Experimental statistics
Overview
The growing demand for new information from numerous stakeholders requires the National Statistical Office to strengthen its information potential, including through the production of experimental statistics. Experimental statistics represent innovations in the statistical production process. They are obtained on the basis of new sources and methods, so as to respond to users' needs in a timely manner.
Experimental statistics are new, innovative statistics that are currently being developed, including those in the evaluation phase. The methods and data sources used in the development of experimental statistics may be changed during the process of improving the methodology based on experience gained, including as a result of feedback and suggestions from users. They are developed and published to inform users about the results obtained during the development of new, innovative statistics that are at the development and testing stage, as well as to involve users and other stakeholders in assessing their necessity and quality at an early stage of development. The ultimate goal is the successful development of new statistics so that they meet the requirements for official statistics in accordance with the European Statistics Code of Practice.
Development approaches
Testing experimental statistics gives producers the opportunity to better understand their quality, including aspects related to accuracy and reliability, as well as their value. Users play a central role in this process—without their involvement, producers will not have a full understanding of the usefulness of statistics.
The conclusion of the development of new statistics will be drawn by the producer based on one of these three results:
- The statistics are of sufficient quality and value to be used in a meaningful way and can therefore be published as official statistics.
- The statistics are of insufficient quality or do not meet user needs, and therefore their production will be discontinued.
- There is insufficient evidence to reach a definite conclusion, or further refinement is needed – manufacturers could continue further development and will clearly present the direction and scope of further testing.
When the NBS has sufficient evidence that the experimental statistics are relevant and useful, and meet the minimum quality requirements, they will be disseminated as official statistics, at the discretion of the institution, and will generally be released on a regular basis.
The label "experimental statistics" cannot be used in the long term for a specific statistical product. The list of experimental statistics will be reviewed annually to assess whether their production should continue or, on the contrary, be discontinued. If the development of certain experimental statistics proves successful, ensuring that the statistics meet quality standards, the label "experimental statistics" will be removed and the statistics will be disseminated as "official statistics," while describing their limitations, where appropriate.
How experimental statistics differ from official statistics?
In general, statistics published by the NBS are considered official statistics. However, experimental statistics are not considered official statistics until they have been thoroughly tested and meet the requirements for official statistics, in accordance with the provisions of Law No. 93/2017 on official statistics and the European Statistics Code of Practice.
It is possible that the methods used to compile experimental statistics have not yet been thoroughly tested, that a new data source is being tested, or that it covers only part of the reference population.
EXPERIMENTAL STATISTICS: | OFFICIAL STATISTICS: |
They are not yet included in the statistics dissemination calendar. They are developed taking into account user needs, as well as available methods and sources. | Are published in accordance with the dissemination calendar for official statistics |
New sources of administrative, private, or geospatial data are used, as well as new methods for estimating/developing statistical data. | Traditional statistical research and administrative sources are used |
Not included in the list of official statistics | Are part of the list of official statistics produced in accordance with the Statistical Work Program |
Not all experimental statistics obtained are comparable with those obtained internationally. | Are developed based on sound methodologies and, in most cases, are internationally comparable |
The methodologies are still in the process of development and testing. | Are produced based on approved methodologies. |
Why does the NBS produce and publish experimental statistics?
The purpose of experimental statistics is: to respond to user requests for new, timely indicators related to the development of society; to improve the quality of statistics, expand the available disaggregations, and streamline the production of statistics, including reducing the information burden on respondents. By using more and more new sources and methods, statistics can be created in less time, in areas where official statistics may have a lag of several months or even years.
The dissemination of experimental statistics, which are preliminary in nature and currently undergoing development and testing, allows for feedback from users and other interested partners before reaching their final form, thus providing a more relevant and reliable product. In the meantime, they offer users the opportunity to familiarize themselves with the new statistics currently being developed.
Since these statistics have not yet reached full maturity in terms of coverage, methodology, or international comparability, they are always marked with a distinct logo and accompanied by methodological notes with the relevant explanations.