diff --git a/sdmetrics/single_table/table_structure.py b/sdmetrics/single_table/table_structure.py
index edbc511e..06a97ede 100644
--- a/sdmetrics/single_table/table_structure.py
+++ b/sdmetrics/single_table/table_structure.py
@@ -36,11 +36,12 @@ def compute_breakdown(cls, real_data, synthetic_data):
             synthetic_data (pandas.DataFrame):
                 The synthetic data.
         """
-        synthetic_columns = set(synthetic_data.columns)
-        real_columns = set(real_data.columns)
-        intersection_columns = real_columns & synthetic_columns
-        union_columns = real_columns | synthetic_columns
-        score = len(intersection_columns) / len(union_columns)
+        real_columns_dtypes = set(zip(real_data.columns, map(str, real_data.dtypes)))
+        synthetic_columns_dtypes = set(zip(synthetic_data.columns, map(str, synthetic_data.dtypes)))
+
+        intersection = real_columns_dtypes & synthetic_columns_dtypes
+        union = real_columns_dtypes | synthetic_columns_dtypes
+        score = len(intersection) / len(union)
 
         return {'score': score}
 
diff --git a/tests/integration/reports/multi_table/test_diagnostic_report.py b/tests/integration/reports/multi_table/test_diagnostic_report.py
index 57ff4cc6..f006915c 100644
--- a/tests/integration/reports/multi_table/test_diagnostic_report.py
+++ b/tests/integration/reports/multi_table/test_diagnostic_report.py
@@ -47,6 +47,7 @@ def test_end_to_end_with_metrics_failing(self):
         """Test the ``DiagnosticReport`` report when some metrics crash.
 
         This test makes fail the 'Boundary' property to check that the report still works.
+        The TableStructure should no longer be 1.0 since there is some dtype mismatch.
         """
         real_data, synthetic_data, metadata = load_demo(modality='multi_table')
         real_data['users']['age'].iloc[0] = 'error_1'
@@ -62,7 +63,7 @@ def test_end_to_end_with_metrics_failing(self):
         # Assert
         expected_properties = pd.DataFrame({
             'Property': ['Data Validity', 'Data Structure', 'Relationship Validity'],
-            'Score': [1.0, 1.0, 1.0],
+            'Score': [1.0, 0.6761904761904761, 1.0],
         })
         expected_details = pd.DataFrame({
             'Table': [
@@ -119,7 +120,7 @@ def test_end_to_end_with_metrics_failing(self):
                 None,
             ],
         })
-        assert results == 1.0
+        assert results == 0.892063492063492
         pd.testing.assert_frame_equal(
             report.get_properties(), expected_properties, check_exact=False, atol=2e-2
         )
diff --git a/tests/unit/single_table/test_table_structure.py b/tests/unit/single_table/test_table_structure.py
index 4d2e4c61..3d6db5e2 100644
--- a/tests/unit/single_table/test_table_structure.py
+++ b/tests/unit/single_table/test_table_structure.py
@@ -50,6 +50,28 @@ def test_compute_breakdown(self, real_data):
         expected_result = {'score': 1.0}
         assert result == expected_result
 
+    def test_compute_breakdown_with_different_dtypes(self, real_data):
+        """Test the ``compute_breakdown`` method with different data types.
+
+        - Real data has 5 columns.
+        - Synthetic data is identical except 'col_1' has a different data type.
+        - Total unique (column_name, dtype) combinations: 6.
+        - Matching combinations: 4.
+        - Expected score: 4 / 6 = 2 / 3.
+        """
+        # Setup
+        synthetic_data = real_data.copy()
+        synthetic_data['col_1'] = synthetic_data['col_1'].astype('float')
+
+        metric = TableStructure()
+
+        # Run
+        result = metric.compute_breakdown(real_data, synthetic_data)
+
+        # Assert
+        expected_result = {'score': 2 / 3}
+        assert result == expected_result
+
     def test_compute_breakdown_with_missing_columns(self, real_data):
         """Test the ``compute_breakdown`` method with missing columns."""
         # Setup