Formal comparisons between claims and indicators. Operational language, not verdicts. Human verification is a separate step.
contradicts Version: confidence 0.55 pending review Zane Oliņa (2017): The competency approach will let pupils prepare better for life and work in a changing world.
PISA — mathematics scores (Latvia) (score): 2003=483 → 2022=483
Claim (Oliņa 2017): the competency approach will let pupils prepare better. Skola 2030 officially launched 2018. Indicator (PISA mathematics): 2018=496* (Latvia's historical peak), 2022=483 (-13, statistically significant decline per PISA's own classification). The promise predicted improvement; the measurement shows significant decline. Confounders: (1) COVID-19 pandemic 2020-2022 affected all OECD countries — Latvia's drop is similar to the OECD average, not unique. (2) Skola 2030 was implemented gradually, with full content rollout 2020-2023, so PISA 2022 measures only partial implementation. Hence relation=contradicts with low confidence (0.55) and mandatory human review.
Δ: (-13 abs)
contradicts Version: confidence 0.45 pending review Zane Oliņa (2017): The competency approach will let pupils prepare better for life and work in a changing world.
PISA — reading scores (Latvia) (score): 2000=458 → 2022=475
Claim (Oliņa 2017): the competency approach will improve preparation. Indicator (PISA reading): 2018=479, 2022=475 (-4). Smaller drop than in math, but still adverse vs the promised improvement. Broader context: Latvia's reading scores have been declining since 2003 (491 → 475 over 22 years) — a secular decline predating Skola 2030. Skola 2030 has not reversed this trend, but reversing reading was not the reform's first-cycle goal. Confidence low (0.45) precisely because of this secular-trend confounder.
Δ: (-4 abs)
ambiguous Version: confidence 0.40 pending review Zane Oliņa (2017): The competency approach will let pupils prepare better for life and work in a changing world.
PISA — science scores (Latvia) (score): 2003=490 → 2022=494
Claim (Oliņa 2017): the competency approach will improve preparation. Indicator (PISA science): 2018=494, 2022=494 (Δ=0). Flat indicator. Two legitimate interpretations: (a) 'Stability = success', since COVID drove all OECD down, but Latvia's science did not fall — Skola 2030 may have acted protectively. (b) 'Promise was improvement, measurement shows no change' — promise unfulfilled. linker.md says such methodology-incompatibility should be classified as ambiguous, not contradicts. Confidence 0.40 — two legitimate optics, awaiting human resolution (or preservation of both as parallel).
Δ: (0 abs)
partially confirms Version: confidence 0.50 pending review TV3 Latvia (2026-02-18): 'Skola 2030' has failed; the ministry hides behind funding-shortage excuses.
PISA — mathematics scores (Latvia) (score): 2003=483 → 2022=483
Claim (TV3 18.02.2026): 'Skola 2030 has failed'. Indicator (PISA mathematics): -13 points drop between 2018 and 2022, statistically significant. This loosely matches the 'failure' frame — math is declining. But 'partially_confirms', not 'confirms', because: (a) the TV3 statement is qualitative/rhetorical, the indicator quantitative — full alignment is methodologically not possible. (b) The drop is partially explained by COVID-19's global impact on PISA. (c) TV3 itself, in the later jauns.lv piece, had VK clarify that the audit did not evaluate Skola 2030 content (claim_vk_clarification) — which weakens the 'failure' premise. Confidence 0.50.
Δ: (-13 abs)
ambiguous Version: confidence 0.50 pending review TV3 Latvia (2026-02-18): 'Skola 2030' has failed; the ministry hides behind funding-shortage excuses.
Education expenditure, % of GDP (% of GDP): 2010=5 → 2023=4.7
The claim (TV3 18.02.2026) frames the MoE position as 'hiding behind funding shortage'. The indicator shows education expenditure as %GDP at 5.8% (2018) → 4.9% (2022) → 4.7% (2023) — a declining trend. This does not refute or confirm the claim, because the claim addresses qualitative rhetoric, not a specific numerical relationship. Hence 'ambiguous': the rhetoric and the trend co-exist, but the claim is not numerically formulated.
ambiguous Version: confidence 0.40 pending review United List (2022): The United List calls to stop annual reviews of the school network and switch to a long-term model with guaranteed state funding.
Education expenditure, % of GDP (% of GDP): 2010=5 → 2023=4.7
The claim (United List programme 2022) demands 'guaranteed state funding' for the school network without specifying a level. The indicator shows 5.8% (2018) → 4.9% (2022) %GDP — a declining trend around the 2022 elections. Ambiguous: the claim is a normative demand without a numerical threshold; the indicator shows a trend, but the claim's 'guaranteed' is not operationalised as a concrete value.