Methodology

What this site measures

"Încotro?" pursues a single question: in which direction is Romania moving, in the domains that matter for people's lives. We do not build a single country score, because any such score would hide arbitrary weighting choices. Instead, we show every indicator with its history, its recent direction, and Romania's position in the EU where comparable data exists.

Where the data comes from

Each indicator has a single primary source, chosen on three criteria: authority (official statistical institutions or reference academic projects), international comparability, and a long historical series. Sources used: Eurostat (most indicators), the World Bank, WHO/UNICEF, the OECD (PISA), V-Dem, Transparency International, Reporters Without Borders, the UN (migrant stock) and International IDEA. Every indicator displays its exact source, dataset code, and the address it can be downloaded from.

Eurostat, World Bank and WHO data, along with the series that come through Our World in Data (V-Dem, PISA, life evaluation), was pulled directly from public programming interfaces (APIs), with no manual transcription. For the few series without such an interface (the Corruption Perceptions Index, voter turnout, Romanians abroad, press freedom), values were taken from the official files published by the source and checked by hand; the raw files and the verification log are published in the site's data repository.

How the trend label is computed

The labels ("improving", "stagnant", "declining"...) are computed mechanically, identically for all indicators, with no editorial intervention:

• We take the last at most 8 available observations. If the series has a methodology break, we use only the post-break segment.
• We estimate the slope with a simple linear regression over that window.
• We normalise the slope by the full series' historical range: what percentage of its historical range the indicator covers in one year.
• Below 0.8% per year: stagnant. Between 0.8% and 2.5%: slow (improvement or decline, by direction). Above 2.5%: clear.
• If the regression explains little (R² below 0.25) and residuals exceed 10% of the historical range, the label becomes volatile: we do not claim a direction that is not there.
• A "slow" trend the regression barely supports (R² below 0.15) is downgraded to stagnant: a slope without statistical backing is not a direction.
• Conversely, a very well-supported "slow" trend (R² of at least 0.6) moving faster than 2.5% per year of the indicator's current level is promoted to clear: normalising by the historical range can understate progress in series that fell enormously in the past (e.g. tuberculosis incidence).
• Fewer than 4 data points: short series. We say we cannot classify.

For every indicator we declare in advance what "good" means: up (e.g. incomes), down (e.g. mortality), or nothing (e.g. fertility, where we impose no judgement). The label on the site combines the statistical direction with that declaration. The thresholds above are editorial choices; we publish them precisely so they can be criticised.

EU position

Where the source covers all member states, we compute Romania's position among the 27, using the most recent year with data for at least 20 countries. Rank 1 always means "best" on the indicator's declared direction.

The limits we accept

• Perception is not reality. Corruption and democracy indicators measure expert and public assessments: the best instruments in existence, but imperfect ones.
• Methodology changes (the 2021 labour-force survey, 2022 e-government, 2022 RSF, 2024 WGI) are marked on charts and handled in the computation.