Highest level of education completed (STATPOP)

| Last update: 21.02.2025

Background and goals

The population’s highest level of education completed serves as crucial indicator for measuring a nation's human capital in terms of qualifications. In the context of demographic shifts such as an ageing population, accurate analysis and monitoring of trends in educational attainment are essential if countries are to remain competitive and meet the increasing demand for skilled labour.

To conduct multi-dimensional analyses, it is essential to have data on the educational level of each member of the permanent resident population. To compile such data, a process needs to be put in place that combines existing data from federal administrative statistics and surveys and uses an imputation model to complete the missing datapoints.

Procedure

This project uses a wide range of data sources each year to determine the level of education of the Swiss permanent resident population: the Population and Household Statistics (STATPOP), the Structural Surveys (RS) since 2010, the Longitudinal Analyses in the Field of Education (LABB) in addition to the School Staff Statistics (SSP) and the Statistics on the Staff of Higher Education Institutions (SHIS-PERS), the Social Assistance Statistics (SAS), the State Secretariat for Economic Affairs (SECO) unemployment statistics, the Swiss Earnings Structure Survey (ESS) and the Swiss National Cohort (SNC) based on the 1990 and 2000 censuses.

Deterministic imputation

First, the data from each source undergoes a validation process. This involves deleting data that is potentially obsolete, for example if the person has returned to education or if the person left Switzerland at an age when they could have returned to education, or if the person claims to have a tertiary level of education when they are under the age of 18. In addition, errors in the source data are corrected using LABB data. Once the data sources have been validated, they undergo a prioritisation process.

Model-based imputation

Structural Survey (RS) data is used to impute the level of education of individuals who have not yet been attributed a level. Our training set is made up of all the people in the Structural Survey who were not attributed a level of education from other data sources. Variables from various registers are used for imputation, including variables from the Central Compensation Office (e.g. salaries), Structural Business Statistics (e.g. NOGA), STATPOP population stock (e.g. gender, nationality, age), STATPOP household data (e.g. household composition), the Buildings and Dwellings Statistics (e.g. apartment blocks/single-family homes), etc.

The highest level of education variable thus obtained is 90% derived from the various data sources and around 10% from imputations based on a model adjusted using the data from the Structural Survey. This variable will be calculated annually, and the methodology for its construction may still evolve.

The processing of the source data and the model's imputation procedures are presented in a technical description available below.