“A new behavioural scale measuring social norms”
[Heiko Rauhut, (University of Zürich)]

Surveys usually measure normative behaviour and attitudes by asking whether some (un)desirable action has been performed in the past, and how this behaviour is evaluated. This may lead to biased estimates if the respondents have an incentive to over-report or under-report due to social desirability. Instead of relying on anonymising techniques to measure prosocial behaviour (e.g. RRT, item count) , we propose a complementing, truly behavioural measure of pro-social behaviour and beliefs. In a series of studies in different populations, we implemented the social value orientation slider measure (SVO) (Murphy et al. 2014) and our new norms scale NS4 (Rauhut et al. 2017) in an online survey. Both measures are based on small tasks in which the respondents distribute a sum of money between themselves and others, and are relatively easy to implement in a survey. Other than the SVO, the NS4 additionally asks for normative and empirical beliefs about what one will and should do. To validate our measures, we use a series of behavioural tasks, e.g. cheating about the outcome of a die or the willingness to answer additional questions, and correlate these control tasks with our new measures as well as with more traditional survey measures. Our results suggest that the new measures are a forceful predictor for constructs usually associated with normative behaviour, such as cheating, cooperation, or political attitudes, and may therefore be a valuable tool for future studies in the social sciences.

Presentation on 07.11.2017

“Studying people and society through large-scale online traces: the case of armed conflicts in online news”
[Robert West, (EPFL)]

Wars and conflicts have constituted major events throughout history. Despite their importance, the general public typically learns about such events only indirectly, through the lens of news media, which necessarily select and distort events before relaying them to readers. Quantifying these processes is important, as they are fundamental to how we see the world, but the task is difficult, as it requires working with large and representative datasets of unstructured news text in many languages. To address these issues, we propose a set of unsupervised methods for compiling and analyzing a multilingual corpus of millions of online news documents about armed conflicts. We then apply our methods to answer a number of research questions: First, how widely are armed conflicts covered by online news media in various languages, and how does this change as conflicts progress? Second, what role does the level of violence of a conflict play? And third, how well informed is a reader when following a limited number of online news sources? We find that coverage levels are different across conflicts, but similar across languages for a given conflict; that Middle Eastern conflicts receive more attention than African conflicts, even when controlling for the level of violence; and that for most languages and conflicts, following very few sources is enough to stay continuously informed. Finally, given the prominence of conflicts in the Middle East, we further analyze them in a detailed case study.
(Joint work with Jürgen Pfeffer from TU München.)

Presentation on 05.12.2017

“Using a mobile app to collect expenditure data: Participation, measurement and respondent experience”
[Annette Jaeckle, (Institute for Social and Economic Research, University of Essex)]

2018

Presentation on 27.02.2018

“Dealing with the complexity of cross-national data: The potential of web probing to assess and explain mis­sing equivalence of comparative national identity measures”
[Katharina Meitinger, (GESIS, Mannheim)]

There has been a tremendous increase in cross-national data production in social science research in recent decades. Before drawing substantive conclusions based on cross-national survey data, researchers need to verify whether the measures are indeed comparable. Quantitative measurement invariance tests are an important control tool for researchers working with cross-national data. However, this approach does not provide explanations if data is found noninvariant and it cannot assess the comparability of single-item measures.

The recently developed qualitative method of web probing can supplement a quantitative assessment and reveal why measures are incomparable. This presentation will introduce the method of web probing and present three applications of web probing for the evaluation of cross-national data: as an assessment tool for single-item measures; in combination with traditional measurement invariance tests using multi-group confirmatory factor analysis; and its contribution to the current discussion on approximate measurement invariance. This presentation uses substantive examples from the field of national identity and will discuss the comparability of the measures of general national pride, patriotism and nationalism as well as patriotic feelings from the 2013 International Social Survey Program (ISSP) module on “National Identity.”

Presentation on 27.03.2018

“Analyse des réseaux sociaux et oeuvres de fiction : lecture distante de systèmes de personnages”
[Yannick Rochat, (University of Lausanne)]

Nous présentons une revue des travaux portant sur les réseaux de personnages de la littérature, soit la modélisation sous forme de données relationnelles de systèmes de personnages. Nous illustrerons larevue à l’aide d’exemples issus de la littérature ainsi qu’avec des résultats originaux.Cette approche de lecture intermédiaire et distante permet à la recherche en humanités numériques de s’intéresser à la représenta tion des personnages avec des corpus de taille importante, nécessitant ainsi l’automatisation des méthodes d’exploration des données. Elle permet de tirer des conclusions sur les configurations exhibées par les entités des œuvres littéraires au travers de relations construites entre elles.

En conclusion de la présentation, nous proposerons un framework visant à classer les types d’analyses déjà réalisées et soulignant les questions de recherche à entrevoir pour l’avenir de cette approche.

Presentation on 24.04.2018

“An Introduction to Factorial Designs Using the Example of Hiring Decisions”
[Robin Samuel, (University of Luxembourg)]

In this contribution, we use a factorial design to explore the moderating role of transaction costs on scarring due to previous unemployment and skills underutilization. Furthermore, we investigate the extent to which the perceived difficulty of recruiting moderates these effects. Factorial designs allow studying respondents’ evaluations as a function of multidimensional stimuli. In this application, we create a pool of hypothetical candidates, where we experimentally vary individual characteristics of young job applicants. We then measure how our respondents, actual recruiters, evaluate the hiring chances of these young people. We further use information provided by the respondents to estimate transaction costs.

Using data from a recent large-scale factorial survey of recruiters in four European countries (N ~ 2, 000) and employing multilevel linear regression
models, we found, overall, scarring due to skills underutilization to exceed scarring due to unemployment. Skills underutilization was especially penalized when recruiting for a particular position was considered easy. Indirect transaction costs, particularly anticipated time required for organizational socialization, were negatively associated with unemployment scarring, but positively with scarring due to skill underutilization. Unemployment spells only had a negative effect on hiring chances for jobs where there were monetary expenses for introductory trainings.

Presentation slides

Presentation on 1.10.2013
Analyser de grands corpus de controverses sur le web: le cas de l’identité nationale en France

[Pascal Marchand (University of Toulouse)]

Le « grand débat sur l’identité nationale » restera sans doute comme l’une des polémiques majeures du quinquennat 2007-2012. Mais s’il a fait couler beaucoup d’encre, ce débat n’avait jamais été vraiment analysé. Le ministre et ses sondeurs se sont attachés à répondre aux détracteurs du débat et à en défendre la légitimité, mais comment les Français ont-ils finalement traité la question « qu’est-ce qu’être Français aujourd’hui ? ». Comment rendre compte des 18 240 contributions publiées sur le site web ? Le ministre avait proposé de les résumer ainsi : « être Français, c’est avoir des droits et des devoirs ». Quant aux sondeurs, ils publiaient un compte-rendu d’une remarquable pauvreté méthodologique et technique. L’informatique et la statistique textuelles permettent d’analyser de tels discours et spécifiquement de gros corpus. Plus précisément, l’analyse par Iramuteq, un logiciel universitaire gratuit, libre et opensource, développé à Toulouse par Pierre Ratinaud, permet de rendre compte du contenu de toutes les contributions, sans piocher au hasard dans la masse, ni faire intervenir nos propres préjugés. Il s’agit de reconnaître et de trier automatiquement le vocabulaire utilisé par les internautes pour obtenir des classes de discours. Et les conclusions sont très différentes de celles du ministre et de ses sondeurs : on découvre des psychologies, des histoires, des expériences, des éruditions et des prises de position d’une rare intensité. Des discours nationalistes et xénophobes ? Il y en a, et sans précautions, ni euphémismes. Certains y applaudissent l’initiative du ministre et de son gouvernement, d’autres dénoncent ce débat et ses sous-entendus politiciens. Certains décrivent une France historique et ses références obligées. D’autres livrent des témoignages poignants d’adhésion aux valeurs d’ouverture et de fraternité. Y compris de la part de personnes qui viennent d’ailleurs et expriment leur attachement à ce pays. Certains ne comprennent pas cette question à l’heure de l’Europe et de la mondialisation, et d’autres marquent leur attachement à leur région. Non, décidément, ce débat ne peut pas être ramené à un slogan simpliste et la statistique textuelle nous invite à cette découvrir cette complexité et à visiter ses mondes lexicaux.

Presentation_Marchand2013

Presentation on 5.11.2013
The Day Reconstruction Method: Linking time-use with emotional well-being

[Michael Ingenhaag (IEMS, UniL)]

The main goal of this study is to describe the association between disability and experienced well-being among older persons from different low and middle income countries. Specifically, we examine differences in the assessment of emotional affects and the allocation of time associated with disability applying the Day Reconstruction Method (Kahneman et al., 2004). Disabled people report lower emotional affects during all activities. Disability is further associated with more time spent in leisure and self-care activities and less time in work-related activities. We then combine these results regarding emotional affects and time allocation associated with disability in order to analyze the relation between time allocation, emotional affects and experienced well-being. Average experienced well-being is lower for disabled individuals. However, the results provide evidence for partially compensatory effects of differences in time allocation and emotional affects associated with disability. We show that the substitution of less pleasant work-related activities by more pleasant leisure activities associated with disability partially compensates the negative effect of lower emotional affects on experienced well-being.

Presentation Ingenhaag 2013

Presentation on 3.12.2013
Principes et analyses du signal électroencéphalographique (EEG): les relations entre les changement des rythmes corticaux et les comportement perceptivo-moteurs

[Jérôme Barral (SSP, UniL)]

L’enregistrement électroencéphalographique (EEG) est une méthode non-invasive qui permet d’étudier la réactivité du signal électrocortical en lien avec des tâches cognitives ou motrices. Ce signal spontané produit par le cerveau correspond à un ensemble d’ondes oscillatoires dont les variations d’amplitude sont interprétées en fonction de la tâche demandées, de la bande de fréquence d’intérêt et de la localisation sur le scalp. Du point de vue expérimental, la qualité de l’interprétation d’un tel signal requiert certaines contraintes méthodologiques. Des exemples sur les mécanismes perceptifs et moteurs viendront illustrer ces aspects de méthodes.
The electroencephalographic recording (EEG) is a non-invasive method that allows studying the electrocortical signal’s reactivity for cognitive- or motor-related tasks. This signal is spontaneously produced by the brain and corresponds to a set of oscillatory waves. Changes in amplitude of these waves can be interpreted according the requested task, the frequency band of interest and the localization on the scalp. From an experimental point of view, the quality of the interpretation of such a signal requires some methodological constraints. This talk will be illustrated with examples from perceptual and motor mechanisms.

Presentation Barral 2013

Presentation on 28.01.2014
Sensitive Questions in Online Surveys: An Experimental Comparison of the RRT and the Crosswise Model

[Ben Jann (UniBE)]

Self-administered online surveys provide a higher level of privacy protection to respondents than surveys administered by an interviewer. Yet, studies show that asking sensitive questions is problematic also in self-administered mode. Because respondents might not be willing to reveal the truth and provide answers that are subject to social desirability bias, the validity of prevalence estimates of sensitive behaviors gained via online surveys can be challenged. A well-known method to combat these problems is the Randomized Response Technique (RRT). However, convincing evidence that the RRT provides more valid estimates than direct questioning in online mode is still lacking. Moreover, an alternative approach called the Crosswise Model (CM) has recently been suggested to overcome some of the deficiencies of the RRT. We therefore conducted an experimental study in which different implementations of the RRT and the CM have been tested and compared to direct questioning. Our study is a large-scale online survey on sensitive behaviors by students such as cheating in exams and paper plagiarism. The results of the study reveal poor performance of the RRT, while the CM yielded significantly higher estimates of sensitive behaviors than direct questioning. We conclude that the CM is a promising approach for asking sensitive questions in self-administered surveys.

Presentation_Jann2014

Presentation on 25.2.2014
Utiliser des méthodes mixtes dans la recherche par étude de cas

[Isabel Valarino (ISS, UniL)]

La présentation porte sur l’utilisation de méthodes mixtes dans la recherche par étude de cas, en se basant sur une recherche qui analyse l’adoption et l’utilisation du congé paternité dans une entreprise romande. Le but est d’une part, à travers cette illustration, présenter quelques uns des enjeux propres à la recherche par étude de cas (ses avantages et inconvénients), ainsi que ceux liés à la sélection du cas d’étude. D’autre part, l’objectif est de montrer l’utilité de recourir à des méthodes mixtes d’analyse (qualitatives et quantitatives) dans la recherche par étude de cas. L’analyse de séquences et de cluster, à partir de données de registre de l’entreprise sur la prise de congé paternité (N=95), est combinée à l’analyse thématique d’entretiens semi-directifs menés avec des employés ayant bénéficié du congé (n=22). Cette combinaison permet d’avoir une compréhension plus approfondie du phénomène étudié, en l’occurrence l’utilisation du congé paternité.

Presentation Valarino 2013

Presentation on 25.3.2014
Mixed methods in action: best and worst practices

[Jörg Stolz (FTSR, UniL)]

Mixed Methods, that is some sort of integration of quantitative and qualitative methods, are increasingly used in the social sciences. While mixed methods seem to be very attractive in theory, to accomplish a successful integration of various kinds of methods and data is often hard. This paper focuses on “best and worst practices” in actual mixed methods research. Using various examples, I argue that successful integration is accomplished by “parallelizing” the design, by using “one logic of inference” and by putting the most important emphasis on the research question (and not the mixing).

Presentation_Stolz2014

Presentation on 29.4.2014
Obtaining Party Positions on Immigration: Comparing Different Methods

[Didier Ruedin (UniNE)]

To what extent it is possible to obtain party positions (on immigration) when there is only a limited amount of political text available? The paper compares different methods (variations of: experts, manual coding, automatic coding), paying special attention to comparability across countries (N=8) and time (N=20).

Presentation Ruedin 2013

Presentation on 27.5.2014
Multiple Imputation of Missing Values: Why, how and the Do’s and Don’ts

[Michael Ochsner (FORS)]

Missing data occur in almost all survey research. Many researchers are often uncomfortable with the treatment of these missing data or ignore the missingness right away. However, inadequate handling of the missing data can lead to biased estimates of parameters such as means or regression coefficients as well as their standard errors and, hence, can distort inference. In this presentation I will demonstrate that “no treatment” of missing data does not exist, in fact, ignoring missingness (i.e. complete case analysis) is a treatment based on rather strong assumptions. I will then focus on multiple imputation as a possible treatment of missing data by introducing the basic idea behind it and presenting some do’s and don’ts of its implementation. Instead of using simulated data, I will work with examples based on ESS data to illustrate the effects of different treatments.

Presentation_Ochsner2014

Presentation on 25.09.2012
Model-based Clustering and Classification; presentation of the Mixmod software
[Gilles Celeux (INRIA, France)]
This talk is devoted to the presentation of the mixture model for cluster analysis, semi supervised and supervised classification. The advantages of this model to answer the selection issues involved in clustering and classification with continuous, discrete or mixed data will be highlighted. EM-like algorithms to derive the parameter estimates will be presented. The numerous versatile and parsimonious useful mixture models will be described. Some emphasis will be placed on the relevant model selection criteria, depending of the focus of the analysis, ground in this model-based setting. The features of the Mixmod software (www.mixmod.org) and its recent interface RMixmod with R (cran.r-project.org/web/packages/Rmixmod/index.html) will be presented.
Presentation Gilles Celeux

Presentation on 30.10.2012
The R-indexes as a complementary measure of data quality? Application to the Swiss ESS 2010 data
[Caroline Roberts and Caroline Vandenplas (FORS/UNIL)]
The only indication of data quality found in many reports is the response rate. It is however known that response rates are a poor estimator of non-response bias and this practice is more and more criticized. Therefore, researchers in the Netherlands have developed an alternative quality measure: R-indexes or R-indicators where “R” stands for representativity. These measures are based on variables known for respondents and non-respondents (typically socio-demographic variables). Since 2010, some FORS surveys have benefited from the Swiss Federal Statistical Office register as a sampling frame. This register contains some information like gender, age or marital status that are thus available for all sampled individuals, respondents and non-respondents. Some of the advantages and remaining issues of this new measure will be illustrated using ESS 2010 data.
Presentation Caroline Vandenplas

Presentation on 27.11.2012
First steps towards a software framework for handling life course survey data in R
[Emmanuel Rousseaux (University of Geneva)]
Population studies strongly rely on survey data and much time is needed to prepare the data. The job is even harder for longitudinal data and network data. From a general point of view, the Dataset project aims at facilitating the management of survey data by providing researchers in social sciences with high-level tools for storing, sharing, exploring and recoding survey data in a secure and efficient way. The toolbox comes as a series of R packages. The software allows to store detailed descriptions for each variable and variable values and then can generate a full codebook directly exported as a PDF file that proves particularly useful for sharing data with others. As key functionalities, the toolbox accepts user-defined missing values, natively account for weights, helps to search for specific variables across the whole data base, and process automatic checks to prevent the loss of representativeness when filtering out cases. The software also provides some specific data-analysis tools such as bivariate association measures, front-end to tree-based methods and a logistic regression method. For all methods provided the toolbox generates nice summary tables easily exportable in PDF format. In this presentation we introduce the main functionalities of the Dataset toolbox, and discuss the functionalities forthcoming for specific handling of panel data, especially for life course analysis. We will use data from the Swiss Household Panel.
Presentation Emmanuel Rousseaux

Presentation on 11.12.2012
Etudier l’engagement militant à l’aide de récits de vie. Quelques réflexions tirées d’une recherche sur l’UDC
[Philippe Gottraux (UNIL)]
Sur la base d’une recherche achevée consacrée à l’engagement militant au sein de l’Union démocratique du centre (Philippe Gottraux et Cécile Péchu, Militants de l’UDC. La diversité sociale et politique des engagés, Lausanne, Antipodes, 2011), nous reviendrons sur diverses opérations de recherche (construction de l’objet ; types de données et modalités d’analyses ; réflexivité, etc.) et accessoirement sur quelques résultats. L’idée sous-jacente sera alors de montrer les apports d’une démarche qualitative pour l’étude d’un tel objet politique, apports trop souvent impensés ou sous-estimés dans la science politique suisse « classique ».
Presentation Philippe Gottraux

Presentation on 29.01.2013
Spatially weighted context data with the R package spacom: Studying the indirect impact of war on well being of young adults in ex‐Yugoslavia
[Sandra Penic (UNIL)]
In multilevel models, geographic space is divided into discrete, often arbitrarily defined units with fixed boundaries. Such a practice ignores the fact that there are usually important social, economic, and institutional ties that link residents from different units. The closer that these units are and/or the stronger the links between their residents, the more they are likely to influence each other. In order to address this issue, a new approach to contextual analysis has been developed: the spatially weighted context data approach. The approach complements classical multilevel analyses, by allowing for the study of spatial dimensions of contextual influences. The R package spacom has been developed by an interdisciplinary team in Lausanne to facilitate the use of the spatially weighted context data approach. Spacom’s functions can be used to construct and describe spatially weighted context data, to introduce spatially weighted contextual indicators in multilevel models (and estimate models through bootstrap procedures that provide robust point estimates and standard errors), and to diagnose spatial dependency in residuals from multilevel models. In my talk, I will illustrate the use of spatially weighted context data with spacom by the study of the impact of collective war victimization and economic exclusion on life satisfaction of young adults in post-war Yugoslavia.
Presentation Sandra Penic

Presentation on 26.3.2013
Cost-benefit estimation when choosing an apprenticeship offer Evidence from a choice experiment
[David Glauser (University of Berne)]
Despite the fact that a high portion of juveniles attending vocational education and training (VET) after compulsory school in German speaking cantons, little is known about why juveniles opt for certain apprenticeships and how they balance pros and cons when they can choose between several apprenticeship offers. The aim of the presentation is to shed light on these questions using a survey-based choice experiment. Data from the DAB-Panel study (http://www.berufswahl.unibe.ch) on vocational and educational decisions at the end of compulsory school is used. The effects of the attributes used in the choice experiment are estimated with conditional logit models.
Presentation David Glauser

Presentation on 30.4.2013
Child development and child care: Marginal treatment effects
[Rafael Lalive (UNIL)]
Many countries are currently expanding access to child care for young children. But are all children equally likely to benefit from such expansions? We address this question by adopting a marginal treatment effects framework. We study the West German setting where high quality center-based care is severely rationed and use within state differences in child care supply as exogenous variation in child care attendance. Data from the German Socio-Economic Panel provides comprehensive information on child development measures along with detailed information on child care, mother-child interactions, and maternal labor supply. Results indicate strong differences in the effects of child care with respect to observed characteristics (children’s age, birth weight and socio-economic background), but less so with respect to unobserved determinants of selection into child care. Underlying mechanisms are a substitution of maternal care with center-based care, an increase in average quality of maternal care, and an increase in maternal earnings.
Presentation Rafael Lalive

[(Flyer 2012/2013)] (PDF – 1 Mo)

[(Presentation Caroline Vandenplas)] (PDF – 13 Mo)

[(Presentation Emmanuel Rousseaux)] (PDF – 599.4 ko)

[(Presentation Lalive)](PDF – 2.9 Mo)

Program Flyer 2011/2012

Presentation on 27.09.2011
Le modèle de Poisson avec sur-représentation de zéros. Une application démographique
Reto Schumacher (SES, Unige)
La modélisation de taux démographiques, définis comme le rapport entre un nombre d’événements démographiques et un nombre de personne-années d’exposition au risque de l’événement en question (event-exposure rate), se fait de manière générale à l’aide de modèles de comptages tels que le modèle de Poisson. Dans cet exposé, il est montré comment la considération de la problématique de la sur-dispersion (overdispersion, ou unobserved heterogeneity), souvent négligé par les démographes, permet d’innover en matière de modélisation démographique. La présentation porte notamment sur le modèle de Poisson avec sur-représentation de zéros. Son application aux taux de fécondité par âge et par rang de naissance atteint permet d’identifier, dans une population transitionnelle, les sous-groupes pratiquant un contrôle d’arrêt de la fécondité.
Presentation Reto Schumacher

Presentation on 26.10.2011
Underrepresentation of Minorities in Swiss Surveys – Evidence to motivate a new Research Initiative
Francesco Laganà (FORS/LINES, Unil)
We present the main results from a FORS-LINES working group that in the last three years has studied the representation of national minorities in general social surveys. Using data from three major Swiss surveys (the Swiss part of the European Social Survey, the Swiss Household Panel, and the Swiss Labor Force Survey), we first address whether national minorities are correctly represented by these surveys and which are the characteristics of the under-represented populations; second we analyze how the solutions proposed in the methodological literature and applied within the surveys correct for such bias. More specifically, we evaluate three potential strategies aimed at reducing bias: strategies to increase response rates, the use of pre- and post-stratification measures taking into account the overall share of foreigners in the national population, and finally offering questionnaires in additional languages. Focusing on cross-sectional bias and attrition, we show that national minorities from non-neighboring countries are largely under-represented and that such under-representation is more severe for those with lower educational levels. We find that the use of measures that are directly targeted at increasing the social and cultural heterogeneity of survey samples are the most promising in order to get an unbiased picture of minority populations. Starting from such findings we present the outline of a more general project aiming at broadening the field of research on other type of minorities that are generally under-represented in social surveys and to open the black box of the different mechanisms producing bias in social surveys.

Presentation on 29.11.2011
Les défis de la complexité : Interdisciplinarité et méthodes mixtes
Michele Poretti, Institut Kurt Bösch
Si l’on s’accorde volontiers sur les opportunités offertes par l’utilisation d’approches interdisciplinaires et de méthodes mixtes – elles permettraient, notamment, de mieux approcher la complexité du réel – les défis pratiques liés à leur mise en œuvre sont rarement analysés. Qu’est-ce que cela veut dire, concrètement, de conduire une recherche interdisciplinaire ? Quels sont les opportunités et les limites liés à l’utilisation de plusieurs cadres théoriques et/ou méthodes ? Comment dépasser les tensions entre différentes conceptions épistémologiques et méthodologiques ? Le papier se propose de contribuer à la réflexion autour de ces questions en partageant l’expérience et les leçons apprises dans le cadre d’un projet de recherche interdisciplinaire financé par le Fonds National Suisse de Recherche Scientifique. Le projet, encore en cours, vise à décrire l’évolution des priorités thématiques de l’agenda international des droits de l’enfant entre 1989 et 2009 et à comprendre les facteurs qui ont influencé les choix des acteurs. Pour répondre aux questions de recherche, le projet a utilisé à la fois des méthodes qualitatives et quantitatives, au point où la frontière entre celles-ci devient difficile à décerner. Tout d’abord, une analyse de contenu herméneutique a été conduite sur un corpus de 327 sources. Ensuite, des perspectives complémentaires ont été acquises à travers des entretiens semi-directifs et en effectuant une analyse du discours sur une portion du corpus. Tout en soulignant l’intérêt de l’approche adoptée dans le cadre du projet, le papier se concentrera sur les méthodes et les outils utilisés, ainsi que sur les défis liés à l’utilisation conjointe de méthodes quantitatives et qualitative.

Presentation on 24.01.2012
Analyse de dispersion de séquences d’états: analyser les liens entre trajectoires et variables explicatives
Matthias Studer (SES, Unige)
L’analyse de séquences est devenue l’une des méthodes privilégiées pour analyser des trajectoires. Elle permet notamment d’étudier dans une perspective holistique les patterns récurrents et la prise en compte de la multiplicité des états possible. Techniquement, cette analyse repose sur une mesure de distance entre trajectoires, ce qui permet de les comparer. En pratique, ces distances sont souvent utilisées pour construire une typologie des trajectoires et identifier des trajectoires types. Au-delà de l’approche descriptive, on s’intéresse généralement à identifier les facteurs qui influencent la construction de la trajectoire. Pour ce faire, il est d’usage de mettre en relation les types obtenus avec d’autres facteurs d’intérêt, tels que le sexe, à l’aide de régressions logistiques ou de tests d’association. Cependant, en se centrant sur les types de trajectoires, on perd de l’information, ce qui peut conduire à des conclusions trompeuses. Nous présentons ici un ensemble de méthodes qui permet d’analyser les liens entre les séquences d’une part et un ou plusieurs facteurs explicatifs d’autre part. Originellement utilisées en écologie, ces méthodes reposent sur la définition d’une mesure de dispersion des séquences et sur une généralisation des principes de l’analyse de variance (ANOVA) à tous types de dissimilarités. Conceptuellement, ces méthodes permettent un changement de paradigme. Plutôt que de se baser sur la recherche de modèles de trajectoires, nous considérons qu’elles sont insérées dans des contextes multiples qui peuvent influencer — chacun à sa manière — la construction d’une trajectoire. Ces méthodes complètent donc l’analyse de séquences traditionnelle, principalement exploratoire, par une approche confirmatoire.
Presentation Matthias Studer
Analyse R Matthias Studer

Presentation on 28.02.2012
Endogeneity: An overlooked threat to validity of cross-sectional research
John Antonakis (Unil/HEC)
John Antonakis will present on the problem of endogeneity. His presentation will focus on a assumption of regression analysis (or structural equation modeling) regarding the fact that the modeled independent variables are assumed to be exogenous (i.e., random and independent of omitted causes from the model). Yet, the problems of endogeneity are not well known to researchers working in many social sciences disciplines. Researchers often conduct studies to determine the relationship between a supposed independent variable, x (e.g., an attitude, choice) and a dependent one, y (e.g., outcome). Estimating such a model—where x has not been exogenously manipulated—may lead to spurious findings because the supposed independent variable may depend on a variable or variables that predict y and correlate with x too. In this seminar, he will give a brief and vivid overview to endogeneity and the conditions under which it will be engendered. He will show, using non-technical language and intuitive explanations that in the presence of endogeneity the observed relationship that is estimated can be very misleading. He will briefly demonstrate how this problem can be solved using procedures borrowed from econometrics. Finally, he will briefly report as to prevalence of endogeneity in a social science discipline using papers published in top journals the previous 10 years, showing that that researchers fail to address at least 66% and up to 90% of design and estimation conditions that that make model estimates invalid.
Presentation John Antonakis
Podcast Endogeneity

Presentation on 27.03.2012
Quality of Online-Panels: An Experiment with Propensity Score Adjustments
Beat Hulliger (School of Business, FHNW)
A battery of possible predictors for the participation in an online-panel was investigated in an experimental recruiting survey. The predictors were based on values, personality traits and rational choice theories. The recruiting survey used a random-quota sample and was carried out with CATI. The later participation in the online-panel of LINK Institute for Market and Social Research was registered and served as the response variable for propensity score models. General linear models and classification trees were used. The objective was to find a model which uses at most five variables in order to establish a parsimonious set of covariates which would be used routinely in recruiting surveys. Direct propensity score adjustment and propensity score matching were tested by comparing the distribution of control variables with the distribution in the recruiting survey. No universal reduction of bias could be found: Clear improvements were only visible, when a variable was heavily biased by differential panel participation and a good model for the panel participation could be established.

Presentation on 24.04.2012
Transnationalism and cosmopolitanism explored through the autobiographical narrative interview methodology
Claudia Vorheyer (UZH)
The talk deals with the qualitative method of narrative interviews. First of all the typical processes of conducting and analyzing narrative interviews are demonstrated. In addition to that its special epistemological potential will be emphasized and using the example of exploring transnationalism and cosmopolitanism demonstrated. Therefore we present our currently research project about the group of so-called “Transnational Mobiles”. The latter are people who were living in at least three different countries for a significant period of time. Here we are interested in their social and biographical backgrounds and related to that the motifs of their transnational careers. Moreover the beginning and further trajectories of their transnational biographies are reconstructed. Finally we focus on their identity formations and cosmopolitan orientations. As a matter of fact investigating transnationalism and cosmopolitism as a mode of self-transformation the use of autobiographical narrative interview methodology allows particularly new insights.

Presentation on 29.05.2012
Effects of prepaid monetary incentives on mail survey response rates and on self-reporting about delinquency – Empirical findings
Rolf Becker (Unibe)
In the context of a survey on criminal behavior a methodological experiment has been carried out in order to investigate whether mail survey response rates can be influenced by monetary incentives. Prepaid monetary incentives are expected to elicit higher response rates, as predicted by social exchange and subjectively expected utility theory. This is true especially in a survey on the respondents’ delinquent behavior, since prepaid incentives are likely to strengthen the respondents’ trust towards the researcher. For surveys in which money is promised once the questionnaire is returned, it is assumed that the response rates to be lower. The empirical findings clearly support our assumptions, even if the social mechanisms of reciprocity which underlie response behavior cannot be observed directly. The analysis also provides that monetary incentives neither evoke desirability nor sponsorship effects which would cause over-reporting of criminal activities. Finally, our results show no systematic variation regarding socio-structural characteristics and correlates of delinquent behavior.

Presentation on 05.06.2012 Please note the different time and location: 11:00-12:15 in room 3032 Anthropole
L’analyse QCA: principes, applications et potentiel en sciences sociales
Benoît Rihoux (Université catholique de Louvain)
L’analyse ’QCA’ (Qualitative Comparative Analysis) s’est développée à la fois comme une approche et comme une série de techniques depuis la fin des années ’80, au confluent de la macrosociologie historique et de la politique comparée. Cet exposé vise, tout d’abord, à présenter les fondements de QCA comme approche, avec ses objectifs, postulats, et conception de la causalité. Il sera démontré que cette approche emprunte à la fois aux approches centrées sur les cas (‘qualitatives’) et centrées sur les variables (‘quantitatives’). Dans un deuxième temps, les principaux usages et techniques de QCA seront présentés ; il sera démontré que QCA peut être utilisé de différentes manières (plus inductives ou hypothético-déductives), en exploitant différents types de données, et suivant des designs de recherche différentes (des « petits N » aux « grands N »). Enfin, quelques tendances récentes dans l’exploitation de QCA seront discutées, telles que la diversification disciplinaire, l’analyse de données à différents niveaux (dont le niveau micro), ou encore la combinaison de QCA avec d’autres techniques.
Presentation Benoît Rihoux

[(Program Flyer 2011/2012)] (PDF – 1 Mb)

[(Presentation Reto Schumacher)] (PDF – 241.3 kb)

[(Presentation Matthias Studer)] (PDF – 1012.4 kb)

[(Analysis Studer)] (texte – 7.6 kb)

[(Presentation John Antonakis)] (PDF – 425.6 kb)

[(Presentation Benoît Rihoux)] (PDF – 136.1 kb)

Program Flyer 2010_2011 (PDF – 762.8 ko)

Presentation on 5.10.2010
Des jeux de hasard à l’évidence statistique : quelques repères historiques
Christiane Ruffieux (CHUV, Lausanne)
Abstract Christiane Ruffieux (PDF – 33 ko)
Presentation Christiane Ruffieux (PDF – 220 ko)

Presentation on 2.11.2010
Evolution d’une structure de pensée Dynamique de l’émergence des représentations sociales
Henry-Arnaud Thévenet (Faculty of Social and Political Sciences, UNIL)
Abstract_Thevenet (PDF – 31.4 ko)
Presentation Henry-Arnaud Thévenet (PDF – 1.2 Mo)

Presentation on 7.12.2010
Maximiser a posteriori la comparabilité de données quantitatives : une illustration avec les violences envers les femmes
Véronique Jaquier (Institut de criminologie et de droit pénal, UNIL)
Abstract Véronique Jaquier (PDF – 31.6 ko)
Presentation Véronique Jaquier (PDF – 364.8 ko)

Presentation on 25.1.2011
Évaluation de l’efficacité des psychothérapies : Apports et limites de la méta-analyse
Grégoire Zimmermann (Faculté des sciences sociales et politiques, UNIL)
Abstract Grégoire Zimmermann (PDF – 31.7 ko)
Presentation Grégoire Zimmermann (PDF – 2.4 Mo)

Presentation on 22.2.2011
Recent developments in the sample survey methodology
Yves Tillé (Institut de statistique ; UNINE)
Abstract Yves Tillé (PDF – 30.3 ko)
Presentation Yves Tillé (PDF – 408 ko)

Presentation on 29.3.2011
Analyse de discours : Théorie et pratique
Véronique Mottier (Faculté des sciences sociales et politiques, UNIL)

Presentation on 19.4.2011
Using a Multilevel Structural Equation Modeling Approach to Explain Cross-Cultural Measurement Noninvariance
Eldad Davidov (Institute of Sociology, UNIZH) and Hermann Dülmer (University of Cologne, Germany)
Abstract Eldad Davidov (PDF – 30.9 ko)

Presentation on 31.5.2011
Introduction aux modèles Exponential Random Graph (ERGM)
Victorin Luisier (Faculté de Psychologie et des Sciences de l’Éducation, Université de Geneve)
Presentation Victorin Luisier (PDF – 2.3 Mo)
Abstract : Comment expliquer que dans un réseau de relations, comme le reseau d’amitié d’une classe d’élève, certaines paires d’individus entretiennent une relation, alors que ce n’est pas le cas des autres ? Derrière cette question intuitive se cache en fait un challenge méthodologique et statistique.

En effet, si l’existence d’une relation interindividuelle peut s’expliquer le biais de variables individuelles (ex : âge), elle dépend aussi d’autres relations dans le réseau (ex : réciprocité, amis en commun). Ces autres relations créent une structure d’autodépendance rendant impossible l’utilisation de modèles statistiques standards.

Les ERGM ont quant à eux justement été développés afin de modéliser l’effet de variables individuelles et de la structure d’autocorrélation sur un réseau de relation. A l’aide d’un réseau d’amitié mesuré dans un gymnase (collège) suisse, nous introduirons ces modèles ERGM. Nous nous intéresserons plus particulièrement à l’effet du genre sur le réseau d’amitié de ce collège.

Plus d’informations
Kathrin Kissau
+41 (0)21 692 37 49
kathrin.kissau@fors.unil.ch

Seminar on Tuesday, 6.10.2009, Vidy 209, at 12:15
New techniques for longitudinal analyses
Florent Dieterlen (UNIL)
A mathematical modelling method is presented, based on dynamical systems. The method consists in constructing a dynamical system (differential equations system) from historical data or from questionnaires. The historical data can be of two types : a) Average data for variables for a population (example : Swiss statistical office (OFS) data for unemployment, voting, etc., for Swiss population) for several years, at regular intervals if possible. b) Data from variables for individuals, countries, etc., which are then stacked in the method, in order to have an average evolution. Here, the averaging is done within the method, instead of being done before (example : FORS data for individuals).

Seminar on Tuesday, 3.11.2009, Extranef 110, at 12:15
TraMineR : un outil pour l’exploration de données séquentielles catégorielles
Gilbert Ritschard (Dpt d’économétrie et Laboratoire de démographie, UNIGE)
TraMineR est une librairie R pour la fouille, la description et la visualisation de séquences d’états et/ou d’événements, et plus généralement de données séquentielles discrètes. Bien que s’appliquant à toutes sortes de séquences chronologiques ou non (textes, séquences ADN, etc.), les fonctions et graphiques proposés ont été conçus essentiellement pour l’analyse de données biographiques longitudinales en sciences sociales, telles que les données décrivant des carrières professionnelles ou des trajectoires de vie familiales. la présentation commencera avec un aperçu des multiples possibilités que propose la librairie aussi bien pour l’exploration de séquences d’états (visualisation, caractéristiques longitudinales des séquences, description synthétique d’un ensemble de séquences, calcul de dissimilarités entre paires de séquences, analyse de variance de séquences, détermination de séquences représentatives) que pour la fouille de séquences d’événements (extraction des sous-séquences fréquentes et parmi elles des sous-séquences les plus discriminantes entre groupes). En s’appuyant sur les données de l’enquête rétrospective réalisée en 2002 par le Panel Suisse de Ménages (PMS), on procède ensuite à une analyse de l’évolution des trajectoires cohabitationnelles et occupationnelles au cours du 20e siècle en Suisse.

Seminar on Tuesday, 1.12.2009, Vidy 209, at 12:15
Mixing modes of data collection in social surveys
Caroline Roberts (MISC, UNIL)
Survey researchers are faced with a number of choices when deciding how to design their studies. Many of these decisions are guided by resource considerations, but the choices made can have important implications for the overall quality of the data collected. In particular, selecting a mode of data collection (e.g. whether to administer the questionnaire by interviewer, in person or over the telephone, or to have respondents read and complete the questions themselves, on paper or on the Internet), is not only relevant to survey costs, it can also influence a range of outcomes relevant to data quality, including the extent of population coverage provided, how well the achieved sample represents the population, and the accuracy of the answers given. Because modes vary by how much they cost and by their unique measurement properties, each one offers different advantages and disadvantages for different types of study. Using a combination of modes in a single study makes it possible to offset the weaknesses of one method with the strengths of another and increasingly, researchers are looking to exploit this potential of mixing modes as a way to reduce specific types of survey error and control survey costs.

However, combining modes has the effect of confounding different sources of error in the data, and the resulting non-equivalence can affect the validity of conclusions drawn by analysts. In this presentation, I describe the challenges to survey quality involved in mixing modes, and in particular, the difficulty of assessing the effects of mixing modes on measurement. I review theories about why different modes of data collection can lead to differences in survey responses and describe the methods typically used to assess mode effects. I then discuss some of the drawbacks associated with these methods, including whether or not mode effects are detected in the first place, and where they are, how to decide whether they matter in practice. The issues raised are illustrated with examples from experimental research conducted by the coordinating team of the European Social Survey, designed to inform decisions about whether to allow mixed mode data collection in its future rounds, but the conclusions drawn have implications for all researchers considering the possibility of mixing modes in their own surveys, or analyzing data from a mixed mode study.

Seminar on Tuesday, 23.2.2010, Vidy 209, at 12:15
’Spatial Metaphors’ or how to make Political Landscapes and Social Spaces Visible
Michael Hermann (SOTOMO, UNIZH)
Maps are powerful instruments to identify and communicate spatial relations. However, they do not necessarily need to be restricted to geographic space, they may also depict abstract metaphorical spaces. Instead of the cardinal directions north, west, south and east metaphorical maps display socio-economic, cultural or political areas of conflict. Although metaphorical maps are maps without direct geographical spatial reference, they are of particular interest for spatial analysis. To visualize geographical entities such as local authorities or statistical zones not in their geographical but in their socio-economic context, means offering new perspectives on the spatial dynamic of development and helps to improve the understanding of its social dimension. With the help of our analysis of Switzerland’s social and political space I will show in the presentation, how to appropriately generate and use metaphorical maps.

Seminar on Tuesday, 30.3.2010, Extranef 110, at 12:15
Traitement de données manquantes : comparaison de plusieurs méthodes
Mélanie Cicognani (IMA, UNIL)
Lors du traitement statistique de données, les données manquantes constituent un problème majeur, puisque l’information est incomplète, ajoutant de l’incertitude. Comme première approche, nous avons comparé par le biais de simulations numériques différentes méthodes existantes pour le traitement de données manquantes. En partant d’un fichier sans aucune donnée manquante, nous avons créé neuf scénarios variant en fonction du nombre de données manquantes et de leur type. Mille ensembles de données ont été générés à partir de chaque scénario, puis les données manquantes ont été traitées selon différentes approches. Les moyennes, écarts-types et corrélations entre variables imputées ont été comparés avec le fichier original sans données manquantes. L’influence du traitement des données manquantes sur un modèle de régression a aussi été évaluée. Nos résultats montrent que les méthodes basées sur l’imputation multiple sont globalement les meilleures. D’autres méthodes, comme par exemple l’imputation simple par régression, permettent aussi l’obtention de résultats intéressants, mais seulement dans certaines situations particulières. Durant ce séminaire, nous aborderons quelques points importants à l’étude des données manquantes ainsi que certaines méthodes, puis nous examinerons les résultats des comparaisons effectuées par simulations.

Seminar on Tuesday, 27.4.2010, Vidy 209, at 12:15
Configurations familiales et santé mentale : une application de l’analyse de réseaux à un suivi longitudinal de patients
Marlène Sapin (FORS)
Cette contribution cherche à explorer l’association entre insertion familiale et santé mentale, une soixantaine d’individus d’un cabinet de psychothérapie privé ont été interrogés tous les trois mois, au cours d’un suivi d’une année et demi, sur leur réseau familial. Ils ont également répondu à un instrument mesurant leurs symptômes psychiques. Les résultats montrent qu’il existe une pluralité d’insertions familiales et de fonctionnements relationnelles, inégalement associés à la détresse psychique des individus. Se basant sur cette recherche, l’objectif de cette présentation est tout d’abord de présenter l’approche, consistant à mesurer l’insertion des individus dans leur réseau personnel (défini par Krackardt (1987) comme « cognitive networks ») ainsi que la méthode utilisée pour déterminer les différentes insertions familiales et fonctionnements relationnels. Il est en second lieu de discuter les points forts et faibles de l’approche et plus spécifiquement de la dimension longitudinale. L’importante variabilité des insertions familiales et des dynamiques relationnelles trouvées chez ces patients fragilisés au cours des vagues a rendu l’analyse longitudinale complexe. Une application de différentes techniques, que j’aimerais discuter, a permis d’identifier des trajectoires d’insertion familiale et de fonctionnement relationnel au cours du suivi.

Seminar on Tuesday, 25.5.2010, Extranef 110, at 12:15
Conducting Research on the Internet
Ulf-Dietrich Reips (Dpt. of Social and Economic Psychology, University of Bilbao)
Methods for Internet-based research are currently one of the hot areas in methodology. Within its brief history (ca. fifteen years), the field has seen a massive increase in the number of studies conducted on the Internet, marking a grass-roots change in how research in the behavioral and social sciences often is conducted. For examples of Web experiments and surveys see the Web experiment and Web survey lists at http://wexlist.net and for tools to conduct Internet-based research see the iScience Server at http://iscience.eu. I will present and discuss findings on methods, techniques, and tools in Internet-based research and will be happy to offer and discuss solutions to many of the imagined and real challenges in using the new methodology.

Seminar on Thursday, 19.3.2009, Internef 125, at 12:15
R et analyses textuelles
Jean-Pierre Müller (IMA)
Des presque 1700 packages créés pour R, une quinzaine ont un lien direct ou indirect avec l’analyse de données textuelles. La première partie de ce séminaire sera consacrée à la présentation et l’exploration de quelques-uns, dont “RQDA”, destiné à l’analyse qualitative, mais aussi de “tm”, “lsa” et “Rweka” destinés à l’analyse quantitative. La seconde partie proposera des solutions aux différents problèmes et difficultés rencontrés lors de l’utilisation de ces librairies sur des textes en français dans le cadre des sciences humaines.

Seminar on Thursday, 2.4.2009, Vidy 209, at 12:15
Panel attrition on the individual and the household level
Oliver Lipps (FORS)
Attrition is mostly caused by not contacted or refusing sample members. On one hand it is well-known that reasons to attrite due to non-contact are different from those that are due to refusal. On the other hand does noncontact most probably affect household attrition, while refusal can be effective on the level of both households and individuals. Knowledge of possible compensating or reinforcing effects of selectivity may give hints how to improve communication, incentive schemes, and fieldwork on both levels in household panel surveys. In this seminar, attrition on both the household and (conditional on household participation) the individual level is analysed in three panel surveys from the Cross National Equivalent File (CNEF): the German Socio- Economic Panel (GSOEP), the British Household Panel Study (BHPS), and the Swiss Household Panel (SHP). To follow households over time, we use a common rule in all three surveys. First, we find different attrition magnitudes and patterns both across the surveys and also on the household and the individual level. Second, there is more evidence for reinforced rather than compensated household level selection effects if the individual level is also taken into account.

Seminar on Thursday, 23.4.2009, Internef 125, at 12:15
Event History Analysis and diffusion models
Jean-Marie Le Goff (ITB)
Les méthodes d’Event History Analysis ont pour objectif l’analyse de la distribution du temps d’occurrence d’un événement dans une population, alors que les modèles de diffusion visent à analyser la diffusion d’une innovation, d’une rumeur ou d’une pratique de personne à personne. Plusieurs travaux ont proposé de faire un lien entre les deux types d’approche, l’événement étant alors l’adoption par une personne de l’innovation (Dieckmann 1987, Greve, Tuma & Strang, 1995). La présente communication aura deux objectifs. Le premier est de spécifier l’un des modèles de diffusion, le modèle d’influence mixte, dans le corpus des méthodes d’Event History Analysis (Mahajan & Peterson, 1985). Le second est de porter un intérêt sur la manière d’appréhender les personnes qui n’adoptent pas l’innovation.

Seminar on Thursday, 7.5.2009, Vidy 209, at 12:15
Spatially weighted contextual variables in multi-level analyses: A new avenue for comparative survey research?
Guy Elcheroth (MISC)
Survey research or electoral studies have repeatedly shown that political attitudes and behaviour are influenced by collective economic and social circumstances, rather than the personal situation of individual citizens. This has lead to an increased interest in multilevel analyses, which are now widely used as a privileged tool for studying the effect of nation-level indicators on individual survey responses. Regrettably, these mainstream procedures convey a series of limitations for creative approaches to contextual data generation, mainly because of highly restrictive implications regarding eligible micro-level data sources. Further, they automatically introduce assumptions about the relationship between the communitarian backgrounds of collective experiences and social representations, which are theoretically undefined. In order to over-come part of these shortcomings, we have started to develop an innovative procedure for generating contextual variables that are spatially weighted. The starting point of our approach is close to geographically weighted regression analysis, sharing its objective to define open-ended social contexts. However, we extend the core notion of (social) distance/proximity to multifaceted social interdependencies and influences, encompassing the territorial, symbolic, and practical dimensions of social communities. A substantive research application, based on comparative survey data collected within transitional societies of the former Yugoslavia, will be presented.

Seminar on Thursday, 28.5.2009, Internef 125, at 12:15
Survey sampling package for R
Alina Matei (UNINE)
L’Institut de Statistique de l’Université de Neuchâtel a entrepris depuis trois ans la réalisation d’un logiciel libre permettant de traiter les enquêtes par sondage au moyen des méthodes statistiques les plus modernes. Initialement, ce projet a vu le jour afin de servir d’outil pédagogique pour des cours avancés sur les méthodes d’échantillonnage organisés par l’Office Fédéral de la Statistique sous l’égide d’Eurostat et de l’Association Européenne de Libre Echange (AELE). Ces cours étaient destinés aux statisticiens des instituts de statistique des pays européens et des pays de la Méditerranée. Aujourd’hui, ce projet, poursuivi dans le cadre de la collaboration avec l’Office Fédéral de la Statistique, a dépassé le cadre strictement pédagogique. L’intérêt de ce logiciel (intitulé ’sampling’) est d’être écrit au moyen du logiciel libre et gratuit R. Il permet de sélectionner des échantillons selon plusieurs méthodes, de traiter les problèmes de non-réponse, d’ajuster des données d’enquêtes sur des données de recensement, et d’évaluer la précision des estimations ainsi obtenues.