This blog post was written by the NDI election team -Michael McNulty and Benjamin Mindes
How “open” are election results data around the world? Answering that question just became much easier. For the first time, the Global Open Data Index 2015 assessed election results data based on whether the results are made available at the polling station level. In previous years, the Index looked at whether election results were available at a higher (constituency/district) level, but not at the polling station level.
As a result, the 2015 Global Open Data Index provides the most useful global assessment to date on which countries are and are not making election results available in an open way. It also highlights specific open data principles that most countries are meeting, as well as principles that most countries are not meeting. This helps inform the reform agenda for open election data advocates in the months and years ahead.
Before we take a look at the findings and possible ways forward, let’s first consider why the Global Open Data Index’s shift from constituency/district level results to polling station results is important. This shift in criteria has shaken up the rankings this year, which has caused some discussion about why polling station-level results matter. Read on to find out!
Why are Polling Station-level Election Results Important?
Meets the open data principle of “granularity”
A commonly accepted open data principle is that data should be made available at the most granular, or “primary,” level — the level at which the source data is collected. (See the 8 Principles of Open Government Data principle on Primary; and the G8 Open Data Charter section on Quality and Quantity.) In the case of election results, the primary level refers to the location where voters cast their ballots — the polling station. (See the Open Election Data Initiative section on Granularity. Polling stations are sometimes called precincts, polling streams, or tables depending on the context) So, if election results are not counted at the polling station level and/or only made available in an aggregate form, such as only at the ward/constituency/district level, then that dataset is not truly open, since it does not meet the principle of granularity. (See the Open Election Data Initiative section on Election Results for more details.)
Promotes transparency and public confidence
Transparency means that each step is open to scrutiny and that there can be an independent verification the process. If results aren’t counted and made public at the polling station level, there is a clear lack of transparency, because there is no way to verify whether the higher-level tabulated results can be trusted. This makes election fraud easier to conceal and mistakes harder to catch, which can undermine public confidence in elections, distort the will of the voter, and, in a close election, even change the outcome.
For example, let’s imagine that a tabulation center is aggregating ballots from 10 polling stations. Independent observers at two of those polling stations reported several people voting multiple times, as well as officials stuffing ballot boxes. If polling station results were made available, observers could check whether the number of ballots cast exceeds the number of registered voters at those polling stations, which would support the observers’ findings of fraud. However, if polling station level results aren’t made available, the results from the two “problem” polling stations would be mixed in with the other eight polling stations. There would be no way to verify what the turnout was at the two problem polling stations, and, thus, no way to cross-check the observers’ findings with the official results.
Election observers can combine their assessment of the election day process with results data to verify or dispel rumors at specific polling stations, but only if polling station-level results are made public.
Aligns with an emerging global norm
Making results available at the polling station level is rapidly becoming a global norm. In most countries, political parties, election observers, the media, and voters have come to expect nothing less than for polling station level results to be posted publicly in a timely way and shared freely.
The 2015 Open Data Index shows how common this practice has become. Of the 122 countries studied, 71 of them (58%) provide election results (including results, registered voters, and number of invalid and spoiled ballots) at the polling station level. There are some significant differences across regions, however. Sub-Saharan Africa and Asia had the lowest percentage of countries that provide polling station level results data (42% and 41% respectively). Eastern Europe and Latin America have the highest percentage of countries with 71% each.
What Does the Index Tell Us about How to Open Up and Use Election Data?
Drawing on the 2015 Global Open Data Index findings and on open election data initiatives at the global, regional and national levels, we’ve highlighted some key priorities below.
1. Advocacy for making polling-station level results publicly available
While most countries make polling-station level results available, over 40% of the 112 countries researched in the Global Open Data Index still do not. At a regional level, Sub Saharan Africa, Asia and the Middle East & North Africa have the furthest to go.
2. Ensuring election results data is truly open
Making election data available is good first step, but it can’t be freely and easily used and redistributed by the public if it is not truly “open.” Election data is open when it is released in a manner that is granular, available for free online, complete and in bulk, analyzable (machine-readable), non-proprietary, non-discriminatory and available to anyone, license-free and permanent. Equally as important, election data must be released in a timely way. For election results, this means near real-time publication of provisional polling station results, with frequent updates.
The Global Open Data Index assesses many of these criteria, and the 2015 findings help highlight which criteria are more and less commonly met across the globe. On the positive side, of the 71 countries that make polling-station level results available, nearly all of them provide the data in a digital (90%), online (89%), public (89%) and free (87%) manner. In addition, 92% of those 71 countries have up-to-date data.
However, there are some significant shortcomings across most countries. Only 46% of the 71 countries provided data that was analyzable (machine readable). Similarly, only 46% of countries studied provided complete, bulk data sets. Western Europe (67%) had the highest percentage of countries providing complete, bulk data, while Middle East & North Africa and Sub Saharan Africa (both 38%) had the lowest percentage of countries doing so.
3. Not just election results: Making other types of election data open
While election results often get the most attention, election data goes far beyond results. It involves information relating to all aspects of the electoral cycle, including the legal framework, decisionmaking processes, electoral boundaries, polling stations, campaign finance, voter registry, ballot qualification, procurement, and complaints and disputes resolution. All of these categories of data are essential assessing the integrity of elections, and open data principles should be applied to all of them.
4. Moving from transparency to accountability
Opening election data helps make elections more transparent, but that’s just the beginning. To unlock the potential of election data, people need to have the knowledge and skills to analyze and use the data to promote greater inclusiveness and public engagement in the process, as well as to hold electoral actors, such as election management bodies and political parties, accountable. For example, with polling station data, citizen election observer groups around the world have used statistics to deploy observers to a random, representative sample of polling stations, giving them a comprehensive, accurate assessment of election day processes. With access to the voters list, many observer groups verify the accuracy of the list and highlight areas for improvement in the next elections.
Despite the increasing availability of election data, in most countries parties, the media and civil society do not yet have the capacity to take full advantage of the possibilities. The National Democratic Institute (NDI) is developing resources and tools to help equip electoral stakeholders, particularly citizen election observers, to use and analyze election data. We encourage more efforts like this so that the use of election data can reach its full potential.