diff --git a/src/qa4sm_reader/img.py b/src/qa4sm_reader/img.py
index e5e61dd..9b56dac 100644
--- a/src/qa4sm_reader/img.py
+++ b/src/qa4sm_reader/img.py
@@ -463,7 +463,7 @@ def _metric_stats(self, metric, id=None) -> list:
         # The number of observations are needed for the averaging of correlation values
         for Var in self._iter_vars(type="metric",
                                    filter_parms={'metric': 'n_obs'}):
-            nobs = Var.values
+            nobs = Var.values[Var.varname]
 
         # get stats by metric
         for Var in self._iter_vars(type="metric", filter_parms=filters):
diff --git a/src/qa4sm_reader/plotting_methods.py b/src/qa4sm_reader/plotting_methods.py
index 7ffa6b4..024657c 100644
--- a/src/qa4sm_reader/plotting_methods.py
+++ b/src/qa4sm_reader/plotting_methods.py
@@ -1624,7 +1624,8 @@ def output_dpi(res,
     return float(dpi)
 
 
-def average_non_additive(values: Union[pd.Series, np.array], nobs: pd.Series) -> float:
+def average_non_additive(values: Union[pd.Series, np.array],
+                         nobs: pd.Series) -> float:
     """
     Calculate the average of non-additive values, such as correlation
     scores, as recommended in:
@@ -1647,7 +1648,6 @@ def average_non_additive(values: Union[pd.Series, np.array], nobs: pd.Series) ->
 
     # Transform to Fisher's z-scores
     z_scores = np.arctanh(values)
-
     # Remove the entries where there are NaNs
     mask = np.isfinite(values) & np.isfinite(nobs)
     z_scores = z_scores[mask]