update scores

Author: SvenKlaassen

doubleml/double_ml_did_cs.py

@@ -301,11 +301,11 @@ def _score_elements(self, y, d, t,
 
         if self.score == 'observational':
             if self.in_sample_normalization:
-                weight_psi_a = np.ones_like(y)   # would be the same with np.divide(d, np.mean(d))
-                weight_g_d1_t1 = np.divide(d, np.mean(d))
-                weight_g_d1_t0 = -1.0 * weight_g_d1_t1
-                weight_g_d0_t1 = -1.0 * weight_g_d1_t1
-                weight_g_d0_t0 = weight_g_d1_t1
+                weight_psi_a = np.divide(d, np.mean(d))
+                weight_g_d1_t1 = weight_psi_a
+                weight_g_d1_t0 = -1.0 * weight_psi_a
+                weight_g_d0_t1 = -1.0 * weight_psi_a
+                weight_g_d0_t0 = weight_psi_a
 
                 weight_resid_d1_t1 = np.divide(np.multiply(d, t),
                                                np.mean(np.multiply(d, t)))

doubleml/tests/_utils_did_cs_manual.py

@@ -160,11 +160,11 @@ def did_cs_score_elements(resid_d0_t0, resid_d0_t1, resid_d1_t0, resid_d1_t1,
 
     if score == 'observational':
         if in_sample_normalization:
-            weight_psi_a = np.ones_like(d)
-            weight_g_d1_t1 = np.divide(d, np.mean(d))
-            weight_g_d1_t0 = -1.0 * weight_g_d1_t1
-            weight_g_d0_t1 = -1.0 * weight_g_d1_t1
-            weight_g_d0_t0 = weight_g_d1_t1
+            weight_psi_a = np.divide(d, np.mean(d))
+            weight_g_d1_t1 = weight_psi_a
+            weight_g_d1_t0 = -1.0 * weight_psi_a
+            weight_g_d0_t1 = -1.0 * weight_psi_a
+            weight_g_d0_t0 = weight_psi_a
 
             weight_resid_d1_t1 = np.divide(np.multiply(d, t),
                                            np.mean(np.multiply(d, t)))