diff --git a/mesmer/stats/_auto_regression.py b/mesmer/stats/_auto_regression.py
index 5f2ddac4..53db871b 100644
--- a/mesmer/stats/_auto_regression.py
+++ b/mesmer/stats/_auto_regression.py
@@ -72,7 +72,6 @@ def _fit_auto_regression_scen_ens(*objs, dim, ens_dim, lags):
         Dimension along which to fit the auto regression.
     ens_dim : str
         Dimension name of the ensemble members, None if no ensemble is provided.
-        If provided, must also have coordinates.
     lags : int
         The number of lags to include in the model.
 
@@ -91,25 +90,19 @@ def _fit_auto_regression_scen_ens(*objs, dim, ens_dim, lags):
     """
 
     ar_params_scen = list()
-
     for obj in objs:
         ar_params = fit_auto_regression(obj, dim=dim, lags=int(lags))
+
+        #TODO: think about weighting! see https://github.com/MESMER-group/mesmer/issues/307
+        if ens_dim in ar_params.dims:
+            ar_params = ar_params.mean(ens_dim)
+
         ar_params_scen.append(ar_params)
 
     ar_params_scen = xr.concat(ar_params_scen, dim="scen")
-    
-    # TODO: think about weighting! see https://github.com/MESMER-group/mesmer/issues/307
-    if ens_dim in ar_params.dims:
-            # mean over ensemble members
-            ar_params_scen = ar_params_scen.mean(dim=ens_dim)
-    
-    # mean over scenarios
-    ar_params = ar_params_scen.mean(dim="scen")
-
-    # clean up
-    ar_params = ar_params.drop_vars("nobs") # don't need it in the result
-    if ens_dim in ar_params.dims:
-        ar_params = ar_params.drop_dims(ens_dim)
+
+    # return the mean over all scenarios
+    ar_params = ar_params_scen.mean("scen")
 
     return ar_params
 
diff --git a/tests/unit/test_auto_regression_scen_ens.py b/tests/unit/test_auto_regression_scen_ens.py
index 1e7a1874..b9840c93 100644
--- a/tests/unit/test_auto_regression_scen_ens.py
+++ b/tests/unit/test_auto_regression_scen_ens.py
@@ -73,12 +73,7 @@ def test_fit_auto_regression_scen_ens_one_scen(std):
     )
 
     expected = mesmer.stats.fit_auto_regression(da, dim="time", lags=3)
-    expected["variance"] = sum(expected.variance * (expected.nobs - 1)) / sum(
-        expected.nobs - 1
-    )
-    expected["coeffs"] = expected.coeffs.mean("ens")
-    expected["intercept"] = expected.intercept.mean("ens")
-    expected = expected.drop_vars(["nobs", "ens"])
+    expected = expected.mean("ens")
 
     xr.testing.assert_allclose(result, expected)
     np.testing.assert_allclose(np.sqrt(result.variance), std, rtol=1e-1)
@@ -98,7 +93,6 @@ def test_fit_auto_regression_scen_ens_multi_scen():
     expected = mesmer.stats.fit_auto_regression(da, dim="time", lags=3)
     expected = expected.unstack("scen_ens")
     expected = expected.mean("ens").mean("scen")
-    expected = expected.drop_vars(["nobs"])
 
     xr.testing.assert_equal(result, expected)
 
@@ -112,6 +106,5 @@ def test_fit_auto_regression_scen_ens_no_ens_dim():
     )
 
     expected = mesmer.stats.fit_auto_regression(da, dim="time", lags=3)
-    expected = expected.drop_vars(["nobs"])
 
     xr.testing.assert_allclose(result, expected)