Storting election, election survey

Name: Storting election, election survey
Topic: Elections

Division for Population Statistics

County

Every fourth year.

Data on micro level, information about sample units, population, and catalogues are stored for later (panel). The data file is transmitted to The Norwegian Social Science Data Services (NSD).

The first election survey conducted by Statistics Norway was in 1969. From 1977 the survey and fieldwork for the Norwegian electoral research programme is a collaboration between Statistics Norway and the Institute for Social Research.

Researchers and students in a wide range of social sciences use the statistics and analyses. Examples are political science, sociology, media and communication. Journalists, the political parties and public management are other important users. Journalists, researchers and student from abroad are also users of the statistics.

No external users have access to the statistics and analyses before they are published and accessible simultaneously for all users on ssb.no at 08 am. Prior to this, a minimum of three months' advance notice is given inthe Statistics Release Calendar. This is one of Statistics Norway’s key principles for ensuring that all users are treated equally.

Access to figures before data is released

Institute for Social Research is granted access to datafiles, figures, publications and reports before data is released by Statistics Norway. Statistics Norway and Institute for Social Research publish figures and results at the same time.

Other statistics about the general election.

Voluntarily

Act on Official Statistics and Statistics Norway §§ 10 and 20.

The Norwegian electorate.

Statistics Norway receives the electoral register from the Norwegian Electoral Directorate before the election. Statistics Norway merges this information with oter registers using a unique identification key for each person in the register. The entire electoral register is the basis for drawing the sample.

From 1969-2013 the sample is drawn according to the current standardised sampling frame used for personal interviewing in Statistics Norway. To use a two stage or a multistage sampling frame gives substantial cost reduction in an interviewer administrated face to face survey. Although details have been changed over the years due to demographic changes and administrative changes in the fieldwork organisation, the sampling frame has in its essence stayed the same. There is a two stage sampling frame where sparsely populated municipalities in the same area with the same (næringsgrunnlag) are collapsed into one psu in stage one, and one of the municipalities is the actual municipality where the sample is drawn. Municipalities with a population over a fixed numbers is not collapsed but is their own psu. Then in stage two a selfweighting sample is drawn. From 1977 the sample is divided into two. (1) About half the sample was part of the sample in the previous round. (2) the other part is new in each round and will be part of the next round. Note that this refer to the actual sampled drawn regardless of response or non response. There is a age cut of at 79 years, people that are over 80 years old are not included in the sample. In order to give the sample cross sectional proportions a proportional “new” sample of young people and immigrants are added to the panel from the previous round.

If a person “move” out from the municipality where he or she was originally sampled the interviewers are instructed to “follow” the person and do the interview in the new municipality. In 1969 there was some examples of substitution du to migration (170). From 1973 substitution is not used. People who migrate to other countries or lives abroad in the fieldwork period is classified as ineligibles. The same is the case for people who dies during the fieldwork period or is a resident of a nursing home.

From 1969 – 1977 the lower age limit for the right to vote was 20 years. From 1981 the lower age limit was set to 18 years.

From 2017 the data collection design is changed to use primely telephone and web. fieldwork design. There is no great cost reduction in using a two stage sampling plan any more, so the new sample is drawn as a simple random sample.

Suppression with less than 25 observations.

Comparable to previous general election surveys

Errors can occur in any survey, irrespective of how comprehensive it is. Errors may arise in connection with collection as well as during processing. Errors during collection, measurements errors, may arise when a respondent gives the wrong answer or when an interviewer register the wrong answer or does not fill in enough data on the computer.

Non-response may lead to sample bias. If the percentage of non-respondents in different groups deviates to a considerable degree, the net sample (the sample of persons interviewed) will only approximately have the same statistical proportion as the gross sample (the sample selected for interviewing)

In principle, statements about bias because of non-response must be linked to variables. If one has found that there is no bias because of non-response in a certain variable, this does not imply that there is no effect of non-response on other variables, and vice versa.

The uncertainty of the findings based on only a part of the population, is often called sample variance. Standard deviation is a measure of this uncertainty. The size of standard deviation depends, among other factors, on the number of observations in the sample, and on the distribution of the current variable in the whole population. The estimate of standard deviation can be computed by using the observation in the sample.