diff --git a/src/tea_tasting/aggr.py b/src/tea_tasting/aggr.py
index d03380c..9b10794 100644
--- a/src/tea_tasting/aggr.py
+++ b/src/tea_tasting/aggr.py
@@ -284,24 +284,24 @@ def read_aggregates(
     demean_cols = tuple({*var_cols, *itertools.chain(*cov_cols)})
     if len(demean_cols) > 0:
         demean_expr = {
-            _DEMEAN.format(col): data[col] - data[col].mean()  # type: ignore
+            _DEMEAN.format(col): data[col] - data[col].cast("float").mean()  # type: ignore
             for col in demean_cols
         }
         grouped_data = data.group_by(group_col) if group_col is not None else data  # type: ignore
         data = grouped_data.mutate(**demean_expr)  # type: ignore
 
     count_expr = {_COUNT: data.count()} if has_count else {}
-    mean_expr = {_MEAN.format(col): data[col].mean() for col in mean_cols}  # type: ignore
+    mean_expr = {_MEAN.format(col): data[col].cast("float").mean() for col in mean_cols}  # type: ignore
     var_expr = {
         _VAR.format(col): (
             data[_DEMEAN.format(col)] * data[_DEMEAN.format(col)]
-        ).sum().cast("float") / (data.count() - 1)  # type: ignore
+        ).sum() / (data.count() - 1)  # type: ignore
         for col in var_cols
     }
     cov_expr = {
         _COV.format(left, right): (
             data[_DEMEAN.format(left)] * data[_DEMEAN.format(right)]
-        ).sum().cast("float") / (data.count() - 1)  # type: ignore
+        ).sum() / (data.count() - 1)  # type: ignore
         for left, right in cov_cols
     }