{
  "about": "CauSciBench native output: 'record' has the shape their evaluation/extract.py reads (a run's records in one list rescore with their evaluation/run_evaluation.py); the method is the tool-family label of the answer call (tool-family accuracy from tool calls, not a method the model named); 'scores' are the per-item terms of evaluation/metrics.py.",
  "answer_ref": "call_ba7ec5d1:coef.treatment_price_normalized",
  "attempt": 1,
  "epoch": 1,
  "gold": {
    "effect": -0.338,
    "effect_as_stored": "-0.338",
    "is_significant": 1,
    "source_id": "causcibench_real.gold.q153",
    "std_error": 0.054,
    "std_error_as_stored": "0.054"
  },
  "label": {
    "family": "iv",
    "method": "iv",
    "rule": "two-stage least squares",
    "tool_id": "infer.iv_2sls"
  },
  "native_id": 153,
  "precision_limited": false,
  "record": {
    "dataset_path": "data/csv_files/realdata/hartford_deepiv.csv",
    "effect": -0.338,
    "method": "iv",
    "name": "Deep IV: A Flexible Approach for Counterfactual Prediction",
    "query": "How does increase in product cost affect customer demand?",
    "result": {
      "final_result": {
        "causal_effect": -0.33807247272434704,
        "covariates": [
          "covariate_emotion_0",
          "covariate_emotion_1",
          "covariate_emotion_2",
          "covariate_emotion_3",
          "covariate_emotion_4",
          "covariate_emotion_5",
          "covariate_emotion_6",
          "temporal_time"
        ],
        "method": "iv",
        "outcome_variable": "outcome_demand",
        "standard_deviation": 0.060156647201061696,
        "treatment_variable": "treatment_price_normalized"
      },
      "retries": null
    }
  },
  "scores": {
    "control_overlap": 1.0,
    "effect_correct": 1,
    "method_correct": 1,
    "outcome_match": 1,
    "relative_error_pct": 0.02144163442219435,
    "treatment_match": 1
  },
  "status": "passed",
  "task_id": "causcibench_real.q153"
}