diff --git a/pandas_ods_reader/parser.py b/pandas_ods_reader/parser.py
index c4d600d..ec9d49f 100644
--- a/pandas_ods_reader/parser.py
+++ b/pandas_ods_reader/parser.py
@@ -21,7 +21,7 @@ def load_ods(doc, sheet, headers=True, columns=None):
         # row is a list of cells
         if headers and i == 0 and not columns:
             # columns as lists in a dictionary
-            df_dict = {cell.value: [] for cell in row if cell.value}
+            df_dict = {cell.value: pd.Series() for cell in row if cell.value}
             # create index for the column headers
             col_index = {
                 j: cell.value for j, cell in enumerate(row) if cell.value}
@@ -30,7 +30,7 @@ def load_ods(doc, sheet, headers=True, columns=None):
             columns = columns if columns else (
                 ["column_%s" % j for j in range(len(row))])
             # columns as lists in a dictionary
-            df_dict = {column: pd.Series([] for column in columns)}
+            df_dict = {column: pd.Series() for column in columns}
             # create index for the column headers
             col_index = {j: column for j, column in enumerate(columns)}
             if headers:
@@ -38,7 +38,7 @@ def load_ods(doc, sheet, headers=True, columns=None):
         for j, cell in enumerate(row):
             if j < len(col_index):
                 # use header instead of column index
-                df_dict[col_index[j]].append(cell.value)
+                df_dict[col_index[j]].append(pd.Series([cell.value]))
             else:
                 continue
     # and convert to a DataFrame