{
  "context": "This criteria evaluates how well the engineer uses the spreg package to implement a spatial error model for handling spatially autocorrelated errors in regression analysis. The focus is entirely on proper usage of spreg's spatial error model classes and methods.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Spatial error class usage",
      "description": "Uses one of spreg's spatial error model classes (GM_Error, GM_Error_Het, GM_Error_Hom, or ML_Error) to estimate the model rather than implementing spatial error correction manually.",
      "max_score": 30
    },
    {
      "name": "Lambda parameter extraction",
      "description": "Correctly extracts the spatial error autocorrelation parameter (lambda) from the fitted model object, typically accessed via model.lam or model.betas attribute.",
      "max_score": 15
    },
    {
      "name": "Coefficient extraction",
      "description": "Properly extracts coefficient estimates from the fitted model object (e.g., using model.betas) and formats them according to the API specification.",
      "max_score": 15
    },
    {
      "name": "Standard errors extraction",
      "description": "Correctly retrieves standard errors from the fitted model (e.g., model.std_err) and includes them in the return value.",
      "max_score": 10
    },
    {
      "name": "Pseudo R-squared",
      "description": "Extracts the pseudo R-squared statistic from the model object (e.g., model.pr2 or model.pr2_e) for model fit assessment.",
      "max_score": 10
    },
    {
      "name": "Spatial weights handling",
      "description": "Passes the spatial weights matrix W to the spatial error model class constructor in the correct format expected by spreg.",
      "max_score": 10
    },
    {
      "name": "Variable names mapping",
      "description": "Properly uses the variable_names parameter to create a meaningful mapping between variable names and coefficients/standard errors in the return dictionary.",
      "max_score": 10
    }
  ]
}