by-individuals matrices where the unit cells describe ties between two individuals. But similar, or even equivalent, blocks can be obtained by classification analysis. Moreover, cross-classification analysis — the Cartesian crossing of a classification of the individuals with a classification of the descriptive variables — permits the generalization of block construction and also identifies the statistically most significant blocks in a data set. Classification Analysis, Blockmodeling, Cross-Classification, Polythetic Classes.
COMMENT ANALYSER LES DONNEES CHRONOLOGIQUES POUR PLAN PARTITIONNE EN SCIENCES SOCIALES
Michel Clavet, François Petry, Jean-Sébastien Brien
How to Analyze Time-Series Cross-Section Data in the Social Sciences. In this paper we present, step by step, a SAS statistical program for the analysis of Time-Series Cross-Section (TSCS) data which gives robust and unbiased regression estimates. Several statistical packages (SHAZAM, SAS, SPSS) offer procedures for the analysis of TSCS data based entirely, or in part, on a particular application of the Generalized Least Squares (GLS) method developed by Parks and Kmenta. In a series of recent articles, Beck and Katz have shown that the GLS estimation method often underestimates substantially the standard error of regression parameters in TSCS models. They suggest a new method for generating adequate correction of the error process in TSCS models. This method consists of using Ordinary Least Squares (OLS) parameter estimates first, and then replacing OLS standards errors (which are biased) with unbiased Panel-Corrected Standard Errors (PCSEs) obtained after correcting both temporal autocorrelation and spatial correlation structures. The paper provides a SAS application of the method suggested by Beck and Katz. The SAS program can be downloaded through our web site. Time-Series Cross-Section (TSCS) Data Analysis, Standard Errors, Regression Parameters, Panel-Corrected Standard Errors.
SOLARIS – L’EXPERIENCE D’UNE REVUE SCIENTIFIQUE SUR INTERNET SOLARIS
Ghislaine Chartron, Jean-Max Noyer, Sylvie Fayet-Scribe
Solaris, The Experience of a Scientific Journal on the Internet. The development of the Internet has involved experiments in scientific communications and, in a more general context, scientific journals are currently confronted with the growth of computer-aided communications and documentation. In this article, we give a brief history and describe aspects of the development of Solaris <http://www.info.unicaen.fr/bnum/jelec/Solaris/>, a French scholarly journal (which also publishes some English articles) in information and communications sciences. After studying the development of other electronic scientific journals, we can identify certain major trends concerning the founders, the financing, the submission of articles, the services available, the citations and other developments. Solaris, Internet, Scientific Publishing.
N. 62, April 1999
NONRESPONSE TO MAIL SURVEYS IN A LOWER-CLASS URBAN AREA – A TWO-STAGE EXPLORATION OF ACCESS FAILURE AND REFUSAL
Harrie Jansen, Tony Hak
Different sources of nonresponse result in different types of bias. In mail surveys, it is usually unknown which proportion of nonreturns must be attributed to access failure and which to refusal. But also within these two categories, further distinctions should be made. In this study, we distinguish between locating failure, contact failure, contingent nonresponse, explicit refusal, hidden refusal, generic refusal, and specific refusal. In a two-stage study of the nonresponse in a lower-class urban district to a mail survey on lifestyle and risk behavior (drinking, smoking, drug-taking, gambling), we first attempted to assess empirically the different components of nonresponse by means of a telephone and face-to-face survey. We found unexpectedly high proportions of localization failure and self-reported generic refusal. In the second stage, a year after the first one, we sent a small respondent-friendly questionnaire on an extremely salient topic to the same sample to assess the stability of response behavior and the validity of self-reported responsiveness. The main findings of this second stage were, first, that self-reported responsiveness did not predict actual response behavior and, second, that, in this lower-class urban district, even the most respondent-friendly and salient survey does not yield a response rate higher than approximately 70%. As it seems impossible to increase response rates above the level that is determined by generic refusal and contingent nonresponse, we conclude that more attention should be given to the development of methods that address avoidable contingent nonresponse. We propose that the success of such attempts to increase response, and the remaining avoidable response bias, be assessed by means of a small respondent-friendly “test” survey on the same topic. Nonresponse, Mail Survey, Differential Accessibility, Differential Responsiveness.
HOW DO SUCCESSFUL AND LESS SUCCESSFUL INTERVIEWERS DIFFER IN TACTICS FOR COMBATING SURVEY NONRESPONSE?
Edith D. de Leeuw
Interviewers play an important role in gaining respondent cooperation in surveys. There is empirical evidence of considerable variation in response rates between interviewers. Despite this, research on the role of the interviewer in nonresponse is relatively scarce. Past research has shown that interviewer experience plays a role in gaining respondent cooperation, and, recently, the importance of the “doorstep” interaction has been emphasized. In an earlier study, we gave an inventory of the tactics reported by experienced interviewers to combat nonresponse. In this paper, we investigate the relationship between favored interviewer tactics and individual interviewer response rates. The effective interviewers emphasized the importance of a combination of basic rules (identify yourself, mention agency almost at once) with more advanced tactics (tailoring, adapting introduction). And above all, they believe in themselves and emphasize the importance of social skills and self-confidence. Those who obtain a lower response rate do not rate the “basic rules” as important, nor do they attach much importance to efficient tailoring and self-confidence. Doorstep Interaction, Persuasion, Survey Participation, Interviewer Strategies, Response Rate.
WHY DID THE POLLS GO WRONG IN THE 1998 QUEBEC ELECTION? THE ANSWER FROM POST ELECTOIN POLLS
Claire Durand, Andre Blais
This article presents the results of a post-election survey conducted among respondents to three pre-election surveys carried out during the last electoral campaing in November 1998 in Quebec, Canada. The surveys, published during the electoral campaign, systematically underestimated vote intentions for the Liberal Party, a center-right federalist party, and overestimated vote intentions for the Parti Quebecois, a center-left sovereignist party, this despite the generally “orthodox” methodology used by Quebec pollsters. The discrepancy between the vote and pre-election voting intentions had been attributed to either a late-campaign shift in favor of the Liberal Party, or a lower participation in the vote among Parti Quebecois voters. The results of this study permit the rejection of both hypotheses. Instead, the discrepancy stems from biases related to the sampling frame and from the different voting behavior of non-respondents to electoral surveys. Post-Election Polls, Surveys, Elections, Sampling Bias.
COMPUTER-ASSISTED INTERVIEWING: THE DESIGN AND APPLICATION OF SURVEY SOFTWARE TO THE WIRED SUBURB PROJECT
Keith N. Hampton
This paper explores the use of Internet and personal computer-based interviewing in the University of Toronto’s Wired Suburb (Netville) Project. The use of computer-assisted interviewing (CAI) in this project differs from other examples in its use of social network questions, of a time-use diary, and of Internet (Web) and personal computer (PC) based interviewing of a small residential population. The purpose of this paper is to develop an understanding of why CAI, specifically Computer Assisted Personal Interviewing (CAPI) and Computerized Self-Administered Interviewing (CSAI), may be more appropriate for some research projects than others, to explore specific problems with the technology and approach used in this study, and to explore specific challenges for the use of CAI in social network and time-use analysis. Internet-Based Interviewing, Personal Computer-Based Interviewing, Computer-Assisted Personal Interviewing (CAPI), Computerized Self-Administered Interviewing (CSAI), Netville Project.
N. 63, July 1999
CLOSED STRUCTURES, OPEN STRUCTURES, STABLE STRUCTURES: EXPLAINING STRUCTURAL FORM AND TEMPORAL STABILITY OF INFORMAL SOCIAL NETWORKS IN ORGANIZATIONS
Hypotheses that address the structural form and stability of informal social networks in organizations are derived from organization theory and dynamic network analysis. An exploratory empirical test of these hypotheses is carried out by reanalyzing a longitudinal network study of trust and distrust relations conducted in a retail sales store in the U.S. during the 1950s. Two problems are addressed. First, it is argued that degree of network closure is a function of sharing arrangements as they emerge from the functional interdependencies in work groups. It is shown that the effects, in part, counteract the processes that would be expected under the more common model of structural balance. Second, it is asked why certain informal network structures remain stable even under conditions where one would expect them to be most prone to dissolution: where the formal incentive structure strongly favors competition rather than cooperation, and where the informal structure, at the level of individual choices, reflects a pattern of restricted rather than generalized exchange. It is demonstrated that the stability of the network under study is significantly enhanced through the presence of a position of third-party intermediaries within the role structure of the system. The results highlight the importance of third parties, occupying the role of intermediaries for the stability of informal social networks and put in perspective the exchange theoretical assumption of the fragile nature of restricted exchange. Intraorganizational Networks, Exchange Theory, Dynamic Network Analysis.
FACTEURS SOCIO-CULTURELS ET PRESENTATION DE SOI DANS DIFFERENTS CONTEXTES D’ENQUETE: ANALYSE D’UN EXEMPLE DE DISCORDANCE
Social Cultural Factors and Presentation of Self in Different Survey Contexts: Analysis of an Example of Disagreement. A qualitative study among homosexual and bisexual men, who participated in a quantitative study, allows the author to show the effect of survey technique individual responses. Fifteen homosexual and bisexual men, who participated in the French national Gay Press 1995 survey on life styles and homosexual men’s reaction to AIDS, were chosen on the basis of their declaration of sexual exposure to risks of HIV infection in the survey’s self-administered questionnaire. During the interviews, the replies of a subgroup of those selected proved to be different from responses recorded on the self-administered questionnaire. The analysis of the differences and the comparison of the conflicting and coherent responses shows the impact of social cultural factors on the possibilities of self presentation that is more or less independent of survey context. Survey Techniques, Quantitative/Qualitative Disagreement, Sexuality, Social Cultural Factors.
AN EVALUATION OF A TYPOLOGY OF RESPONDENTS WITH A MULTILEVEL-MULTINOMIAL LOGIT MODEL
Jan Pickery, Geert Loosveldt
In this article, we evaluate a typology of survey respondents, based on the report of the interviewer. In his or her report, the interviewer evaluates the respondent’s motivation and ability to answer questions. In a previous article, this information was used to create a typology with three types of respondents. The first type has ample cognitive skills and is highly motivated. The second type resembles the first type, but the ability and motivation of the respondents of this type is slightly lower. The third type has problems doing what is expected of the respondent. It was possible to validate this typology empirically by relating it to background characteristics of the respondents and interview data quality indicators. We validate this typology further with a multilevel analysis, thus controlling for interviewer bias. Multilevel analysis takes into account the hierarchical structure of survey data and offers accordingly the best possibilities to analyze interviewer effects. We use a multilevel-multinomial logit model, with respondent characteristics and answering behavior as independent variables, the typology as dependent variable and the interviewer as a level 2 identifier. Our results show significant interviewer variability, which raises doubts about the usefulness of the report of the interviewer to create a respondent typology. Multilevel Analysis, Interviewer Effects, Response Behavior, Respondent Typology, Multinomial Logit Model.
NEW TECHNOLOGIES IN SOCIOLOGICAL RESEARCH: BMS’ ON-LINE EXPERIENCE, ISA’S EMAIL QUESTIONNAIRE & THE COMPUTER-LITERACY GAP
Karl M. van Meter
In its last two issues, the Bulletin de Méthodologie Sociologique (BMS) published presentations from the Special Session, “New Technologies in Sociological Research, Documentation, Publishing and Teaching”, of the 1998 World Congress of Sociology, organized in Montreal, Canada, by the International Sociological Association (ISA), which also organized a Publications Committee Special Session on related topics. In this article, I summarize the four remaining presentations of the “New Technologies” session and describe the results of related work presented in the Publications Committee session, including results of an early BMS-ISA Internet survey by questionnaire. Sociology, Sociological Research, Internet, New Information Technologies.
N. 64, October 1999
MODELE LINEAIRE CONTRE MODELE LOGISTIQUE EN REGRESSION SUR DONNEES QUALITATIVESModèle linéaire contre modèle logistique en régression sur données qualitatives
The Linear Model Versus The Logistic Model in Regression on Qualitative Data. Regression on qualitative data is usually done by using a logistic model. By examining data where “everything is otherwise equal”, one can show that the linear model is quite compatible with this type of data. Results of an in-deep analysis of cross-tabulation data (called tabular analysis), and of regression using linear and logistic models, are quite similar (in terms of percentage effects). The theoretical question concerning the possibility of a linear model remains to be examined: it seems that the specific situation of a reference category introduces constraints which make the linear model possible. Qualitative Data, Tabular Analysis, Logistic Regression, Linear Regression.
METHODOLOGICAL ISSUES IN SURVEY RESEARCH: A HISTORICAL REVIEW (1)
Wim de Heer, Edith D. de Leeuw, Johannes van der Zouwen
In this paper, we present a historical overview of social surveys and describe the historical development of scientific survey methodology and survey statistics. The origins of survey research can be traced back to the early 19th century and the first scientific survey was conducted in England in 1912 by Bowley. Modern survey methods have their roots in the 1930’s, when three major aspects of the survey process were developed and refined: sampling techniques, data collection techniques, and statistical methods for data analysis. In the fifties, the scientific face-to-face interview was established, and handbooks were published. In the early seventies, telephone interviewing was introduced and mail-survey methods were updated and refined. This was followed by the development of computer-assisted interviewing and electronic surveys in the eighties and nineties. At the same time, more emphasis was given to concepts such as measurement error, nonresponse, and total quality control. We are now at the brink of a new millennium. Although the basic principles of the scientific survey still hold, survey statisticians from the last millennium, like Bowley, would be surprised to see the design of a survey in 2012. Undoubtedly, this design will incorporate new technologies and will focus on reducing the respondent’s burden, while improving data quality. Bowley would certainly be pleased to see that sophisticated methods for the reduction of nonresponse and measurement error have been incorporated in a total quality survey design! Survey Methods and Statistics, Social Survey, Scientific Survey, New Technologies, Data Quality, Nonresponse, Respondent Burden, TQM, CBM.
LES “RAFALES”, UNE METHODE POUR IDENTIFIER LES DIFFERENTS EPISODES D’UN RECIT: CONTRIBUTION AU TRAITEMENT ET A L’INTERPRETATION DES ENTRETIENS NON-DIRECTIFS DE RECHERCHE Les , une méthode pour identifier les différents épisodes d_un récit : contribution au traitement et à l_interprétation des entretiens non-directifs de recherche.(1)
Mathieu Brugidou, Pierre Le Quéau
“Spurts” (rafales), A Method for Identifying the Different Episodes in Speech – A Contribution to the Analysis and Interpretation of Empirical Open-Ended Interviews. This article describes research with a method for the analysis of open-ended interviews. The identification of a series of often-repeated terms during a short period (“spurts” or “rafales”, in French) allows us to reveal “packets of meaning” which constitute episodes of speech. By following these spurts of terms, it seems possible to identify thematic breaks (the end of a series of spurts) and the frontiers where the themes of conversation unravel while a new theme (or speech episode) develops (a new series of spurts). The algorithm for identifying spurts has been integrated into the TROPES semantic analysis software and we attempt to show the contribution it can make in the analysis of open-ended interviews. This article uses and compares the lexicometric approach (spurts based on graphic patterns) and the semantic approach (spurts based on equivalence classes of terms with similar meanings). Analysis of Open-Ended Interviews, Speech, Structural Analysis of Speech, Discourse Analysis, Lexicometry, Spurts (Rafales).
HOW TO USE SPSS TO STUDY EGO-CENTERED NETWORKS (1)
Christoph Müller, Barry Wellman, Alexandra Marin
SPSS’s data-handling facilities make it useful for studying ego-centered networks. Such networks are useful for studying a variety of subjects, such as social support, personal community, and an organization’s relationships. We show how analyses of ego-centered networks are best done by starting with two data sets: (1) focal individuals and their ego-centered networks; (2) network members and their ties with focal individuals. We show how to link these two data sets to: (a) calculate summary information about each ego-centered network; (b) combine focal individual, tie and network data. SPSS, Ego-Centered Networks, Focal Individuals.
N. 65, January 2000
THE GEOMETRIC ANALYSIS OF QUESTIONNAIRES: THE LESSON OF BOURDIEU’S LA DISTINCTION 
Henry Rouanet, Werner Ackermann, Brigitte Le Roux
This paper investigates the use of Correspondence Analysis (CA) in Bourdieu’s La Distinction (1976/79), having in mind the sociologists who wish to analyze their data in a way in harmony with that of Bourdieu. For Bourdieu, CA is not simply a handy tool among others for visualizing data, but a unique instrument apt to uncover the two related spaces of individuals and of properties. A careful reading of La Distinction reveals a thoughtful and creative use of CA, applied to Individuals x Properties tables, with its main aids to interpretation. The major conclusions from this “lesson” remain valid today and generally apply to Individuals x Properties tables whether analyzed by CA or more specifically by Multiple Correspondence Analysis (MCA). In the analysis of questionnaires, doing correspondence analyses is not enough to do “analyses à la Bourdieu”. The fundamental space must be constructed from a set of relevant variables ample enough to allow the full multidimensional display of individuals. Bourdieu’s La Distinction, Geometric Data Analysis, Individuals x Properties Tables, Correspondence Analysis, Cloud of Individuals, Simultaneous Representation.
THE USE OF MULTIDIMENSIONAL PARTIAL-ORDER SCALOGRAM ANALYSIS WITH BASE COORDINATES (MPOSAC) IN PORTRAYING A PARTIALLY-ORDERED TYPOLOGY OF CITY WARDS BY SOCIAL-MEDICAL CRITERIA The use of Multidimensional Partial-Order Scalogram Analysis with base Coordinates (MPOSAC) in portraying a partially-ordered Typology of City Wards by social-medical criteria 
Shlomit Levy, Reuven Amar
Scientific problems, especially in the social sciences, are usually of a multivariate nature. In particular, a frequently encountered variety of typological problem is that which results when a given population is classified simultaneously by several criteria, each being ordered in a sense common to all criteria. The new technique of Multidimensional Partial-Order Scalogram Analysis with base Coordinates (MPOSAC) is proposed for presenting a partially-ordered typology of city wards, relating social-residential factors to a medical problem. The data consist of seven variables characterizing each of the 21 wards of the city of Hull, England. Four variables are social criteria and three are rates of incidence of infectious diseases. Each of the seven variables is ordered from high to low well-being. Two dimensions were insufficient to represent the partially-ordered typology. It turned out that three dimensions were needed for a satisfactory solution. Two dimensions were generated by the social variables differentiating between “overcrowding” and “no inside toilets” as base coordinates, while the third dimension, namely the third base coordinate, partitions the three-dimensional space into layers according to the infectious diseases. Structuple (profile), Order Comparable, Order Non-Comparable, Partial Order, POSAC, MPOSAC, Base Coordinates, Item Diagram.
UNE METHODE D’ANALYSE DES SEQUENCES Une mÚthode d’analyse des sÚquences
A Sequence Analysis Method. Data that can be expressed as successive states — temporal or not — of variable length, depending on the subjects, can be described as “sequences”. The analysis of such entities should allow for their automatic classification in categories as homogeneous as possible, but also the most differentiated possible, while allowing for their interpretation via an ideal-type characterization of the categories. Secondly, such an analysis should allow for differentiated weightings of the states and their separations to avoid the null hypothesis of semantic equivalence of all states and all separations. A mathematical model of accumulated density matrices, integrated into a software program, permits us to attain these objectives simultaneously and, moreover, results in a method of colored graphic representation of categories of sequences. The data used in this example come from research on the professional trajectories of 520 continuing-education teachers which used an earlier version of this method. Sequences, Longitudinal Data, Classification, Ideal Types, Graphic Representation.
VERBATIM: UNE EXPERIENCE DE CAPITALISATION D’ENTRETIENS QUALITATIFS
Dominique Le Roux, Jean Vidal
Verbatim, An Experiment in Capitalizing on Quantitative Interviews. Archiving qualitative data for use in secondary analysis is an established part of Anglo-Saxon sociology. In this article, we present encouraging preliminary results of an experiment in this domain in France and carried out in a business environment. Verbatim, Qualitative Data Archiving, Secondary Analysis.
N. 66, April 2000
REMARQUES SUR LA COMPARAISON ENTRE LES MODELES LINEAIRE ET LOGIT
Emmanuel Aris, Jacques Hagenaars
On the Comparison between Linear Models and Logit Models. The aim of this research note is to provide additional information about the comparison between the linear model and the logit model for a categorical dependent variable presented by Ph. Cibois in the BMS (n. 64, October 1999). Special attention is paid to the conceptual differences between the two models. Some relevant literature about models used with categorical dependent variables is also given. Categorical Dependent Variables, Logit Model, Linear Model, Interaction Effect.
CURRICULUM VITAE ET CONNAISSANCE PREALABLE DES PERSONNES: LEUR INTERET POUR LA CONDUITE DES ENTRETIENS BIOGRAPHIQUES
Curriculum Vitae and Previous Knowledge of Interviewees – Its Interest for Conducting Biographical Interviews. The conduct of biographical interviews usually neglects the importance of precise knowledge of people before interviewing them. This paper suggests that personal names play a decisive role in the building of personal identities and trajectories, and examines the methodological consequences of such a statement. It shows to what extent the use of curriculum vitae and of any biographical information is an appropriate device to prepare and conduct the interviews. As a customized interviewing guide, biographical knowledge is presented as a means of eliciting an interviewee’s key life periods. Biographical Interviews, Curriculum Vitae, Knowledge of Interviewees.
L’ANALYSE DE SITUATIONS A L’EPREUVE DES SCENARIOS: L’EXEMPLE DES ACTIONS QUALITE DANS L’ORGANISATION
The Analysis of Situations by Scenarios – The Example of Quality Campaigns in Organizations. Scenarios are brief discourses submitted to a survey population. This method constitutes a useful manner for obtaining statements and discourse from the respondents. It helps reveal norms and rules of action about which individuals talk little and for which it is difficult to obtain information. The method can complement and enrich data obtained by more classic methods such as interviews. Through the comments and arguments of these respondents or actors, the method obtains access to the different logics of action revealed by the responses and the judgements provided in the framework of the proposed situations. The method can also reveal the capabilities mobilized by respondents in relation to the different problems encountered during work and also access the different forms of expertise used by each member of an organization in daily activity. Scenarios, Logics of Action in Surveys, Reference Framework, Tacit Rules, Debate between Actors, Confronting Judgements.
CHILDREN AS RESPONDENTS IN SURVEY RESEARCH: COGNITIVE DEVELOPMENT AND RESPONSE QUALITY Children as Respondents in Survey Research: Cognitive Development and Response Quality. 1
Natacha Borgers, Edith de Leeuw, Joop Hox
Although children are no longer a neglected minority in official statistics and surveys, methodological knowledge on how to survey children is still scarce. Researchers have to rely mainly on ad-hoc knowledge from such diverse fields as child psychiatry and educational testing, or extrapolate from methodological knowledge on how to survey adults. In this article, we review the available literature on children as respondents and present the first results of a secondary analysis of the influence of cognitive development on response quality. We end with recommendations for surveying children. Children, Cognitive Development, Data Quality, Questionnaires, Interviews.
N. 67, July 2000
QUEL PARTI VA GAGNER LES ELECTIONS? AVANTAGES ET FAIBLESSES D’UNE QUESTION NUMERIQUE 
Which Party Will Win? Advantages and Weaknesses of a Numerical Question. This article evaluates the numerical question used in the 1997 Canadian Election Study which measures electors’ perceptions of parties’ chances of winning the election. At first, this question appears inappropriate for reliable research. At least three important weaknesses are associated with the question. First, the formulation contains some ambiguities. Second, the literature provides many pieces of evidences regarding the limited capacities of people to deal with probabilities. Finally, responses to the 1997 Canadian Election Study are not consistent with researchers’ expectations regarding the form of these answers. However, the question provides reliable answers. Two empirical tests demonstrate that respondents give sensible answers. First, their perceptions follow the evolution of polls, and these perceptions also affect their voting behaviour. Methodology, Surveys, Elections, Strategic Voting.
A PROPOS DE L’ECHANTILLON. DE L’UTILITE DE QUELQUES MISES AU POINT
Concerning Samples – The Utility of Certain Clarifications. The article concerns qualitative samples and qualitative sampling. The subject is first treated from an historical point of view by looking at the work of the Chicago school of thought and the birth of the qualitative-quantitative opposition. This quarrel over methods is attenuated by distinctions made in this article between: (a) locality and subject of a study; (b) statistical representativeness and sociological (or theoretical) representativeness; (c) descriptive theory and explicative theory. The construction of a qualitative sample can be formulated as a calculation involving explicitly and uniquely defined and regulated operations. Sample, Qualitative Sampling, Chicago School, Quebec.
AN INTERVIEWER NETWORK: CONSTRUCTING A PROCEDURE TO EVALUATE INTERVIEWERS
Social research in general, and data collection in particular, can be viewed as a production process where “Total Quality Management” can be applied. Using the concept of TQM, we evaluate the quality of one important phase of the survey process: data collection by face-to-face interviews. In this article, we present the main results of a study of a procedure to evaluate interviewers working in a permanent network of interviewers. The work is based on a doctoral thesis (Carton, 1999). Given the philosophy of TQM, we end with some recommendations for improving the quality of the network and take account of the specific social context of the network. Total Quality Management, Interviewer Selection, Interviewer Training, Interviewer Supervision, Data Quality.
OBSERVATION ET MODELE LINEAIRE OU LOGISTIQUE: REPONSE A ARIS ET HAGENAARS
Observation and Linear or Logistic Models – Reply to Aris and Hagenaars. The answer proposed here to Aris and Hagenaars (2000) stresses the fact that inadequate application of logistic modeling in analyzing data only concerns marginal effects in percentages. We call for the abandon of the indicator in favor of either logistic regression odds ratio estimations, or observations of tabular analysis (or their estimation by marginal effects in percentage of a linear regression). Qualitative Data, Tabular Analysis, Logistic Regression, Linear Regression.
N. 68, October 2000
RC33 President’s Report
Edith D. de Leeuw
Report on the ISA Research Committee Council Business Meeting held July 30, 2000 in Montreal, Canada
RC33 Information Coordinator’s Report
Karl M. van Meter
RC33 2000 Conference Organizer’s Report
Survey Research Methods in the 21st Century: New Threats, New Concepts and New Opportunities
Don A. Dillman
Early Roots of Statistical Modeling
Willem J. Heiser
Report on the first Unfolding Session at the RC33 2000 Conference in Cologne
Wijbrandt H. van Schuur
Report on the Sessions and Round Table about Trends in Social Research Methodology
Hans van der Zouwen
Session “Translation and Validation of Measurement Instruments: Can Items Have Stable Meanings?”
Ethical, Social & Political Processes in Social Research
Variants of Automatic Coding of Text Data
Cluster Analysis of Network Data
Nonparametric Item Response Theory
Gideon J. Mellenbergh
Session, “Design and Analysis of Data in Social Surveys – Comparative Perspectives across Time and Space”
Methodology for Handling Complex Societal Problems
Dorien J. DeTombe
Geometric Data Analysis
Methodological Issues in Feminist Research
Capitolina Diaz Martinez
Sensitive Topics – Sensitive Questions: Overview of the Sociological Research Literature
Karl M. van Meter
Survey research by questionnaire in any form — fact-to-face, pencil-and-paper, by telephone, by computer — is confronted by problems associated with asking questions about sensitive topics. To provide an overview and a possible structure for “sensitive topics” and “sensitive questions” research, we analyze the contents of the Social Research Methods (SRM) and the Sociological Abstracts data bases for these terms. Sensitive Topics, Sensitive Questions, Social Research Methods Data Base, Sociological Abstracts, Research Literature.
Outline of sessions “Cognitive Methodological Research for Testing of Questionnaires” — I. Principles and II. Applications
Ger Snijkers, Tony Hak
Analysis of Structured Data
Report of the organisers of the sessions on Agent-Based Social Simulation
Klaus G. Troitzsch
Report on the session “Estimation of Differential Equations and Stochastic Processes: Methods and Applications”
Report of the session “From Opposition of Quantitative-Qualitative Approaches to a New Modality”
Bridging Qualitative & Quantitative Methods
N. 69; January 2001
IS IT FITTING TO USE THE TELEPHONE DIRECTORY AS A FRAME POPULATION IN SURVEYS? AN ANALYSIS OF THE CASE IN SPAIN.
Vidal Diaz de Rada
When the target population is specialized collectives (doctors, lawyers, agency directors, etc.), it is common to use the professional directories of each individual collective. However, what can be done if the general population is being targeted? The aim of this paper is to try to answer this question with an analysis of the extent to which the use of the telephone directory as a frame population for mail surveys is valid as it is for telephone or face-to-face surveys. With this aim in mind, the results of a mail survey carried out in Navarra, a region in the North of Spain in the Spring of 1998 will be outlined. The hypothesis to be tested is that the main problem in the use of the telephone directory as a frame population is the difficulty of achieving an adequate representation of the real target population. Mail Survey, Telephone Survey, Frame Population, Non-Coverage.
LES METHODES DE SELECTION DES PREMIERS SONDAGES AMERICAINES
Methods of Selection of the First American Surveys. This article describes the methods of selection used in the first American partial surveys done between the middle of the XIXth century and the 1920s. Three procedures, invented by the Department of Agriculture, are examined successively. We once again come across the specific role played by the selection of individuals in the generalization to the total population permitted by surveys considered to be “representative”. History, United States, Agriculture, Surveys, Selection, Generalization.
INTRODUCTION AUX METHODES STATISTIQUES EN SOCIOLOGIE DE BLOSS & GROSSETTI ET LEUR ENSEIGNEMENT EN FRANCE
Introduction aux Méthodes Statistiques en Sociologie by Bloss & Grossetti and Teaching Statistics in France. In this “Ongoing Research” section, we first present this book intended for undergraduate students in France. This is followed by a commentary by Philippe Cibois concerning the book and teaching statistics in sociology in France and especially the meaning of Chi-square. Cibois’ commentary is followed by a reply by the book’s authors. Teaching Statistics, Chi-square, Sociology in France.
FAIRE COMPRENDRE LE KHIDEUX. A PROPOS DE L’INTRODUCTION AUX METHODES STATISTIQUES EN SOCIOLOGIE DE THIERRY BLOSS ET MICHEL GROSSETTI
Understanding Chi-square – Concerning Introduction aux Méthodes Statistiques en Sociologie by Thierry Blöss and Michel Grossetti. Related to this introductory statistics book, several questions on how to teach the Chi-square are examined. The author proposes to not using the Chi-square test as a simple recipe but to help students understand that the Chi-square is homogeneous over a given sample and that basic attention should be made in its application. Chi-square, Null Hypothesis, First Order Risk, Yates’ Correction.
COMMENT PARLER DES TESTS STATISTIQUES AUX ETUDIANTS FRANCAIS DE SOCIOLOGIE? REPONSE A PHILIPPE CIBOIS
Thierry Bloss, Michel Grossetti
How to Teach Statistical Tests to French Sociology Students? – Reply to Philippe Cibois. The authors of the book, Introduction aux Methodes Statistiques en Sociologie, reply to the preceding critique of their work by Philippe Cibois. Teaching Statistics, French Sociology, Chi-square.
N. 70, April 2001
MODAL AUXILIARIES IN TEXT ANALYSIS
Modal auxiliary verbs (e.g., ought, want, etc.) convey information about the intentions of the semantic subjects within the clauses in the text in which they appear. For example, in asserting that a person ought to act in a certain way, one contrasts the action with the person’s potential intention to act otherwise. This quality can be used in research. Semantic text analysis allows coding modal auxiliaries. This paper investigates how to make modal auxiliaries remain visible when network text analysis is used. This type of text analysis allows one to deal with rather complex argumentation. The answer shows that in this situation network text analysis should not be used, the investigator should stay with semantic text analysis. Semantic Text Analysis, Network Text Analysis, Modal Auxiliary Verbs.
LE DIRE ET LE DIT DANS LES ENTRETIENS: ELEMENTS POUR LE TRAITEMENT DE LA COMPLEXITE DU LANGAGE
Saying and What’s Said in Interviews: Elements for the Treatment of Complexity in Language. This article attempts to pose the principles for the use of certain tools, forged in linguistics, for sociological analysis, particularly for the analysis of interviews. Far from being a simple collection of information, the interviews with migrants analyzed here are an arena for intense language activity. With the use of concepts such as annunciation, pragmatism and natural logic, various language functions are identified that sociologists should take into consideration: the construction of representations, negotiation of one’s position in an interaction. Contrary to the spontaneous sociology of certain linguistics, language practices are not those of an actor free of all social determinants. On the contrary, by taking into consideration the complexity of language, sociology can construct a plural, heterogeneous and even divided actor. Language Practices, Language Complexity, Interviews, Social Actors.
OVERVIEW OF TEXT ANALYSIS SOFTWARE
In the last few years, quite a few computer programs for the analysis of texts have appeared on the market. This overview provides information on this software, and it also attempts to classify them. The short descriptions contain general and technical information, as well as whether or not a test, demo, or trial version is available. The information given here is constantly updated and can be found at http:www.intext.de/TEXTANAE.HTM. Text Analysis, Content Analysis, Qualitative Data Analysis, Computer Programs.
COMPTE-RENDU DE LA JOURNEE D’ETUDES DU CIDSP: LES NOUVELLES METHODES D’ANALYSE DES ENTRETIENS – ANALYSE ASSISTEE PAR ORDINATEUR ET CAPITALISATION DES ENTRETIENS NON-DIRECTIFS DE RECHERCHE
Mathieu Brugidou, Bruno Cautrès
Report on the CIDSP Conference: New Methods for Analyzing Interviews – Computer-Assisted Analysis and Secondary Analysis of Non-Directive Research Interviews. This conference was organized by the authors on 9 March 2001, at the Maison des Sciences de l’Homme in Grenoble, France. Computer-Assisted Qualitative Data Analysis Systems (CAQDAS), Interviews, Open-Ended Questions, Interview Data Archiving.
N. 71, July 2001
PROBLEMS IN COMPARATIVE AND TRIANGULATED HOMELESSNESS RESEARCH1
In 1993, the French National Institute of Demographics (INED) in Paris started a research program on homelessness, combining in-depth interviews of homeless people and service providers, and several statistical surveys of the Paris homeless, the two main ones being a survey of 591 homeless adults (1995) and a survey of 461 homeless youth aged 16 to 24 (1998). The methodology of those surveys was close to the one used by several American studies, namely the Urban Institute 1987 national survey, the RTI 1991 survey of the Washington homeless and the Bureau of the Census 1996 national survey of homelessness. Hence some comparisons are possible. This report focuses on two issues. First, drawing on the examples of the homeless in Paris and the United States, I will discuss the problems of international comparisons which arise from differences in methodology, definitions, field of the surveys, etc., making the elaboration of comparable results difficult; and from differences in policies designed to relieve poverty, to help homeless people, to deal with immigration or mental illness, etc., also making interpretation difficult. Second, I will examine the use of triangulation from statistical surveys and in-depth interviews, with examples taken from our research program. This last question will be mostly dealt with from the standpoint of someone conducting a statistical survey, but the contribution of quantitative data to qualitative investigations will also be investigated. Homeless, Statistical Survey, In-Depth Interviews.
QUALITY OF LIFE INDEXES FOR NATIONAL POLICY: REVIEW AND AGENDA FOR RESEARCH1
Michael R. Hagerty, Robert Cummins, Abbott L. Ferriss, Kenneth Land, Alex C. Michalos, Mark Peterson, Andrew Sharpe, Joseph Sirgy, Joachim Vogel
A number of governments and public policy institutes have developed “Quality of Life Indexes” (QOL) statistics that attempt to measure the quality of life for entire states or regions. We develop 14 criteria for determining the validity and usefulness of such QOL indexes to public policy. We then review 22 of the most-used QOL indexes from around the world. We conclude that many of the indexes are successful in that they are reliable, have established time series measures, and can be disaggregated to study subpopulations. However, many fall short in four areas: (1) indexes vary greatly in their coverage and definitions of domains of QOL, (2) none of the indexes distinguish among the concepts of input, throughput, and output that are used by public policy analysts, (3) they fail to show how QOL outputs are sensitive to public policy inputs, and (4) none have examined convergent validity against each other. We conclude that many of these indexes are potentially very useful for public policy, and recommend research to further improve them. Quality of Life (QOL), QOL Index Comparison, Comparative Research.
SCIENTOMETRICS AND SCIENCE STUDIES
Authors cannot be reduced to texts and texts cannot be reduced to authors. Cognitions emerge and can be recognized in the development of complex networks of texts and authors operating on one another. The study of these multivariate networks over time provides the subject of “scientometrics”. How can one describe and/or measure the sciences so that the cognitive content can be made visible? First, one needs non-parametric statistics because idealization would abstract from the cognitive substance. The substance is shaped and reproduced at each moment in terms of socio-cognitive interactions (“the static model”). Second, the interactions can be evaluated along the time axis in terms of what the interactions mean with reference to agency, on one hand, and with reference to cognitive development at the network level, on the other (“the dynamics of science”). The discursive mediation has been refined by scientific codification; for example, in terms of citation practices, indexing, and abstracting services. The codification along different axes drives the developments reflexively toward increasing “self-organization” of the knowledge-based systems. Autopoiesis, Self-Organization, Entropy Statistics, Scientometrics, Measurement, Recognition, Science.
THOUGHTS ON THE CLARIFICATION OF SOCIOLOGICAL CONCEPTS
Fred W. Riggs
Concepts are units of thought or knowledge, not words used to represent them. The words found in our lexicon often if not always can represent more than one concept. We usually understand an author’s intentions because, in context, only one of the possible meanings of a word makes sense. When we are not sure which of a word’s meanings is intended, we may turn to a dictionary or a glossary where its different senses are defined and select the one that seems most relevant. A growing number of specialized glossaries seek to keep pace with the proliferation of technical terminology and, fortunately, many of them can be consulted on-line. However, these and other lexicographic tools all suffer from their inability to place concepts in the context of the theories and systems of concepts that make them significant. They need to be supplemented by both semantic and onomantic analysis which the author presents. The Internet now permits us to speed and improve this ancient process by systematizing conceptual innovation and the author gives several examples including a Terminology site for Ethnicity, a well-developed glossary for Linguistics. the Online Computing Dictionary. Sociological Concepts, Dictionaries, Glossaries, Semantics, Onomantics, Internet.
N. 72, October 2001
THE DOMAINS OF THE QUALITY OF LIFE
Abbott L. Ferriss
Cultural values define “the good life” which is measured by both subjective and objective domains. Domains are variously identified through: (a) “elitist” scholarly identification of universal values; (b) sample surveys of persons in different cultures describing their QOL (Quality of Life); (c) psychological dimensions of well-being; (d) dimensions of social well-being; and, (e) satisfaction with sectors of daily life as represented by social institutions. These various nomenclatures are combined to compose a set of ten domains that identify major sectors of “the good life”. Hypotheses are suggested for improving several specific QOL domains. Values, Psychological Well-Being, Social Well-Being, Satisfaction, Life Concerns, Institutions Sector Measurement.
SURVEY DATA FUSION
Toma Aluja-Banet, Santiago Thiô
This paper concerns the problem of merging information coming from independent sources. Surveys are one of the most used ways to collect quantitative information, very often at great expense. Also, the reliability of the collected information in modern societies is decreasing due to the “survey pressure” on respondents, thus making it interesting to avoid direct requests to respondents. This is a challenging endeavour at the limits of what statistics can do. Any fusion can be reduced to its canonical situation, which is defined by a complete file, which we call donor, with variables x’s and y’s, and another file, which we call recipient, with only variables x. The fusion consists of transfering the y variables from the donor file to the recipient. It is worthwhile to point out that the fusion does not create new information, it just profits better from that existing. The objective is to estimate individual data valuable in an aggregated way. There are two main methodologies for data fusion: One is modelling the relation of the y variables with respect to the x’s in the donor file and applying this to the recipient file. The other is the hot deck methodology, which involves finding for each individual of the recipient file one or more similar individuals in the donor file and transfering the y values of these individuals to the recipient. Although the problem is much more general, here we consider only the problem of combining survey data coming from TV-audience and consumption surveys. The hot-deck by k-nearest neighbours fusion is currently being employed in media surveys, but in general it greatly reduces the variability of imputed variables. Our objective is compare the accuracy of k-nn fusion with PCAR imputation and PLS regression to assess the advantages of these methodologies. Fusion, Imputation, Hot Deck, Modeling, K-Nearest Neighbours, PCAR, PLS Regression.
WWW — SEARCH ENGINES, AN ALTERNATIVE TO THE SOCIAL SCIENCE DATABASES?
In the age of the World Wide Web, users of information services are faced not only with a highly decentralized, heterogeneous data space with different data sources and with “value-added” components in information systems, but also with the more fundamental question of whether they still should, for instance, search (gray) literature in the specialized databases of information centers like the IZ (InformationZentrum Sozialwissenschaften) in Bonn, Germany, or whether they should use general WWW search engines like AltaVista or Fast. Does the WWW displace information centers with their high quality databases or will it only fill the information gaps left? What are the results we can obtain nowadays using general WWW search engines? How do the search engines work and what are the differences with respect to the information retrieval of specialized databases? World Wide Web (WWW), Search Engines, Heterogeneity, Information Retrieval.
N. 73, January 2002
LANDMARK PAPERS IN SURVEY STATISTICS
This article is a version of the author’s introduction to the International Association of Survey Statisticians (IASS) Jubilee commemorative volume, Landmark Papers in Survey Statisticis, which was published late last year. It is reprinted here with the authorization of the IASS which retains full copyright authority. It also includes the references to the cited papers. IASS, Survey Statistics, History of Statistics.
SYMPOSIUM ON CELEBRATING CLASSIC SOCIOLOGY: PIONEERS OF BRITISH QUALITATIVE RESEARCH
Report from the Symposium on Celebrating Classic Sociology: Pioneers of British Qualitative Research, which was held on 5-6 July 2001 at the University of Essex. The full program is also included along with the available abstracts. Qualitative Research, British Sociology, History of Sociology.
Multi-Level Analysis. This article is based on a one-day course that the author has presented in Lille and in Paris to French-speaking sociologists who are just now becoming familiar with the method. Multi-Level Analysis, Statistical Modeling, Statistical Education.
LA BOITE A MOUSTACHES POUR SENSIBILISER A LA STATISTIQUE
Monique Le Guen
Box and Whiskers Plot for Developing Statistical Sensitivity. The Box & Whiskers Plot is an invention of Tukey (1977) for schematic representation of a distribution. This graphic representation can serve as a means of familiarization with abstract statistical concepts. In this article, we examine the need to rethink initiation to statistics and base our work on new knowledge in neuron sciences which propose placing the student in a situation of discovery, while using real data and exploratory data analysis computer programs. In the second part of this article, we explain in detail how to interpret box and whiskers plots, and show how students can discover, while exploring data, certain properties of the median and the means. In the references, we provide Internet addresses for further work with box and whiskers plots. This article is intended for faculty members and those who work with applied statistics. Statistics Education, Interactivity, Visualization, Exploratory Data Analysis, J. W. Tukey, Box and Whiskers Plot.
N. 74, April 2002
L’ANALYSE DE DONNEES TABULAIRES AVEC LES MODELES LINEAIRES ET LOGLINEAIRES
Emmanuel Aris, Henri Aris
Analysis of Cross-Tabulation Data with Linear and Loglinear Models. Several models have been proposed to evaluate the strength of the relationships between observed variables, some dependent on others. In particular, when the variables are categorical, models in which the effects are considered additive (linear models), or multiplicative (loglinear/logit models) are often used. In the most complete version of these models, dependence relationships between several categorical dependent and independent variables can be modelled. The aim of this article is to provide basic knowledge of these models to be able to apply them satisfactorily. The fundamental dependency equations, the construction of the complete models, and recent developments are presented. Further, references to programs that may be used to apply these models, and to more technical works, are provided. This presentation is based on the analysis of an example. Categorical Dependent Variables, Linear Models, Logit Models, Loglinear Models
MESURER LA SENSIBILITE AUX CONDITIONS INITIALES DANS LES PROCESSUS INDIVIDUELS DE PREMIERE INSERTION PROFESSIONNELLE: PROPOSITION DE PROCEDURE ET APPLICATION*
Measuring the Sensitivity of Initial Conditions in Individual Procedures for Finding First Professional Employment – Proposal for a Process and Application. Individual procedures for obtaining first professional employment are influenced simultaneously by social determinism, chance and individual strategies. The question is how to unravel these different factors. Starting with the central opposition between chance and determinism — where strategies create possibilities — this study proposes an analysis within the framework of bounded chaos. First, a method is proposed for the measurement of sensitive initial conditions (SCI) based on Lyapunov powers associated with obtaining first professional employment. Second, this study shows that the SCI measure permits one to examine the respective roles of chance, determinism and strategy. It then suggests that the SCI measure is a better tool for counseling on scholastic and professional orientation than surveys of entire classes concerning first jobs. The latter remain better adapted to counseling and evaluating educational institutions and curriculums. First Jobs, Social Determinism, Individual Strategies, Bound Chaos.
N. 75, July 2002
AN INNOVATIVE TECHNIQUE FOR ASKING SENSITIVE QUESTIONS: THE THREE-CARD METHOD1
Judith A. Droitcour, Eric M. Larson
The three-card method is a survey technique that is being developed by the U.S. General Accounting Office (GAO) to collect sensitive data in large-scale personal-interview surveys. The technique is intended to allow estimation of the needed statistics while maximizing response privacy and reducing “question threat.” The technique was originally devised to collect data on respondents’ immigration status; however, it might also prove useful in a variety of other sensitive question areas. Sensitive Topics, Sensitive Questions, Social Research Methods Data Base, Sociological Abstracts, Research Literature
HONEST BY CHANCE: A QUALITATIVE INTERVIEW STUDY TO CLARIFY RESPONDENTS’ (NON-)COMPLIANCE WITH COMPUTER-ASSISTED RANDOMIZED RESPONSE
Hennie Boeije, Gerty Lensvelt-Mulders
The randomized response method (RRM) is intended to minimize evasive responses to sensitive social questions. This study examines how respondents participate in the RRM and how their participation contributes to (in)adequate answers. Eleven respondents, all social service clients, answered twelve questions in a computer-assisted survey. The forced response technique was used as the RRM, and computerised dice (internal randomiser) were alternated with manual dice (external randomiser). Respondents’ behaviour was video taped. Afterwards, a qualitative interview was conducted which was audio-taped. The results show that the RRM caused very few problems. All respondents had trouble giving false answers forced by dice and, as a consequence, some cheated. Moreover, their individual interest, capping of their social service allowance, determined their compliance since it overruled all other arguments even when the respondents understood the RRM. Randomized Response, Social Security Benefits, Sensitive Questions, Surveys.
EXTERNAL VARIABLES AS POINTS IN SMALLEST SPACE ANALYSIS: A THEORETICAL, MATHEMATICAL AND COMPUTER-BASED CONTRIBUTION
Erik H. Cohen, Reuven Amar
Given a configuration of n points in a Euclidean space and given a vector of similarity coefficients of an external object E with those n points, the problem is how to locate E among the n fixed points such that the higher the similarity of E with a point, the lower its distance from this point. Although this problem has been technically resolved by means of several types of procedures discussed below, it is nonetheless generally neglected. In the present article, we offer a theoretical and mathematical contribution to resolving this problem. Smallest Space Analysis, External Variables, Similarity, Distances.
SEMINAIRE LASMAS-IDL DES 7 ET 8 JANVIER 2002 – LES EXPERIENCES NATIONALES D’ETUDES DE COHORTES ET LEUR APPORT AUX SCIENCES SOCIALES 1
LASMAS-IDL 7-8 January 2002 Seminar – National Experiences with Cohort Studies and Their Contribution to the Social Sciences : The objective of this seminar was the presentation of foreign surveys in order to evaluate the pertinence of extending educational panel studies for those leaving the educatioinal system. The first day included presentations on research involving cohorts, panels or retrospective surveys done in the United States, Great Britain and Germany. The second day was intended for sharing experiences with foreign colleagues concerning the necessity, the conception and the operationalization of cohorts studies and the various difficulties encountered in maintaing contact with the individuals involved over long periods of time. Cohorts, Educational Panels Studies, USA, Great Britain, Germany, France.
N. 76, October 2002
RESO COMME OUTIL DE STRUCTURATION DES CATEGORIES SOCIOPROFESSIONNELLES
Reso as a Tool for Structuring Social-Professional Categories: For purposes of readability ande comparison, how can one reduce the social mobility tables that are as detailed as those of the 1985 French FQP survey, and, at the same time, not sacrifice the nomenclature but show the continuity/discontinuity between categoris? Contrary to work in “stratification” which perceives the nomenclature as a hierarchy and qualifies mobility as “ascending”, we model each table in the form of a graph and reveal the social-professional categories as typologies with the software Réso which uses strong connectedness to characterize the impossibility of constructing an induced order of mobility on the 32 categories. Réso constructs proximities between categories by finding and following lines of flow and fracture in a center(s)/periphery(ies) model with points of flexion and points of weakness. These proximities vary from one table to another and cannot be expressed in terms of “stratification”. Social Mobility, Social Professional Categories, Stratification, Graphes, Strongely Connected Components.
THE EFFECT OF INTERVIEWER ATTITUDE ON SURVEY COOPERATION
Michael Lemay, Claire Durand
Research to improve the quality of social surveys has shown that methodological factors have a significant impact on respondent cooperation. The research presented here is inspired by the sociology of work and studies the impact of interviewers’ sense of motivation and “self-efficacy” in convincing respondents to answer surveys. To do so, a survey of interviewers working for three private pollsters was conducted during the Canadian federal election campaign of November 2000. The results indicate some evidence of relationships between motivation, sense of self-efficacy (and associated behaviour), and interviewer performance. However, some of the results could be better explained by differences in training. In addition, the lack of a better measure of performance may also explain the weakness of the results. Future studies should examine performance measurement and try take into account how performance changes over time. Surveys, Interviewers, Interviewer Performance, Interviewer Attitudes, Measurement of Interviewer Performance.
COMPARAISON DE L’ANALYSE STRUCTURALE ET DE L’ANALYSE SCIENTOMETRIQUE DANS L’ETUDE DE LA PRODUCTION SCIENTIFIQUE
Comparison of Structural Analysis and Scientometric Analysis in the Study of Scientific Production: This article is based on the study of networks of advice, coauthorship and cocitation among a certain category of French cancer researchers. We show that the study of scientific production based on strictly quantitative indicators, characteristic of scientometry, is not sufficient. It is also necessary to take into account social networks constructed among such researchers to have a more complete view of a scientific domain and how it is organized. In a comparative process, we present the structural approach which answers this requirement through the use of hypotheses concerning intra-organization of interdependent entrepreneurs that are tested at the inter-organizational level. The two approaches – scientometrics and structural analysis – are then shown to be strongly complementary in their results which lead to a better understanding of how a scientific field of research functions and the roles of the different actors. Scientific Production, Structural Analysis, Scientometric Analysis, Elites, Inter-Organizational Analysis.
N. 77, January 2003
TENSION, ADVENTURE, AND RISK (TAR) AND THE CLASSIFICATION OF OCCUPATIONS: A MULTIDIMENSIONAL ANALYSIS
Erik H. Cohen, Elchanan I. Meir, Hila Segal, Reuven Amar
The vast array of occupations can be classified in many ways. The present study employed Roe’s (1956) classification, based on psychological differences, which posits eight occupational fields: Business, Organization, General-Cultural, Service, Arts and Entertainment, Outdoor, Science, and Technology. These fields have been found, in more than 25 studies (see Meir, 1975), to display a circular structure. However, an additional group of occupations, entitled Tension, Adventure, and Risk (TAR), does not figure as a separate field in existing occupational classifications, including Roe’s. The objective of this investigation was to determine whether TAR constitutes a separate field (Hypothesis 1) or a separate dimension, a differentiation within other fields (Hypothesis 2). Following procedures designed by L. Guttman (1968) and others, namely the SSA and the External Variables module, the findings support the second hypothesis. Theoretical and practical implications of the results are discussed. Smallest Space Analysis, External Variable, Multidimensional Analysis, Guttman, Occupational Classification, Tension, Adventure, Risk.
LE DEMI REVENU MEDIAN, MARTINGALE DU SEUIL DE PAUVRETE?
The Half Median Salary, A Toy for Defining Poverty?: In France, the commonly accepted definition of (economic) poverty level is a relative measure, half the median salary. Computer simulations permit us to question the pertinence of this criteria and show, on one hand, that the poverty level and the rate of exit from poverty cannot be evaluated by themselves but in relation to the salary structure of the entire population and also in relation to the structure of salary movements, even though this is not often rendered explicit in work on poverty. On the other hand, there are other statistical indicators which could better describe the phenomenon of poverty. Poverty Level, Half Median Salary, Exit from Poverty, Simulations, Salary Structure, Structure of Salary Movements.
QUESTIONNAIRE DEVELOPMENT, EVALUATION, AND TESTING METHODS (QDET) Conference – Good Discussions, Debates, and Dining in Charleston
Jennifer M. Rothgeb
The International Conference on Questionnaire Development, Evaluation, and Testing Methods (QDET) took place in Charleston, South Carolina, on November 14-17, 2002. There were 32 sessions with 22 invited papers and 54 contributed papers. In addition, there were 15 poster presentations. The author describes the meeting, presents the entire program (sessions / posters, authors, titles) and provides the abstracts for the invited papers. Surveys, Questionnaires, Testing & Development Methods.
TABLEAUX CROISES ET DIAGRAMMES EN MOSAIQUE, POUR VISUALISER LES PROBABILITES MARGINALES ET CONDITIONNELLES
Monique Le Guen1
Cross-Tabulations and Mosaic Diagrams For Visualizing Marginal and Conditional Probabilities: This article is part of a campaign to encourage awareness to various aspect of statistics. Visualizing information with graphic methods, when based on New Information and Communication Technologies (NICT), is a promising means to improving the comprehension of abstract concepts in statistics. We address the question of the visual representation of cross-tabulations by mosaic diagram. In the framework of statistical graphics, we present an example of double entry cross-tabulations which permits us to introduce the vocabulary and the different statistical elements — sample size, marginal probabilities, conditional probabilities — found in cross-tabulations. We show the information provided by visual representations as bar diagrams, spine plots and mosaic diagrams. We close with a discussion of ongoing developments and software. The references provide links to articles and software accessible by Internet. Visualization, NICT, Cross-Tabulations, Marginal Probabilities, Conditional Probabilities, Bar Diagrams, Spine Plots, Mosaic Diagrams.
N. 78, April 2003
DE LA SOCIOLOGIE DE LA SCIENCE AUX REVUES DE SOCIOLOGIE
From the Sociology of Science to Sociological Journals: In France, in the social sciences, the evaluation of published research is a poorly structured field. But evaluation is possible when following a rigorous procedure. The method developed here by the author consists to crossing several authoritative data bases of periodicals, then constructing lists of reviews which a widely submitted to the scientific community for comment, and also to a few experts. A journal is then considered scientific when it is so judged to be by the researchers of its scientific community. A case study is proposed here concerning the field of sociology. The survey results (almost 200 French and foreigh journals and half a thousand sociologists questioned) are analyzed here. Sociology, Research, Evaluation, Reviews, Bibliometrics.
INTERNATIONAL CONFERENCE ON METHODOLOGY AND STATISTICS, LJUBLJANA, 15-18 SEPTEMBER 2002
Valentina Hlebec, Anuska Ferligoj
This was the twentieth annual conference in Slovenia on methodology and statistics in the social sciences and included invited presentations by Patrick Doreian, John C. Gover, Robin Henderson and Donald B. Rubin. There were two plenary lectures, four invited presentations and a total of 78 papers presented. Social Sciences, Methodology, Statistics, Slovenia.
SECOND RC33 COLOGNE 2000 SOCIAL SCIENCE METHODOLOGY CONFERENCE REPORT
Karl M. van Meter
With the second edition of the conference CD, we analyze the contents of the CD and the evolution of the production of the conference which increases from 167 papers (19,8 Mo) to 241 papers (34,2 Mo), and compare it with what we presented in the BMS n. 68 (October 2000). Social Sciences, Methodology, Statistics, RC33, Cologne.
N. 79, July 2003
MARLOWE, PROSPERO & LA TECHNOLOGIE LITTERAIRE
Francis Chateauraynaud, Bernard Reber, Karl van Meter
Marlowe, Prospero and Literary Technology: An author, an experimenter and a reviewer of the computer programs, Marlowe and Prospero, contribute in this article, respectively and successively, an extensive and detailed presentation of Marlowe, a report on what it is like to interact with Marlowe and a review of the recently-published work, by the first author, describing the development of Prospero for the Analyse of complex dossiers of texts concerning a social controversy, and Prospero’s extension and adaptation with Marlowe to direct natural language dialog with researchers concerning specific complex dossiers. Artificial Intelligence, Text Analyse, Content Analyse, Social Controversies, Language-Based Data.
NONRESPONSE IN SURVEYS: DETERMINING THE RESEARCH AGENDA FOR THE FUTURE
Joop Hox, Edith de Leeuw, Ger Snijkers
In this paper, we describe research questions and their priorities as reported by experienced researchers in the field of nonresponse gathered at the international nonresponse workshop in Copenhagen. These experts give highest priority to research on statistical non-response adjustment and measures to reduce non-response, including monitoring the effects of these measures. Nonresponse, Adjustment, Reduction, Future Research, Concept Mapping.
PROXY NETWORKS – ANALYZING ONE NETWORK TO REVEAL ANOTHER
Data from Web-based book retailers, such as Amazon, often provide, for an given book, the list of other books that were also purchased by customers along with the first given book. This article uses this network tie information to construct social networks of “buddy books”. Although the actual political affiliation of each book purchaser is not known, the structure of the buddy book network shows that there are two clearly divided groups: a larger and more diffuse left-of-center readership, and a smaller and more closely tied right-of-center readership. Types or networks of readers linked to a specific author are also studied. Online Data, Social Network Analysis, Joint Readership, Book Purchasers.