{"args":{},"checks":[{"abs_tol":null,"bind":"any","column":null,"failure_class":null,"field":"value","interval":false,"key":null,"kind":"answer_backed","min_lre":null,"note":"every number traced to a tool call or the question","open_interval":false,"rel_tol":null,"required":true,"source_id":null,"table":null,"tool":null,"value":null,"weight":1.0,"where":null},{"abs_tol":null,"bind":"answer_call","column":null,"failure_class":null,"field":"value","interval":true,"key":null,"kind":"answer_figure_close","min_lre":null,"note":"CRB estimate_within_solution_ci: the answer figure inside the gold 95% interval effect \u00b1 1.96\u00b7|SE| of source crb_papers.q047.gold (closed, evaluate_result.py; no --auto-rescale)","open_interval":false,"rel_tol":null,"required":true,"source_id":null,"table":null,"tool":null,"value":[-4.525834998078203,23.042285244253353],"weight":1.0,"where":null},{"abs_tol":7.0326837352886625e-06,"bind":"answer_call","column":null,"failure_class":null,"field":"value","interval":false,"key":null,"kind":"answer_figure_close","min_lre":null,"note":"replication cell: the gold effect to rel 1e-6 (abs floor 1e-6\u00b7SE); never required","open_interval":false,"rel_tol":1e-06,"required":false,"source_id":"crb_papers.q047.gold","table":null,"tool":null,"value":9.258225123087575,"weight":1.0,"where":null},{"abs_tol":7.0326837352886625e-06,"bind":"answer_call","column":null,"failure_class":null,"field":"se","interval":false,"key":null,"kind":"answer_figure_close","min_lre":null,"note":"replication cell: the gold SE to rel 1e-6; never required","open_interval":false,"rel_tol":1e-06,"required":false,"source_id":"crb_papers.q047.gold","table":null,"tool":null,"value":7.032683735288662,"weight":1.0,"where":null}],"datasets":{"ds_0025":{"columns":null,"description":"| Variable Name | Description |\n| --- | --- |\n| year | Gregorian calendar year of observation. One record per country-year. Coverage varies by country but spans much of 1828\u20132015. |\n| Countryid | Numeric identifier for countries in the panel. Serves as a stable code to group observations by country across years. |\n| primratio | Annual primary-school enrollment rate, expressed as a percentage of the resident population ages 5\u201314 enrolled in primary school (public and, where sources report it, private). Series are compiled from historical administrative counts and harmonized to a common 5\u201314 denominator; values are continuous percentages (0\u2013100). |\n| treatment | Civil-war exposure indicator at the country-year level. Coded 1 from the first year of the earliest intrastate war that directly involved the state onward, and 0 otherwise; war timing is drawn from the Correlates of War intrastate war records. |\n\nDataset Context:\nThis is an annual country\u2013year panel that brings together long-run measures of primary-school enrollment coverage and the timing of internal armed conflict for European and Latin American countries. The schooling series were constructed from historical administrative sources, anchored in International Historical Statistics and cross-checked against widely used decennial benchmarks, then extended backward with national statistical yearbooks, education reports compiled by the U.S. Bureau of Education (1872\u20131915), and regional data handbooks. The resulting enrollment measure reflects the share of children ages 5\u201314 enrolled in primary school each year and is available far earlier than most modern datasets, enabling analysis across the nineteenth, twentieth, and early twenty-first centuries.\n\nCivil conflict timing is drawn from the Correlates of War intrastate war dataset, focusing on episodes where the central state was a combatant. For each country, the panel flags the first qualifying episode and then maintains an \u201cever-exposed since onset\u201d indicator thereafter, aligning conflict timing with the annual schooling series.\n\nBecause the coverage begins as early as the 1820s for some countries and extends to 2015 for others, the panel spans periods of early state involvement in schooling, subsequent expansion and standardization of primary education, and the modern era of administrative reporting. Start years and continuity differ by country, reflecting when governments first recorded enrollment statistics and when national primary education systems took form. The dataset is suitable for historical description of how primary education coverage evolved over time in relation to episodes of internal conflict documented in comparative records.","file_sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","format":"csv","frame_sha256":null,"generator":null,"kind":"file","licence":"redistributed by CausalReasoningBenchmark (MIT) under the original replication data's terms","member":null,"object":null,"params":{},"path":null,"read_kw":{},"records":null,"role":"agent","sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","transforms":[],"url":"https://huggingface.co/datasets/syrgkanislab/CausalReasoningBenchmark/resolve/edd0fe1c1d1c5c0a4f0e3e67a50481ac679fb888/data/ds_0025/data/data.csv"},"ref_adopt":{"columns":null,"description":"Each country's first year with treatment == 1, among complete rows","file_sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","format":"csv","frame_sha256":null,"generator":null,"kind":"file","licence":"redistributed by CausalReasoningBenchmark (MIT) under the original replication data's terms","member":null,"object":null,"params":{},"path":null,"read_kw":{},"records":null,"role":"reference","sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","transforms":[{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"select"},{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"dropna"},{"args":{"query":"treatment == 1"},"op":"filter"},{"args":{"aggs":[{"column":"year","fn":"min","name":"adopt"}],"by":["Countryid"]},"op":"collapse"}],"url":"https://huggingface.co/datasets/syrgkanislab/CausalReasoningBenchmark/resolve/edd0fe1c1d1c5c0a4f0e3e67a50481ac679fb888/data/ds_0025/data/data.csv"},"ref_panel":{"columns":null,"description":"Country rows of 1848 and 1859: the 1849 cohort and the countries untreated in both years","file_sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","format":"csv","frame_sha256":null,"generator":null,"kind":"file","licence":"redistributed by CausalReasoningBenchmark (MIT) under the original replication data's terms","member":null,"object":null,"params":{},"path":null,"read_kw":{},"records":null,"role":"reference","sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","transforms":[{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"select"},{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"dropna"},{"args":{"query":"year in [1848, 1859]"},"op":"filter"},{"args":{"how":"left","on":["Countryid"],"with":"ref_years"},"op":"merge"},{"args":{"how":"left","on":["Countryid"],"with":"ref_adopt"},"op":"merge"},{"args":{"how":"left","on":["Countryid"],"with":"ref_pre"},"op":"merge"},{"args":{"how":"left","on":["Countryid"],"with":"ref_post"},"op":"merge"},{"args":{"query":"nyears == 2 and t_pre == 0 and ((adopt == 1849 and t_post == 1) or t_post == 0)"},"op":"filter"},{"args":{"args":{"cmp":"eq","column":"t_post","value":1},"name":"TreatCohort","op":"indicator"},"op":"derive"},{"args":{"args":{"cmp":"eq","column":"year","value":1859},"name":"Post","op":"indicator"},"op":"derive"},{"args":{"args":{"columns":["TreatCohort","Post"]},"name":"Interaction","op":"interaction"},"op":"derive"}],"url":"https://huggingface.co/datasets/syrgkanislab/CausalReasoningBenchmark/resolve/edd0fe1c1d1c5c0a4f0e3e67a50481ac679fb888/data/ds_0025/data/data.csv"},"ref_post":{"columns":null,"description":"Each country's treatment in 1859 (t_post)","file_sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","format":"csv","frame_sha256":null,"generator":null,"kind":"file","licence":"redistributed by CausalReasoningBenchmark (MIT) under the original replication data's terms","member":null,"object":null,"params":{},"path":null,"read_kw":{},"records":null,"role":"reference","sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","transforms":[{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"select"},{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"dropna"},{"args":{"query":"year == 1859"},"op":"filter"},{"args":{"columns":["Countryid","treatment"]},"op":"select"},{"args":{"map":{"treatment":"t_post"}},"op":"rename"}],"url":"https://huggingface.co/datasets/syrgkanislab/CausalReasoningBenchmark/resolve/edd0fe1c1d1c5c0a4f0e3e67a50481ac679fb888/data/ds_0025/data/data.csv"},"ref_pre":{"columns":null,"description":"Each country's treatment in 1848 (t_pre)","file_sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","format":"csv","frame_sha256":null,"generator":null,"kind":"file","licence":"redistributed by CausalReasoningBenchmark (MIT) under the original replication data's terms","member":null,"object":null,"params":{},"path":null,"read_kw":{},"records":null,"role":"reference","sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","transforms":[{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"select"},{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"dropna"},{"args":{"query":"year == 1848"},"op":"filter"},{"args":{"columns":["Countryid","treatment"]},"op":"select"},{"args":{"map":{"treatment":"t_pre"}},"op":"rename"}],"url":"https://huggingface.co/datasets/syrgkanislab/CausalReasoningBenchmark/resolve/edd0fe1c1d1c5c0a4f0e3e67a50481ac679fb888/data/ds_0025/data/data.csv"},"ref_years":{"columns":null,"description":"How many of 1848 and 1859 each country is observed in","file_sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","format":"csv","frame_sha256":null,"generator":null,"kind":"file","licence":"redistributed by CausalReasoningBenchmark (MIT) under the original replication data's terms","member":null,"object":null,"params":{},"path":null,"read_kw":{},"records":null,"role":"reference","sha256":"c4a713c165661449eafadc60e8646b778b01b8c8bff567f77bc6adaa39ba2fb0","transforms":[{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"select"},{"args":{"columns":["year","Countryid","primratio","treatment"]},"op":"dropna"},{"args":{"query":"year in [1848, 1859]"},"op":"filter"},{"args":{"aggs":[{"column":"primratio","fn":"count","name":"n"}],"by":["Countryid","year"]},"op":"collapse"},{"args":{"aggs":[{"column":"n","fn":"count","name":"nyears"}],"by":["Countryid"]},"op":"collapse"}],"url":"https://huggingface.co/datasets/syrgkanislab/CausalReasoningBenchmark/resolve/edd0fe1c1d1c5c0a4f0e3e67a50481ac679fb888/data/ds_0025/data/data.csv"}},"difficulty":"M","id":"crb_papers.q047","level":"T2","meta":{"agent_reach":{"blocked_by":["reference_frame:ref_adopt","reference_frame:ref_panel","reference_frame:ref_post","reference_frame:ref_pre","reference_frame:ref_years"],"dtypes":{"ds_0025":{"primratio":"float64","treatment":"float64"}},"named":{"outcome":["primratio"],"treatment":["treatment"]},"reachable":false,"rule":"magi_bench.reach/1"},"answer_format":{"figure_transform":"identity","round_to":null,"unit_scale":1.0},"cluster":"ds_0025","crb":{"data_path":"data/ds_0025/data/data.csv","dataset_id":"ds_0025","gold":{"ci95":[-4.525834998078203,23.042285244253353],"ci_rule":"effect \u00b1 1.96\u00b7|standard_error| (evaluate_result.ci_bounds)","effect":9.258225123087575,"standard_error":7.032683735288662},"identification":{"causal_quantity":"ATT","group_variable":"Countryid","minimal_controlling_set":[],"outcomes":["primratio"],"reason_for_minimal_controlling_set":"The paper\u2019s event-study DiD specification includes country and year effects that absorb confounding from time-invariant differences and common shocks: \u201c\u03b3_i accounts for permanent observable and unobservable country characteristics\u2026 \u03d5_t accounts for any common time shocks that affect countries with and without civil war\u201d (Cross-national Evidence section, p. 1248). The identifying pre-trend diagnostics show no differential trends before treatment: \u201cthe estimated coefficients for n < 0 are close to zero and not statistically different from zero, lending credibility to the identifying assumption\u201d (p. 1249). Given only year, Countryid, treatment, and primratio are available, no additional covariates beyond these fixed effects are needed for identification in the 2\u00d72 cohort vs not-yet-treated comparisons.","strategy":"Difference-in-Differences","time_variable":"year","treatments":["treatment"]},"identification_strategy":"Difference-in-Differences"},"headline_keys":{"infer.ols":["coef.Interaction"]},"item_caveats":[{"kind":"implicit_window","note":"the question leaves implicit the comparison years the gold's script uses (pre year 1848)"}],"naive":{"dataset":"ds_0025","treatment":"treatment","y":"primratio"},"native_id":"47","reference":{"cite":1,"note":"Reference path transcribed from the item's own estimation script.","steps":[{"args":{"cluster":"Countryid","controls":["Post","TreatCohort"],"dataset":"ref_panel","x":["Interaction"],"y":"primratio"},"tool":"infer.ols","why":"sm.OLS(primratio, add_constant([Post, TreatCohort, Interaction])).fit(cov_type='cluster', groups=Countryid) (CR1 errors with statsmodels' default G/(G-1)\u00b7(N-1)/(N-K) correction)"}]},"reference_path":{"calls":[{"args":{"cluster":"Countryid","controls":["Post","TreatCohort"],"dataset":"ref_panel","x":["Interaction"],"y":"primratio"},"tool":"infer.ols"}],"needs":[],"provenance":"benchmark_script","quotes":{},"reference_datasets":["ref_adopt","ref_panel","ref_post","ref_pre","ref_years"],"source_id":"crb_papers.q047.script"}},"question":"How did primary-school enrollment (percent of ages 5-14) evolve after the onset of a country's first civil war in 1849 compared with countries that had not experienced civil war by 1859?","requires":[],"runner":null,"sources":["crb_papers.q047.gold","crb_papers.q047.script"],"split":"public","suite":"crb_papers","tags":["econ-bench","external","crb","did","ref_benchmark_script"],"target":"agent","tool":null,"truth":"CausalReasoningBenchmark gold at the pinned revision: causal_queries.json effect and standard_error (the value the item's estimation.py prints), scored as evaluate_result.py scores estimate_within_solution_ci"}