Due to maintenance work, ssb.no will be unavailable between 5pm and 7pm on Saturday 15 December. We apologise for any inconvenience this may cause.
Statistical methods and standards
Research in this area seeks to improve methods and standards of data collection and the production and communication of statistics with a view to enhancing statistical reliability. The development of general statistical theory, where particular attention is paid to sample surveys, remains an ongoing activity.
Theory and methods for dealing with non-response and imputation seek to reducenon-response and increase representativeness of observed data, reduce bias and uncertainty of estimates, and ease the production of statistics and publicly available databases for scientists and policy makers.
Sampling design addresses sampling errors, data editing non-sampling errors. Sampling errors need to be minimized given user requirements, response burden and cost. Data editing looks at methods to identify error sources and identify and correct major errors to improve the quality of statistical products.
Small area statistics typically use indirect data to reduce uncertainty stemming from a lack of direct data. All available information is applied across population domains and/or time. Small area statistics are increasingly used to aid decision-making in the public and private sector, in resource distribution and regional planning.
Research in this area aims to ascertain how information on statistical units may be disclosed through publication or micro data transmission. Methods to prevent disclosure are under development.
One area of research deals with theoretical considerations of various approaches to the analysis of sample surveys. Another main research activity is concerned with developing estimation methods and methods for variance estimation adapted to surveys conducted at Statistics Norway.
Research in this area deals mainly with foundational issues and statistical methods for prediction, problems, with applications to the theory on sample surveys. Likelihood theory is the underlying approach to this research.
Research on time series models is used in for making prognoses, usually for economic time series, but also for population forecasts. Seasonal adjustment removes calendar and seasonal variation effects and eases detection of changes in the economic time series. Research seeks to align methods to Norwegian conditions, including, i.e., correcting for moveable holidays.
News about statistical methods and standards
Theory and methods in survey sampling
The next round of census will be completely register-based in all the Nordic countries. Household is a key statistical unit in this context. In a Discussion Paper Li-Chun Zhang outline a unit-error theory which provides a framework for evaluating the statistical accuracy of these register-based household statistics, and illustrate its use through an application to the Norwegian register household data.