Replication data for: When data lie — correcting a conflict variable reverses a significant finding in Chad's agricultural time series (1961–2024)

Variables: millet, sorghum and maize yields and areas (FAOSTAT, QCL domain); mean temperature, temperature anomaly, precipitation and SPI (CRU-TS 4.10 via the World Bank Climate Change Knowledge Portal); nitrogen fertiliser use (FAOSTAT, RFN domain); rural population (WDI); and two binary conflict variables: conflit_original (initial coding, 1965–1987 only) and conflit_corrige (corrected against the UCDP/PRIO Armed Conflict Dataset v26.1).

The two conflict variables differ in 32 of 64 years: 27 conflict years coded as peace (false zeros) and 5 years coded as conflict but not recorded by UCDP/PRIO (1965, 1973, 1974, 1975, 1985). Version 2 (September 2026): the replication do-file now reports all estimates from Stata 17 (newey, Student-t inference), uses identical specifications before and after the conflict correction (Table 4), adds robustness checks (linear trend, second autoregressive lag) and the year-by-year reconciliation of the conflict variable.