The Methods and Research meetings will take place at 12:15-13.30 in room 5799 of the Geopolis building in the Mouline quarter. Also see here
Presentation on 04.10.2016
“Methodological challenges in international research with centenarians”
[Daniela Jopp, (UNIL)]
The number of very old individuals, including centenarians, has increased substantially over the past decade, and this trend is likely to continue, as every second child born after the year 2000 is expected to reach age 100. Yet, research with very old individuals involves various challenges, including usability of standard measures with individuals who may have somewhat limited cognitive capacity, or inclusion of individuals who lack any communicative skills. This presentation will highlight methodological issues related to centenarian research with respect to design (focus on centenarians, focus on proxy informants, use of both centenarian and proxy), assessment approach (mixed methods, adjustment of standard measures), and analysis (suitable control group), and will make proposals how to deal with these issues on the basis of the experiences collected in the context of an international network of centenarian studies established in 2012.
Presentation on 01.11.2016
“Prediction of the quality of survey questions using the program SQP: Background and applications”
[Melanie Revilla, University Pompeu Fabra, Barcelona]
On average, 50% of the observed variance in answers to survey questions is due to measurement error (Alwin, 2007). However, the size of the error can vary a lot depending on the exact formulation of the survey questions and the language. One approach to estimate the size of measurement errors is the Survey Quality Predictor (SQP). This software allows predicting the characteristics of survey questions by coding their formal and linguistic characteristics.
In this presentation, I will introduce the problem of measurement errors, before briefly presenting standard approaches to measure such errors. Then, I will present the background behind the SQP software, as well as its uses. Finally, I will illustrate how to correct for measurement errors and the effect on the substantive results.
Presentation on 06.12.2016
“A New Data Set for the Analysis of Migration and Integration in Switzerland”
[Ilka Steiner, University of Geneva]
To fully understand migratory flows, their determinants and their consequences researchers increasingly use methods based on life course – or longitudinal – approaches. The introduction of the social security number in 2010 in the Swiss administrative data and a new decree, which came into force in 2013, regulating data linkage for statistical purpose, allows today the development of longitudinal statistics based on population registers and official surveys. This presentation reviews the process leading to the creation of a new longitudinal data set for the analysis of migration and integration in Switzerland. We first describe its conceptual framework, by presenting the different available registers and the target population. Second, we discuss the data linkage and the validation procedures. Finally, we show an example of a possible application.
Presentation on 28.02.2017
“The design and practice of research with Qualitative Comparative Analysis (QCA)”
[Martino Maggetti, University of Lausanne]
Recent years have witnessed a host of innovations for Qualitative Comparative Analysis (QCA). This presentation (mostly based on current work by Eva Thomann and myself) aims at reviewing, systematizing and clarifying different approaches, challenges, and tools regarding the design of QCA-oriented research. In this regard, I will first introduce the nuts and bolts of QCA as an approach and as a technique. Afterwards, I will turn to the trade-offs that characterize different emerging approaches to QCA, as well as to the importance of doing justice to the nature and goals of QCA in each specific research context.
Presentation on 28.03.2017
“A data mining approach to longitudinal risk assessment, with application in cognitive epidemiology”
[Stephen Aichele, University of Geneva]
Research on psychological aging is easily confounded by problems related to participant attrition (e.g., due to dropout or death) and by complex interrelations among the variables investigated. Numerous statistical methods are available to address these issues (e.g., imputation to handle missing data, multi-level models for estimation of longitudinal change), but investigators in the behavioral sciences have only recently begun to discover the benefits of including machine learning techniques alongside more traditional approaches. One such technique is random forest analysis (RFA), which is especially effective for estimating the relative influence of numerous predictors in relation to an outcome of interest. The first part of this presentation will provide an overview of RFA as a valuable analytic tool. The second part of the presentation will demonstrate the use of RFA for predicting mortality risk and cerebral white matter lesion burden in a large sample (N = 6203; n = 112 with brain data) longitudinal study (> 20 years) of older adults.
Presentation on 25.04.2017
“Experimental designs for studying social norms”
[Heiko Rauhut, University of Zurich]