Research Project: Electoral Control in Eastern Europe

Report on Project Overview and Discussions from Workshop 2 and Specialized Course, “Parliamentarians, Candidates and Elections in Eastern Europe: From Theory to Empirics”

prepared by Joshua Kjerulf Dubrow, Associate Professor at IFiS-PAN and Principal Investigator of the Research Project

March 25, 2014

This event provides a forum for scholars interested in the research project, and for PhD students interested in the administration of research projects and in turning ideas into manageable research projects. We begin with an overview of the project.

Overview of the Project

This project, “Who Wins and Who Loses in the Parliamentary Elections? From Formal Theory to Empirical Analysis,” is funded by Poland’s National Science Centre (decision number 2012/05/E/HS6/03556) .Theory tells us that by conditioning their ballots on policy outcomes, voters can use elections to control politicians. Presumably, politicians anticipate that they will be sanctioned for poor party-performance, and thus have an incentive to implement policies, through their parties and other political units, that correspond to the preferences of the electorate. Does the system of repeated elections function as a mechanism of electoral control, and if it does, what factors influence its effectiveness? We consider this question in a broad context of studies on parliamentary elections.

The main activities of this research project are: (a) Data collection, preparation and distribution, (b) Data promotion: generate interest from PhD students to established scholars, (c) Train PhDs in data collection and administration, and (e) Academic publications — an edited book, as well as articles — on both substantive and methodological issues.

Review of Workshop 1

Workshop 1, “Winners and Losers in the Elections of Eastern Europe,” brought together PhD students and young and established scholars from sociology and political science from the U.S and Europe. The workshop introduced the data collection effort for the East European Parliamentarian and Candidate Database (EAST-PaC), a dataset on parliamentary candidates for all elections matched across time since 1989 in Poland, Ukraine and Hungary. We presented substantive and methodological issues of the EAST-PaC data, encouraged discussion of these topics, and facilitated the emergence of a network of scholars.

Overview of Data Collection, Preparation and Distribution

The main focus is on Poland, Hungary and Ukraine.

Data are of two types:

1. The universe of candidates — not only winners, but also election losers – who stood for elective office in the national legislature after the end of State Socialism (and, if possible, before), including their publicly available sociodemographics and their political biography such as political party, district, and vote totals.

2. Contextual data for each election year, including different electoral/party systems and outcomes, unemployment rates and other factors.

We call these data, the East European Parliamentarian and Candidate Data, or EAST PaC, for short. In the last Workshop scholars described the data collection effort for Poland and Ukraine. The Polish data is closest to completion; it has gone through a rigorous data cleaning procedure as described by Prof. Zbyszek Sawinski in the last Workshop, and needs a final look before we create the codebook and prepare it for archiving. The Polish contextual data for the most recent elections has already been collected. We will provide a description and release of these data soon. In short, these data are collected at the voivodship level (N = 41) for the years 2007 and 2011 and contains data on population, unemployment, crime, and social services, specifically medical, including hospital and doctor and nurse availability. We plan to collect more of these data that extends further back in time.

The Ukrainian candidate data is currently being cleaned. The Polish team are working with Dr. Natalia Pohorila to prepare a final version of these data. We are planning to collect contextual data at the oblast level. In a separate presentation, Dr. Zsofia Papp discussed the Hungarian candidate data collection effort (download the report, Candidate data collection in Hungary). We have contextual data for Hungarian administrative districts from 1991 to 2004. We plan to update these data.

Through this Workshop, and through the network of scholars, we can build a better understanding of the substantive issues and methodological challenges involved in electoral control in Eastern Europe.

Some suggested that voter turnout be a contextual data. It would be good to collect these data at the administrative district level for each election.

And elsewhere: In the course of this project, we have been in contact with researchers and students who are interested in candidate data outside of our target countries of Poland, Hungary and Ukraine. Currently, we are working with Mihail Chiru and Marina Popescu on the collection of Romanian candidate data that improves upon and updates the previous effort from the University of Essex that was initially led by Marina Popescu. They have sent a report to me on their efforts. Unfortunately, they could not attend this Workshop, so I will summarize their report::

They are attempting to collect candidate data from every election between 1990 and 2012. At this point, the data they have been able to collect varies by election. In short, the closer to 2012, the better the data. As they describe it:

The Romanian Permanent Electoral Authority (ROAEP) told the Romanian team to go to the county tribunals, which should have archived the candidate lists, according to ROAEP’s electoral law interpretation. In mid January they sent freedom of information (FOI) requests to all the 43 county tribunals in Romania asking for the electronic copies of the 1990-2004 electoral lists used in those counties, or for the possibility to copy them ourselves. As of March 17, 2014 they received 14 substantive answers, but the news are rather mixed. Some claim to have destroyed the lists following a law that apparently makes mandatory the destruction of voting materials three months after the elections. Others say that they have sent the documents to the National Archives. Finally, only two courts did send them the lists, while two others invited them to visit their archives to make copies. Of the two courts that sent them the lists one has made illegible much of the personal information (e.g. profession, year of birth). We will continue to work with the Romanian team, providing assistance as appropriate.

After I presented this report from the Romanian team, Dr. Claudiu Tufis offered interesting advice on the data collection effort for Romania. In his opinion, the destruction of such records seems an unfortunate decision. Perhaps information from previous elections can be collected from newspapers, party archives and other outlets that publish voting bulletins. Candidates may be required to report their role in the government prior to 1989, and these reports may be publicly available on-line.

There was great interest in the Czech candidate data. These data had been collected for prior elections, but over the years since we have not had the relevant contacts to follow-up. However, the project administration agreed to further examine how to make these data available. Project participants suggested that these data should be available in any form, and that interested scholars can update these data. The Electoral Control project can offer some data management, networking and methodological support.

We hope to extend into Asia. We have a PhD student from GSSR who is interested in candidate data from Nepal (Sanjay Chaudhary). In fact, he has already collected the foundations of it — the last two elections in Nepal — and we plan to help him turn it into a manageable dataset. Our project will likely play a limited role, but we hope that, by investigating the possibility, future funding can be secured or, at least, it provides information for the international scientific community to collect these data.

The Polish team, led by Prof. Sawinski and Dorota Laskowska from ORBS, here at PAN, has amassed considerable methodological expertise in collecting and cleaning candidate data. We plan to work with them to provide detailed descriptions of their methods and their software so that scholars in other countries can better collect and manage their own candidate data.

Harmonization, Data Files and Data Sharing

A key issue is the harmonization of data across datasets, countries, districts and time. There are many methodological challenges in analyzing these data, as well as in harmonization. At root, harmonization starts with a specific research interest that will guide the harmonization process, including what to harmonize. We take a bottom-up approach by providing researchers with the best possible data and the best possible information to conduct their research, which will in turn inform our project on how best to proceed. As such, this project will not provide a complete harmonized dataset. Instead, our intention is to provide researchers with as much information and support as possible to harmonize these data as they see fit. We are communicating with the data harmonization team, led by Prof. Kazimierz M. Slomczynski here at IFiS, whose efforts will provide suggestions on how, methodologically, to harmonize such data. The only harmonization we provide at this point is matching candidates through time.

At this point, we are creating separate data files for candidate data and contextual data. We are exploring the creation of a party data file, but that, too, will be kept as a separate file. The party file would have unique party identifiers that enable its harmonization with the candidate data that would have the same unique party IDs.

We also need to create a protocol for data sharing, including a proper citation. This would encourage users to cite the data and use the protocol.

Virtual Networking and Communication

Since this is an international project, there now arises a need for participants to communicate directly, but via channels established by the Electoral Control project such that all participants can benefit from the network and communications. The Project plans to hire PhD student who will assist with this problem.

The following are notes on conducting a cross-national project across different time zones and geographies. There are some general principles.

First, we want to use free cloud-based software.

Second, we want a simple, intuitive interface.

Third, the software must solve more problems than it creates.

Fourth, the software should improve our cross-national work process and increase productivity.

Fifth, the software should allow us to keep an enduring record of the research activity.

The direct outcome of using this software should be (a) a smoother workflow and (b) an excellent electronic record of the process that led us from our ideas to the finished product. A by-product could be that the software allows us to create a useful and enduring free website that not only has our private electronic record of the process, but also a more limited visible public version.

Google Drive is useful for many reasons: its a great cloud-based document editor that handles all of the document files we use, as well as Excel/spreadsheet files. But it is not sufficient. Google Sites would allow for an easier management of files and notes, mostly because it is designed to be a website, and Google Drive isn’t. Google Sites interface is not intuitive like WordPress. WordPress, however, is designed to make “pretty” websites, and is useless for keeping an automatic work process record. Google Sites keeps a nice record, but there is a learning curve. Perhaps there are other software solutions, such as a project wiki.

Proposed Research Projects

Participants presented to their Workshop and specialized course colleagues a preliminary plan for a research article featuring the analyses of EAST-PaC data. In this presentation, participants will outline the research questions they would like to address, the country or countries in the EAST-PaC data they would like to analyze, and a sketch of the methodological issues involved in addressing their research questions. The Workshop and specialized course organizers encourage participants to be creative in their approach to the theories and empirics — including the possibility of merging EAST-PaC with other data — and to be open minded with regard to suggestions from fellow Workshop and specialized course colleagues on the possible directions of the research project. The purpose is to generate ideas that can be turned into manageable and publishable research products. Each presenter had a maximum ten minutes to present their research, followed by ten minutes of discussion about it.

Other Discussion Items

— Need to provide comprehensive list of published articles that have used previous versions of EAST PaC data.

— A document on changes to the electoral systems in each country would be helpful. Currently, we have overviews of the electoral systems and election outcomes. This can be enhanced with discussions of changes to electoral laws over time.

— The next workshop (Workshop 3) should be presentations of papers that have analyzed these data. Papers would be submitted 10 days before the Workshop, and the Workshop would be devoted less on presentations and more on informed discussions and constructive criticisms of the papers. Sessions should be thematic.

— Publication outlets include a book and special issues of Electoral Studies or Party Politics or some other journals that would be interested in presenting papers of EAST PaC data. The book could present expanded methodological discussions of articles published in journals. Some suggested a book devoted to methodology.