1# Verification checks (P4)23- Clean re-run: deleted out/results.json, re-ran analysis.py, byte-identical to previous output (cmp): PASS4- No CIs/SEs produced (no uncertainty propagation required), so CI-contains-estimate and SE>0 checks are not applicable.5- Information rule: every input is a printed cell of the cited extract (line numbers in locators); E4 baseline not substituted.6- n_harmonised equals count of non-null values: PASS7- n_not_harmonisable equals count harmonisable=false: PASS8- n_harmonised + n_not_harmonisable == n_analysed == 5: PASS9- null value <=> harmonisable false <=> missing_quantity non-null: PASS10- E1 computed rounds to printed -35.0%: PASS11- E1 sign negative like log-point ATT: PASS12- E2 recompute 100*1.47/75.0: PASS13- E3 recompute 100*-0.388*0.047: PASS14- E3 magnitude close to abstract 'about 2% employment decline': PASS15- E5 recompute -0.0962*100: PASS16- all non-null values finite: PASS17- random_seed null, task_id correct: PASS18- n_input_rows matches wc -l (66,91,58,134,16) and 5 items: PASS19- exclusions: 5 raw - 0 removed = 5 analysed: PASS20