Methodology & background

Explanations

How are the overviews on OpenRCV produced, what do the terms and metrics mean, and where are the limits of the data? Here are the key points.

What is this?

OpenRCV makes the roll-call votes of the European Parliament searchable and traceable — with the voting record of every individual Member. All information comes straight from the Parliament's official Open Data Portal. No unofficial third-party sources, no estimates and no AI summaries are used.

What is a roll-call vote?

Parliament votes in several ways: by show of hands, electronically (totals only) or by roll call (RCV). Only a roll-call vote records how each individual Member voted.

OpenRCV only covers these roll-call votes. Show-of-hands votes have no individual records and cannot be shown here.

Where does the data come from?

The sole source is the European Parliament's EP Open Data Portal. Titles, topics (EuroVoc), committees and results are provided there officially and in many languages — so no machine translation is needed.

The data is available under CC BY 4.0. The prepared dataset can be downloaded freely (Data).

Main vote and sub-votes

A file often involves several roll-call votes: on individual amendments, paragraphs (§), recitals — and the final vote on the text as a whole. OpenRCV identifies the final vote (the "main vote") from the official reference text and highlights it; the result at the top of a detail page refers to it.

If a file has no roll-call final vote — for instance because the final vote was by show of hands and only an amendment was voted by roll call — no vote is marked as the main vote, and a note points this out.

Vote positions

  • In favour / Against / Abstention — the vote cast.
  • Excused — the Member was officially excused on that sitting day.
  • Did not vote — in mandate but did not take part in this vote (without an official excuse).

Vote corrections

After a vote, Members may put on record that they intended to vote differently (e.g. after a button error). OpenRCV shows such corrections transparently — the original vote stays visible, the intended one is marked.

Important: the official result is unaffected by corrections; they do not change the official vote counts. Recording these corrections is a distinguishing feature of this dataset.

The metrics explained

  • Turnout — share of votes during the mandate in which the Member took part.
  • Cohesion — how unified a group votes: on average, the share taking the same position (in favour, against or abstention).
  • On the winning side — share of clearly decided votes where the majority position (of a group or a Member) matches the chamber's result.
  • Dissent rate — share of votes in which a Member votes against the majority of their own group. The Member's own vote is excluded from the group majority.
  • Agreement / alliances — how often two groups take the same majority position.

Mandate basis & substitutes

Turnout and dissent rates only refer to votes within the respective mandate. Members who join mid-term receive no values for votes before their mandate began — keeping the comparison fair.

How current is it?

The site is rebuilt automatically every day; the time is shown in the footer ("Last updated"). New plenary weeks appear as soon as the data is available in the portal. Vote corrections are sometimes filed days later — they are included automatically.

Limits & notes

  • The primary final votes per file are covered; individual amendment/split votes are not listed separately in the aggregates.
  • Group and country assignment follow the current mandate status and may differ from the affiliation at the time of the vote.
  • OpenRCV is an independent project. The official result of the European Parliament always prevails; for details see Legal notice.