The V-Dem Project

Varieties of Democracy

The Varieties of Democracy (V-Dem) Research Project takes a comprehensive approach to understanding democratization. This approach encompasses multiple core principles: electoral, liberal, majoritarian, consensual, participatory, deliberative, and egalitarian. Each Principle is represented by a separate index, and each is regarded as a separate outcome in the proposed study. In this manner we reconceptualize democracy from a single outcome to a set of outcomes.

In addition, we break down each core principle into its constituent components, each to be measured separately. Components include features such as free and fair elections, civil liberties, judicial independence, executive constraints, gender equality, media freedom, and civil society. Finally, each component is disaggregated into specific indicators.

This fundamentally different approach to democratization is made possible by the V-Dem Database, which measures 600+ indicators annually from 1789 to the present for all countries of the world.

The V-Dem approach stands out, first, as a large global collaboration among scholars with diverse areas of expertise; second, as the first project attempting to explain different varieties of democracy; and third, thanks to the highly disaggregated V-Dem data, the first project to explore causal mechanisms linking different aspects of democracy together.

With five Principal Investigators, 23 Project Managers with special responsibility for issue areas covered in the V-Dem dataset, 33 Regional Managers, 134 Country Coordinators and almost 4000 Country Experts, the V-Dem project is one of the world’s largest social science data collection projects on democracy.

The V-Dem Institute

The Headquarters is based at the V-Dem Institute at the Department of Political Science, University of Gothenburg, Sweden. In the early years of V-Dem, the Kellogg Institute for International Studies at the University of Notre Dame, one of the two founding institutions for the project, played an instrumental part in building and establishing V-Dem. It was initially responsible for data collection in the Western Hemisphere, hosted workshops, and funded many students working on the project, as well as one of the Project Coordinators. In addition, the Center for Research Computing at Notre Dame developed the research database and the web interfaces that were used from 2011 to fall 2014.

As the project grew, the V-Dem Institute in Gothenburg progressively assumed responsibility for these functions and became, in effect, the headquarters for the project. Recognizing the shifting roles, in 2018 the Kellogg Institute formalized its current role as the V-Dem Regional Center in North America, which supports research projects using V-Dem data and hosts speakers and occasional conferences and workshops.

V-Dem Methodology

V-Dem has developed innovative methods for aggregating expert judgments in a way that produces valid and reliable estimates of difficult-to-observe concepts. This aspect of the project is critical because many key features of democracy are not directly observable. We continually review our methodology—and occasionally adjust it—with the goal of improving the quality of V-Dem indicators and indices.

Country Experts

V-Dem has a global network of more than 3,700 Country Experts supporting the data collection process by providing expert information via our online surveys. This network helps V-Dem obtain detailed, local knowledge from qualified experts familiar with political developments in a given country. This unique approach allows V-Dem to detect subtle changes to the institutional environments of countries from one year to the next.

Country Expert contributions enable nuanced analysis of aspects of democracy by researchers, policy makers and everyone in the world interested in democracy and political systems in their full complexity.

Experts are usually academics or professionals with specialist and evidenced knowledge in one or more domains. Approximately two-thirds are nationals or residents of the country they provide information on.

We endeavor to have a minimum of five experts for each indicator per country. This typically means we have twenty-five or more experts per country, since each expert only codes indicators in his/her areas of expertise. The quality and impartiality of the data is highly dependent on the Country Experts.

Consequently, we pay a great deal of attention to their recruitment and use the following selection criteria:

• Expertise – Validated expert knowledge in the country and specific areas to be coded
• Local, In-Depth Knowledge - By design, two-thirds of Country Experts providing data on a country should be nationals or permanent residents of that country
• Seriousness of purpose - Willingness to devote time to the project and to deliberate carefully over the questions in the survey
• Impartiality - V-Dem aims to recruit Country Experts who will answer survey questions in an impartial manner
• Diversity - In professional background among the Country Experts chosen for a particular country.

We do not reveal the identity of our Country Experts and preserve Country Expert confidentiality according to a strict policy. All personal identifying information is stored separately and securely. The strict policy does not apply to Historical Country Experts who contribute data on the pre-1900 period. Historical Country Experts are given the option to be publicly acknowledged as an expert for their country or to remain anonymous.

Expert-coded Data

Expert-coded data raise concerns regarding comparability across time and space. Rating complex concepts requires judgment, which may vary across experts and cases. Moreover, because even equally knowledgeable experts may disagree, it is imperative to report measurement error to the user. We address these issues using both cutting-edge theory and methods, resulting in valid estimates of concepts relating to democracy.

A Bayesian Item-Response Theory (IRT) Estimation Strategy

Pemstein et al. (2018) have developed a Bayesian Item-Response Theory (IRT) estimation strategy that accounts for many concerns regarding problems with the expert-coded data, while also providing estimates of remaining random measurement error. We use this strategy to convert the ordinal responses experts provide into continuous estimates of the concepts being measured. The basic logic behind these models is that an unobserved latent trait exists, but we are only able to see imperfect manifestations of this trait. By taking all of these manifest items (in our case, expert ratings) together, we are able to provide an estimate of the trait. In the dataset, we present the user with a best estimate of the value for an observation (the point estimate), as well as an estimate of uncertainty (the credible regions, a Bayesian corollary of confidence intervals).

The IRT models we use allow for the possibility that experts have different thresholds for their ratings. These thresholds are estimated based on patterns in the data, and then incorporated into the final latent estimate. In this way, we are able to correct for the previously-discussed concern that one expert’s “some-what” may be another expert’s “weakly” (a concept known as Differential Item Functioning). Apart from experts holding different thresholds for each category, we also allow for their reliability (in IRT terminology, their “discrimination parameter”) to idiosyncratically vary in the IRT models, based on the degree to which they agree with other experts. Experts with higher reliability have a greater influence on concept estimation, accounting for the concern that not all experts are equally expert on all concepts and cases.