Pre-registration
UK post-1945 Attlee reforms (NHS, nationalisation of coal/rail/steel, expanded public housing) delivered measurable improvements in life expectancy and child mortality without undermining subsequent 1950s-1960s growth.
Falsification criterion — what would disprove this
This hypothesis is considered falsified if:
PRIMARY (dispositive, three sub-tests, all must hold for SUPPORTED): (1) UK life-expectancy at birth gain 1948→1969 ≥ +3.0 years AND UK gain does NOT lag the 8-peer (DEU,FRA,NLD,BEL,ITA,SWE,DNK,NOR,CHE minus DEU where data missing) mean gain by more than 1.0 year. (2) UK infant-mortality proportional reduction 1950→1969 ≥ 40% AND does NOT lag the peer-mean reduction by more than 10 percentage points. (3) UK 1950s real GDP per capita growth (Maddison mpd2020 mean annual log-growth 1950-1959) ≥ +2.0%/yr. Verdict mapping: 3/3 → SUPPORTED. 2/3 → partial. ≤1/3 → refuted. INFORMATIVE (non-gating): UK SD-distance from peer mean for both health metrics; per-country peer growth comparison. METHOD_VALID: peer coverage ≥ 6/9 for life-expectancy, ≥ 5/8 for infant-mortality; UK series present at all anchor years.
formal test & threshold
test: uk_attlee_three_primary_health_growth_check_1948_1969 threshold: PRIMARY: (uk_le_gain >= 3.0y AND uk_le_lag <= 1.0y) AND (uk_imr_reduction >= 0.40 AND uk_imr_lag <= 0.10) AND (uk_1950s_mean_log_growth >= 0.020).
Method
- Template
descriptive- Sample
- 10 countries · 1945 – 1969
- Evidence type
- associational
UK 1945-1969 single-country time-series compared against a 9-country continental peer set (DEU, FRA, NLD, BEL, ITA, SWE, DNK, NOR, CHE) on life-expectancy gain, infant-mortality reduction, and 1950s real-GDP-per-capita growth. Anchor years 1948 (post-war stabilization, coincides with NHS launch) and 1950 (broad IMR peer coverage). Not causal — postwar global growth tailwinds and welfare-state convergence confound clean attribution to the Attlee bundle. Peer-relative comparison partially controls for global tailwinds; it cannot control for selection on prior trend because UK and most peers also expanded social-insurance / health-service coverage in the same window (NL Sickness Benefit Act 1947, French Sécurité Sociale 1945, etc.).
Data
| Variable | Source | Transform |
|---|---|---|
life_expectancy_at_birth outcome | owid:life-expectancytier 2 | level |
infant_mortality outcome | owid:infant-mortalitytier 2 | level |
real_gdp_per_capita outcome | maddison:mpd2020tier 3 | log |
attlee_reform_indicator treatment | constructed:indicator = 1 for years >= 1945 (Attlee government inauguration); also marks NHS launch 1948 and major nationalisations tier 5 | indicator |
log_population control | world_bank_wdi:SP.POP.TOTLtier 2 | log |
● ready · ● pending · ● reconstruct-needed
Detailed result card
UK Attlee-era reforms: health outcomes & growth path (1945–1969)
Verdict: refuted — Only 0 of 3 primaries hold. Failed: life-expectancy, infant-mortality, 1950s growth. UK 1950s growth +1.69%/yr; LE gain +3.29y (peer-mean +4.45y); IMR cut 41.4% (peer-mean 53.5%).
Summary
- Life expectancy at birth, UK 1948→1969: 68.37 → 71.66 years (gain +3.29y). Peer-mean gain across 8 continental peers: +4.45y (SD 1.80y). UK lag vs peer-mean: +1.16y; UK SD-distance from peer mean: -0.64.
- PRIMARY 1 (FAIL): UK gain ≥ 3.0y AND UK lag ≤ 1.0y.
- Infant mortality, UK 1950→1969: 3.15 → 1.84 (OWID per-100 units), a 41.4% reduction. Peer-mean reduction (8 peers, DEU dropped — coverage starts 1968): 53.5%. UK lag: +12.1pp; UK SD-distance: -1.73.
- PRIMARY 2 (FAIL): UK reduction ≥ 40% AND UK lag ≤ 10pp.
- 1950s real GDP per capita growth (Maddison 1950→1959 mean YoY log-growth), UK: +1.69%/yr. Peer-mean: +3.74%/yr (range +2.41% to +7.83%).
- PRIMARY 3 (FAIL): UK ≥ 2.0%/yr.
Method
Three pre-registered primary statistics, all dispositive:
- UK life-expectancy gain 1948→1969 vs the 9-country continental peer mean (DEU,FRA,NLD,BEL,ITA,SWE,DNK,NOR,CHE). Anchor 1948 instead of 1945 to avoid wartime-mortality asymmetry (NLD 1944 famine, FRA German occupation, ITA front, etc.) — using 1945 would mechanically advantage UK vs continental peers.
- UK infant-mortality proportional reduction 1950→1969 vs peer-mean reduction. DEU is dropped from this peer set (OWID infant-mortality coverage starts 1968 only). Anchor 1950 rather than 1949 broadens peer coverage from 3 to 8. The OWID series appears to be expressed per-100 live births rather than the more common per-1000 — this does not affect proportional-reduction comparisons.
- UK Maddison real GDP per capita YoY log-growth, mean over 1950-1959, against the spec's stated 2%/yr threshold.
Verdict logic: 3/3 → SUPPORTED, 2/3 → partial, ≤1/3 → refuted. INFORMATIVE (non-gating): UK SD-distance vs peer-mean for both health metrics; peer-country growth comparison.
Important caveat (in spec disclosure): none of these primary tests are causal. Postwar global growth tailwinds and welfare-state convergence confound clean attribution to the Attlee bundle (NHS, nationalisations, public housing, National Insurance). Even if all three primaries pass, this evidence is consistent with the democratic-socialist claim, not proof of it.
Data
- owid:life-expectancy (1543–2023, broad coverage)
- owid:infant-mortality (1949+ for most peers; DEU 1968+)
- maddison:mpd2020 (gdppc, real GDP per capita)
- world_bank_wdi:SP.POP.TOTL (population control, manifest only)
Notes
Stub seeded from a democratic-socialist school prediction about Attlee-era reforms. The 1945-1969 growth path is hard to attribute to any single reform; needs human review of counterfactual specification.