diff --git a/recopytex/store/filesystem/loader.py b/recopytex/store/filesystem/loader.py
index d6dfdab..b86fff7 100644
--- a/recopytex/store/filesystem/loader.py
+++ b/recopytex/store/filesystem/loader.py
@@ -156,7 +156,7 @@ class CSVLoader(Loader):
         >>> questions = loader.get_exam_questions(exams)
         >>> scores = loader.get_questions_scores(questions)
         >>> scores.columns
-        Index(['id', 'origin', 'student_name', 'score'], dtype='object')
+        Index(['origin', 'question_id', 'student_name', 'score'], dtype='object')
         >>> scores["student_name"].unique()
         array(['Star Tice', 'Umberto Dingate', 'Starlin Crangle',
                'Humbert Bourcq', 'Gabriella Handyside', 'Stewart Eaves',
@@ -170,20 +170,27 @@ class CSVLoader(Loader):
         group_questions = questions.groupby("origin")
         for origin, questions_df in group_questions:
             scores_df = pd.read_csv(origin)
+            student_names = [
+                c
+                for c in scores_df.columns
+                if c not in self._config["csv_fields"].values()
+            ]
+
             common_columns = [c for c in questions_df.columns if c in scores_df.columns]
             scores_df = pd.merge(scores_df, questions_df, on=common_columns)
             scores_df = scores_df.drop(
                 columns=list(self._config["csv_fields"].values()) + ["exam_id"]
             )
-            kept_columns = ["id", "origin"]
-            student_names = [c for c in scores_df.columns if c not in kept_columns]
+
+            kept_columns = [c for c in scores_df if c not in student_names]
             scores_df = pd.melt(
                 scores_df,
                 id_vars=kept_columns,
                 value_vars=student_names,
                 var_name="student_name",
                 value_name="score",
-            )
+            ).rename(columns={"id": "question_id"})
+
             scores.append(scores_df)
 
         return pd.concat(scores)