Our internal meetings are held on Mondays at 16h00 on the RECSM facilities, unless otherwise stated. In them, members of the RECSM and guests present their current research to colleagues. Public is invited to attend prior confirmation to recsm@upf.edu.
PROGRAM
PAST MEETINGS
January 31, 2012 - 09h30
"Giving the right options, getting better responses: Split Ballot MTMM Experiments on Response Options in Surveys"
by Sedef Turper (University of Twente)
Given the increasing use of survey data the need for improving the data quality and therefore the need for methodological studies gain even more importance. While the need for enhancement of statistical analysis research has been acknowledged in the field, studies on the data collection still requires further scholarly attention. While designing the survey questionnaires researchers need to make many choices on various aspects of a survey item. These decisions can concern various aspects such as the wording and ordering of the response requests, general layout of the questionnaire form, data collection method and response choices offered for each survey item. The current study intends to elaborate on the choices made regarding the response options. Accordingly the preliminary results from the split ballot experiments on the response scale length, ordering of responses in ascending and descending order, the labeling of the response options and lastly constructing response scales in unipolar and bipolar format will be presented.
October 10, 2011 - 12h00
"The potential outcome framework for causal inference in observational studies: multilevel settings and interference"
by Bruno Arpino (Universitat Pompeu Fabra)
In the social sciences, experiments are very rare because of ethical or practical reasons and the estimation of causal effects of treatments/programs has to rely on observational studies. The validity of inference will then strictly depend on the plausibility of the assumptions underlying the employed statistical techniques. In the first part of the talk, I will introduce the potential outcome framework for causal inference and provide some examples of its application in the social sciences. Then, I will discuss two specific developments: the estimation of causal effects with propensity score matching in multilevel settings and in the presence of interactions among units.
July 8, 2011 - 16h00
"Assessing Bias Against Muslims and Christian Immigrants: Results from a List Experiment"
by Mathew J Creighton (Universitat Pompeu Fabra) and Amaney Jamal (Princeton University, U.S.A.)
Migrants could be subject to targeted and generalized opposition by a receiving society. Targeted opposition is characterized by opposition to particular immigrant groups based on observed, partially observed or unobserved characteristics of the migrant population. Generalized opposition does not distinguish between groups and is uniformly expressed toward all immigrants. Measuring and distinguishing between targeted or generalized opposition is problematic as respondents to survey questions that could be interpreted as intolerant often mask their true opinion. Termed “social desirability bias”, this difference between held and expressed opinions inhibits accurate assessment of a variety of controversial topics. We use an experimental design and recently developed multivariate techniques to test at a population level for both the presence and magnitude of social desirability bias. We focus on opposition to citizenship for two particular immigrant populations in the United States – Muslims and Christians. We find that U.S. citizens openly express significantly greater opposition to citizenship for Muslim immigrants relative to Christian immigrants. However, when respondents are allowed to mask their opinions, we find no evidence that Muslim immigrants are targeted, suggesting that opposition to citizenship for Muslims is not more widely held, but is more openly expressed.
July 6, 2011 - 16h00
"Measuring the Nations that Exist"
by Eduard Bonet (Universitat Pompeu Fabra)
This article proposes a new measurement model of national identities with survey data. The literature on national identities points to the fact that national identities are formed by a combination of political and cultural traits, ethnic and civic features. Nevertheless it is unclear how that combination is articulated in a measurement model that captures the collective aspect of national identities and the individual one. The author suggests that for any given nation we should expect competing views on the nation among its citizens, but also there should also be certain degree of commonality in the way citizens conceive their nation. Results of empirical research with survey data are ambivalent and technically inadequate. This article presents a cumulative unidimensional measurement model to build a scale of national identity that combines ethnic and civic traits and, more importantly, identifies the shared and the disputed traits among the citizens. The model, a nonparametric probabilistic Mokken model, is tested for 31 countries using the eight classical ISSP items used to measure civic-ethnic national identities. The results show that unidimensional cumulative scales can be formed in 22 countries combining ethnic and civic items in each. The article also points towards the weaknesses of the ISSP items and comments on the implications of the findings.
May 23, 2011 - 16h00
"Cultural Frame Schitching: How Culture and Language Cues Affect the Behavior of Biculturals and Bilinguals"
by Veronica Benet-Martínez (Universitat Pompeu Fabra)
This presentation will review socio-cognitive experimental work showing that individuals can simultaneously hold two or more cultural orientations and shift between them in response to cultural/linguistic cues, a process called cultural frame-switching (CFS; Hong, Morris, Chiu, & Benet-Martinez, 2000). CFS effects have been documented in multiple behavioral domains such as attribution (Benet-Martínez, Leu, Lee, & Morris, 2002), personality self-views (Ramirez-Esparza, Gosling, Benet-Martínez, & Pennebaker, 2006), ethnic identity (Verkuyten & Pouliasi, 2002), emotional experience (Perunovic, Heller, & Rafaeli, 2007), values (Verkuyten & Pouliasi, 2006), cooperation (Wong & Hong, 2005), auto-biographical memory (Bender & Ng, 2009), and decision making (Briley, Morris, & Simonson, 2005) among others. CFS is usually unconscious and automatic, and in other for a particular cultural cue to influence behavior, the relevant cultural schemas have to be (1) cognitively available (i.e., the individual has internalized values, norms, attitudes, and emotional associations relevant to that culture/language), (2) cognitively accessible (the schemas have been recently activated by explicit or implicit contextual cues), and (3) applicable to the situation.
May 16, 2011 - 10h00
"From Documenting Survey Questionnaires to Documenting Survey Data Sets"
by Rainer Schnell (University of Duisburg-Essen, Germany)
In a series of projects funded by the German Research Foundation a Java-based tool for documenting the development of questionnaires has been developed by a workgroup lead by the author. This questionnaire documentation system (QDDS) has been extended to allow the search on multiple QDDS-Files. This web-based feature is intended to search questions in large scale survey depositories. More than 50 surveys have been entered into the database up to now. Features and limitations of QDDS will be described in details. QDDS was not indended for data analysis support in daily research practice on existing data bases. Such an universal tool for searching in questionnaires, meta-data and data files suited for complex data sets like the ESS or GSOEP is still missing. The talk will include a description of a recent research proposal for the development of such a tool.
April 27, 2011 - 15h00
"Robust Estimation of Poverty and Inequality Indicators with SILC Data"
by Beat Hulliger (University of Applied Sciences Northwestern Switzerland, Switzerland)
The European indicators on poverty and inequality are based on the incomes collected through the surveys of the Statistics on Income and Living Conditions (EU-SILC). The income measured by SILC is composed by numerous components at person and household level. These income components are subject to intensive statistical data preparation and finally aggregated to the personal disposable income. The income components and the disposable income still contain outliers and robust methods should be used to estimate poverty and inequality indicators. Robust methods developed under the EU FP7 project “Advanced Methodology of European Laeken Indicators”, in particular the Quintile Share Ratio, and simulation results are presented.
April 11, 2011 - 17h00
"Life satisfaction and political participation: discussing a preliminary model"
by André Pirralha (RECSM - Universitat Pompeu Fabra)
A growing literature across the social sciences examines how political factors contribute to individuals’ happiness or subjective life satisfaction. Amongst these studies, the impact of individuals’ political engagement on their satisfaction has received comparatively little empirical attention. Furthermore, the very small literature that have assessed the relationship between political participation and life satisfaction have mostly come to inconclusive results. In this presentation, we will propose a first version of a model designed to explain the link between political participation and life satisfaction, using data from a Russian panel study.
April 11, 2011 - 16h00
"An analytical Framework for the study of Change in the National Identification"
by María José Hierro (RECSM - Universitat Pompeu Fabra)
This paper focuses on the study of the changes that occur in the individuals' national identification over time. The paper qualifies recent research on the study of ethnicity, and proposes a definition of individual national identification. From here onwards, the paper examines the conditions which make the individuals' national identification susceptible to change, and defends that research dealing with the change in the national identification should distinguish four kinds of changes: changes in the meanings individuals attribute to their national identification, changes in the individuals' category of self-identification, changes in the salience of the national identification, and finally changes in the strength of the attachment. All in all, this paper aims at offering an analytical framework that contributes to improve our understanding of individuals' national identification and helps researchers to examine the causes of change. Case studies and historical examples illustrate the paper's argument, thereby enriching the analytical framework.
March 28, 2011 - 16h00
"Avoiding deviations in questionnaires across countries through SQP"
by Diana Zavala Rojas (RECSM - Universitat Pompeu Fabra)
SQP is a semi automatic program for the prediction of the quality of survey items. It is also helpful to detect deviations in formal (no-linguistic) characteristics across countries due to differences in translation. Previous research shows that the deviations detected by SQP affect the comparability among countries and the quality of the instruments. Correction of those deviations before field work decreases this source of error. This presentation shows some findings from the coding process of 27 items of the main and supplementary questionnaires used in the Round 5 of the ESS in 20 countries.
March 21, 2011 - 16h00
"The split-ballot mutlitrait-multimethod approach: The importance of the third group"
by Melanie Revilla (RECSM - Universitat Pompeu Fabra)
In 2004, Saris, Satorra and Coenders proposed a new approach to estimate the quality of survey questions, combining the advantages of two existing approaches: the multitrait-multimethod (MTMM) approach and the split-ballot (SB) approach. Implemented in practice, this approach led to frequent problems of non-convergence and improper solutions. This paper uses Monte Carlo simulations to try to understand why the SB-MTMM is working well in some cases but not in others, and when it does not work properly, it looks for solutions. It shows that increasing the sample size may solve the problems, but sometimes the needed sample size to get proper estimates is really large. So we conclude that a 3-group design should be used as often as possible. Over many replications, even an extremely small third group improves a lot the performance of the estimation, but when focusing on one specific dataset, it is not enough to guarantee proper estimates, and a larger third group is needed.
November 9, 2010 - 15h00
"Understanding subjective wellbeing"
by André Pirralha (RECSM - Universitat Pompeu Fabra)
This presentation will provide an overview of the growing literature on subjective wellbeing, also commonly known as "happiness". Wellbeing has been traditionally measured by objective dimensions such as material progress, defined by income or GDP. Presently however, wellbeing is seen as a multi-dimensional concept, evolving several aspects of human life that cannot be captured by just using the more traditional objective measures. The main findings about the determinants of subjective wellbeing will be discussed, as well as some avenues for further research.
October 26, 2010 - 16h00
"Quality in unimode and mixed-mode designs: a multitrait-multimethod approach"
by Melanie Revilla (RECSM - Universitat Pompeu Fabra)
Most surveys used so far face-to-face or telephone questionnaires in order to collect data. But the costs of achieving a survey using these traditional modes increase. At the same time, the response rates decrease, making the idea of switching mode very attractive. Because each mode has its own drawbacks and advantages, the idea of mixing modes of data collection is becoming more and more popular. Nevertheless, combining different modes of data collection may be problematic if people answer differently depending on the mode. Also, a switch from a unimode to a mixed-mode design may threaten the comparability of the data across time. This paper focuses on comparing the quality of questions asked in a unimode and two mixed-mode surveys. Using data of the European Social Survey (ESS) in the Netherlands, and following a MultiTrait-MultiMethod approach (MTMM), few differences are found between the unimode and mixed-mode designs in terms of quality. Looking at the differences across modes lead to slightly less similarities, but overall the quality does not change much.
September 22, 2010 - 17h00
"The Transformation of Social Democracy"
by Iseult Larkin (Universitat Pompeu Fabra)
This paper examines how Social Democratic parties in Denmark, Sweden, Norway, the Netherlands, Austria, Germany, the UK, Greece, Spain, Portugal and France responded to societal, political and economic developments in the 1970s and 1980s, transforming into catch all-parties which appeal to the centre rather than defined class constituencies. The paper argues that voter dissatisfaction with this transformation has been expressed by increasing support for parties at both the extreme left and right of the political spectrum. It suggests that the current decline of many social democratic parties is part of a long-term trend that may be understood as a result of the changed ideological nature of the parties.
January 18, 2010 - 16h00
"Attitudes Toward Democracy and the Mechanisms of Voting Intermediation"
by Gerardo Maldonado (RECSM - Universitat Pompeu Fabra)
We study how some individual political attitudes might affect substantially how different citizens face the electoral process and the corresponding flow of political information, producing different types of voters and citizens. Our main question is: are critical citizens (Dalton, 2004; Norris, 1999 and 2010) or disaffected democrats (Torcal, 2002 and 2007) more responsible citizens, having a more active role in searching for information, obtaining more heterogeneous sources of information, and controlling better, as a result, incumbent representatives? Based on data from the Comparative National Election Project (CNEP), first, we construct three different typologies of citizens based on some of the most well know political attitudes ---support for democracy, democratic satisfaction, and political disaffection--- and, second, we analyse how the quantity and plurality of exposition to political intermediation are related with these typologies and the level of individual political knowledge.
December 14, 2009 - 16h00
"The relationship between issues and left-right ideology"
by Wiebke K. Weber (RECSM - Universitat Pompeu Fabra)
The major body of the literature on the relationship between the left-right self-orientation and its value or issue-based component assumes that individuals’ values and attitudes towards different issues will determine their position on the left-right dimension. Regarding values, it is assumed that these are stable over (a long period of) time and hence, affect the left-right self-orientation. Yet, the impact of issues seems to be less clear, in particular as it was found that issues differ over time and cross-nationally.
Given the dynamic of changing issues, it seems reasonable to expect them to have an effect on individuals’ left-right orientation but one can expect the opposite effect as well. Moreover, not all issues are equally important to different individuals and thus, the relationship between issues and left-right ideology might vary. This will be tested by using structural equation modelling with data from the first round of the European Social Survey (2002/03) from the Netherlands.
November 30, 2009 - 16h00
"The paradox of support for democracy in Venezuela"
by Paolo Moncagatta (RECSM - Universitat Pompeu Fabra)
New democracies have not showed the same patterns of unconditional and generalised support for the democratic regime established, advanced democracies have. It is thus of crucial importance to understand under which particularities are new democracies maintained and legitimized. For this purpose, it is necessary to understand what is influencing mass support for democracy in them. "Support for democracy" has been widely measured in surveys, as it is regarded by many as an indicator of democratic legitimacy. Several questions have aimed to measure citizens' support of their regimes, seeking in this way a deeper understanding of the quality of democracy.
Using data from the Latin American Public Opinion Project (LAPOP) 2008 round, I use structural equation modelling to assess the nature of the relationship between trust in the president and authoritative institutions and support for the democratic regime in Venezuela. Two distinct groups emerge from the data: those who show no trust at all towards president Chávez and institutions and those who have great trust towards them. The paradox is that both groups show remarkably high levels of support for the democratic regime. This suggests these different groups have alternative conceptions of what democracy is, and implies the need for the development of new approaches to the measurement of support for democracy.
November 23, 2009 - 16h00
"How to estimate measurement error in categorical survey questions"
by Daniel Oberski (RECSM - Universitat Pompeu Fabra)
The present paper suggests a statistical method, the latent class MTMM model, of estimating the quality of single questions while making fewer assumptions than have been made so far in such evaluations. The method is a combination of the multitrait-multimethod research design of Campbell and Fiske (1959), the basic response model for single questions of Saris and Andrews (1991), and the latent class factor model of Vermunt et al. (2004). The latent class MTMM model is thus not novel in itself, but combines an existing design, model, and method to improve the analysis of single questions in survey research.
A real experiment from the European Social Survey (ESS) is analyzed and the results are discussed at length, yielding valuable insights into the functioning of these questions.
November 16, 2009 - 16h00
"Changes in national identification"
by María José Hierro (RECSM - Universitat Pompeu Fabra)
Generally, individual national identification has been assumed to be very stable over time. Even if this may be truth for some contexts, there are some other contexts in which the existence of a national cleavage and the presence of competing ideas on the nation make possible the occurrence of individual changes. Changes in the individual national identification can also take place due to a migratory experience. Despite this, the study of changes in the individual national identification has been limited so far by the absence of panel data.
This paper addresses this issue focusing on a highly contested period of time during the negotiation of a new self-government law in Catalonia. The debate on the new law was an expression of the identity politics in Spain. Being against the reform was framed as "being against the interests of the Catalan People" or, more generally, "being anti-Catalan". Conversely, being in favor of the reform was considered to be evidence of the Catalan's lack of solidarity towards other territories in Spain and as an attempt to break up the Spanish nation. Bearing this context in mind, one would expect that individuals reacted to the political parties' national mobilization changing their national identification.
The paper tests this hypothesis using a three-wave panel study ("Panel de Desigualtats a Catalunya, 2001/02 - 2005") and contrast the effect that some socio-demographic characteristics and party identification had on the occurrence of changes. The analysis is performed for different groups (natives, immigrants, and second generations).
November 9, 2009 - 16h00
"Voter turnout, turnout inequality, and economic inequality"
by Aina Gallego (RECSM - Universitat Pompeu Fabra)
This paper examines if economic inequality depresses voter turnout and increases turnout inequality with contextual and survey data for 34 countries in the period 1996-2008. There is a negative relationship between voter turnout and income inequality. A structural equations model suggests that the causality runs from inequality to turnout rather than from turnout to inequality. In the analysis with individual and contextual data it is found that while the poor and the poorly educated vote on average less frequently in the countries examined, the gradient of the relationship between income or education, and voter participation is not intensified by income inequality. Thus while social inequality might depress voter participation it does not exacerbate turnout inequality.