Marina Aleksandrova spoke at the conference of the European Sociological Research Association at the University of Milano-Bicocca
Marina Aleksandrova presented at the conference the results of her dissertation research on the topic: “The Application of Natural Language Processing Methods for prediction of item nonresponse: evidence from the ESS”.
The topic of the international conference of the European Sociological Research Association (ESRA), which was held in Milan from July 17 to 21, was related to the challenges, opportunities and new directions of survey research in times of crisis in University of Milano-Bicocca.
The conference program turned out to be very rich, which is why we often had to make difficult choices. Every day of the conference, it was possible to attend the speeches of researchers, which allowed them to learn about insufficiently familiar methods of data analysis, ways to assess the quality of measurement in sociological surveys, features of conducting a survey among representatives of socially vulnerable groups, to get acquainted with original methodological approaches to solving research problems, with new available data that are not actively used in research in Russia, although they may also be of interest to researchers. So, the following sections were visited:
- "Survey Nonresponse Trends and Trust in Surveys: A global perspective on the current survey climate".
- "Reducing and measuring nonresponse bias in times of crisis: Challenges, opportunities, and new directions".
- "Combining Data Science and Survey Research to Improve (Training) Data Quality".
- "Assessing the Quality of Survey Data".
- "Questionnaire translation and the global crises – new challenges and opportunities".
Marina Alexandrova's presentation took place at the section "Methods for questionnaire development: SQP, Cognitive Interviewing, and others". The session was led by Cornelia Neuert (PhD, GESIS - Leibniz Institute for the Social Sciences) and Lydia Repke (PhD, GESIS - Leibniz Institute for the Social Sciences).
The session was devoted to methods and procedures that can be used to assess the quality of survey tools in order to identify problems in the questionnaires being developed. This direction remains relevant, since such problems can lead to biases in the data obtained and, accordingly, unreliable and invalid results, in addition, there is still no consensus among researchers regarding the practices used to assess the quality of questionnaire questions.
Marina Alexandrova's report suggested one of the possible options for the development of existing methods for assessing the quality of questionnaire questions using natural language processing methods to solve the more specific problem of omissions in survey data arising from partial non-responses. The results presented in the report are based on a dissertation research conducted over several years. The report provoked a positive reaction among the researchers who visited the section, useful comments were received on possible options for the development of dissertation research and solutions to new research problems identified during the empirical stage.
In addition, we managed to take part in a strategic session "Strategic session on the future of survey practices: How has the pandemic affected survey practices - which are worth keeping in the long term?" dedicated to discussing the impact of the pandemic on survey research, as well as which of the developed practices, according to researchers, will remain with us after it, and which will not.
Acknowledgments section
The presentation of the report would have been impossible without the financial and organizational support of the Scientific Commission of the Faculty of Social Sciences, employees of the School of Sociology and our Laboratory.