The voting archive includes votes from the Storting's electronic voting system and covers the period from and including the 1979/80 session. The period is available as an unbroken time series. The data files are organized so that the votes are units in the files. The voting results, all case and reference information, party codes and personal information are stored as variables. The database provides the opportunity to search for votes on various subject areas and periods.
What binds these tables together is the variable id_vot, which identifies the individual vote. Id_vot is based on the Storting's own numbering, but takes into account that this numbering has varied over the years. In addition, id_vot is sorting-friendly in the sense that the sorting order gives the correct order of the votes:
The voting archive contains all non-unanimous votes that have been taken using the Storting's electronic voting system. This means that all unanimous votes are excluded, and this is because they are not initially voted on using the system. Some votes where it is expected that only a small number of representatives will vote differently from the majority are also carried out without using the voting system. This primarily applies to personal proposals from small parties that are not expected to receive support. In the period 1977-85, both the Socialist Left Party and the Liberal Party were small parties in the Storting, and the Storting proceedings show that the votes on these parties' proposals were often carried out by the representatives standing up. The same was the case for the Progress Party in the period 1981-1989. In the period 1993-97, the same applied to RV, Venstre and the independent representatives. Studies of cooperation patterns in the Storting must therefore take into account that these parties in the specified periods will appear to agree more with the rest of the Storting than was actually the case. At the same time, the other parties will appear to be less in agreement with each other than was actually the situation.
The number of votes per session increased considerably in the 1990s compared to the early 1980s. While in the first half of the 1980s there were approximately 3-400 votes a year, in the 1990s there have been over 1,200 votes each year. Part of this increase can be explained by changes in voting procedures. While early in the period it was common to vote on a personal proposal alternatively with the setting, from the latter half of the 1980s it has become more common to vote on personal proposals and the setting separately. This transition from alternative to sequential voting creates a direct increase in the number of votes, but at the same time there has been a real increase in all types of voting. On 19 June 1997, the Storting adopted a budget reform as a trial scheme for the period 1997-2001. This has led to the number of votes being more than halved.
The period from the 1979/80 session up to and including the 1988/89 session is registered and organized by NSD on the basis of paper printouts from the voting system. We have also, based on the parliamentary proceedings, registered case information about the individual votes. Starting with the 1989/90 session, NSD has had data transferred from the Storting's electronic voting system at the end of each session. NSD prepares data for analysis purposes and supplements the material with additional case and reference information from the Stortingstidende. Subject codes will also be added in relation to the Storting's thesaurus. With the exception of the period 1981-1989 (session 126-133), the data also includes information on the representatives of the Storting who have voted at the polls.
The data quality is mostly good for the votes in the archive. Proposals from smaller parties in the Storting have, to varying degrees, been "hammered through" by the president against the votes of these smaller parties, without having gone through the electronic voting system. This means that these votes are missing from the archive and that distances between smaller and larger parties will be affected.
Voting data is most often used to construct agreement and disagreement indices. The starting point is to create a measure that can say something about the distance between pairs of parties in a simple way. To achieve this, we start from the representatives' votes in the Storting, we distribute the representatives' votes for and against the parties, and then calculate the parties' percentage of votes for and against. By comparing the votes for (or the votes against) for two parties, you will be able to say something about the (political) distance between these parties. If all the representatives of both parties have voted in favor in a vote, the percentage of votes in favor will be 100% for both parties and there will be no gap between the parties. We will get the same result - no disagreement - if no one in the two parties voted for it. The distance between two parties consequently reaches its maximum when all the representatives in one party vote for a proposal, while all the representatives in the other party vote against a proposal.
Mathematically, we calculate the distance - the index of disagreement - by taking the difference (absolute value) between the proportion of votes for between the pairs of parties being studied. The table below shows some examples.
| Voting | Parti A (%) | Parti B (%) | Disagreement index |
Agreement index |
|---|---|---|---|---|
| No. 1 | 100 | 100 | 0 | 100 |
| No. 2 | 0 | 100 | 100 | 0 |
| No. 3 | 100 | 0 | 100 | 0 |
| No. 4 | 60 | 40 | 20 | 80 |
All the representatives within one party usually vote either for or against. This means that most disagreement indices between pairs of parties will be either 0, no disagreement or 100 complete disagreement. As the degree of agreement is easier to perceive, we have reversed the index in the presentations by taking the difference between 100 and the disagreement index. Cf. the agreement index column in the table.
The agreement indices can then be used to create aggregated statistics to say something about party distances in general and not just for each individual vote. In the presentations, this is done by calculating the average distance divided by session, with possibly further subdivisions by subject, committee or type of case. The data files are organized so that the individual representative's (party in the period 1981-1989) vote in each case are units in the files. In particular, there are two different file types that are delivered to users: