{
  "context": "This evaluation criteria assesses how well the engineer uses the spreg package to compute model fit statistics (R-squared and pseudo R-squared) for regression models. The focus is on correctly using the OLS and TSLS classes and accessing their fit metric attributes.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "OLS class usage",
      "description": "Uses spreg.OLS class to fit ordinary least squares regression models for computing R-squared metrics",
      "max_score": 25
    },
    {
      "name": "R-squared access",
      "description": "Correctly accesses the r2 attribute from fitted OLS model objects to retrieve the R-squared value",
      "max_score": 15
    },
    {
      "name": "Adjusted R-squared access",
      "description": "Correctly accesses the ar2 attribute from fitted OLS model objects to retrieve the adjusted R-squared value",
      "max_score": 15
    },
    {
      "name": "TSLS class usage",
      "description": "Uses spreg.TSLS class with proper parameters (y, x, yend, q) to fit two-stage least squares models with endogenous variables and instruments",
      "max_score": 25
    },
    {
      "name": "Pseudo R-squared access",
      "description": "Correctly accesses the pr2 attribute from fitted TSLS model objects to retrieve the pseudo R-squared value",
      "max_score": 15
    },
    {
      "name": "Array conversion",
      "description": "Properly converts input data to numpy arrays with correct shapes (nx1 for y, nxk for x) as required by spreg methods",
      "max_score": 5
    }
  ]
}