Mindenkit szeretettel várunk az MTA TK "Lendület" RECENS  hálózati előadás-sorozatának következő alkalmára 2016. szeptember 13-án (kedden), melyen Dr. Rudas Tamás főigazgató úr tart előadást "Surveys, Statistics and Computational Social Science - Between Two Paradigms” címmel. 

Az előadás 15 órakor kezdődik, helyszíne szervezés alatt.

Az előadás kivonata:

Empirical social science undergoes rapid changes these days, to the point that some approaches are labeled now as computational social science. We are working with massive data sets, but the essence is not that the data are big, rather that they are process produced. The changes in the practice of social science research only mirror the societal changes which are taking place, so our happiness about finally being able to answer many of the long-standing questions of sociology, should be confined by our responsibility to find the questions which are relevant now and, perhaps, remain so in the near future. The presentation, unavoidably restricted in scope by the limited understanding which is characteristic of every transitional period, will discuss what remains useful from the tradition of statistical analysis of survey data. It will be argued that from the two main components of statistics, namely finding meaningful summaries of data, and inferring from a sample to a population, the first remains of major importance. So does one of the fundamental principles of statistical practice, stating that interpretation of the results should always be done in cooperation with the substantive scientist. The views expressed will go strongly against the widely held belief that the data can speak for themselves and also against the implied practice that any analysis which is technically possible, will provide meaningful and interpretable results.  Seeing the often missing substantive context of many data mining and machine learning activities, social scientists have become more sensitive to this issue, which is a very agreeable change of view after decades of linear regression being imposed on tens of thousands of data sets, which would have deserved better.