diff --git a/river/stats/kolmogorov_smirnov.py b/river/stats/kolmogorov_smirnov.py
index 873e3dd731..e3689a9613 100644
--- a/river/stats/kolmogorov_smirnov.py
+++ b/river/stats/kolmogorov_smirnov.py
@@ -226,14 +226,6 @@ def __init__(self, statistic="ks"):
         self.n_samples = 0
         self.statistic = statistic
 
-    @staticmethod
-    def ca(p_value):
-        return (-0.5 * math.log(p_value)) ** 0.5
-
-    @classmethod
-    def ks_threshold(cls, p_value, n_samples):
-        return cls.ca(p_value) * (2.0 * n_samples / n_samples**2)
-
     def update(self, x, y):
         keys = ((x, 0), (y, 1))
 
@@ -282,7 +274,11 @@ def get(self):
         else:
             raise ValueError(f"Unknown statistic {self.statistic}, expected one of: ks, kuiper")
 
-    def test_ks_threshold(self, ca):
+    @staticmethod
+    def _ca(p_value):
+        return (-0.5 * math.log(p_value)) ** 0.5
+
+    def _test_ks_threshold(self, ca):
         """
         Test whether the reference and sliding window follows the same or different probability distribution.
         This test will return `True` if we **reject** the null hypothesis that
diff --git a/river/stats/test_ks.py b/river/stats/test_kolmogorov_smirnov.py
similarity index 93%
rename from river/stats/test_ks.py
rename to river/stats/test_kolmogorov_smirnov.py
index baa979189e..24b2fd4f97 100644
--- a/river/stats/test_ks.py
+++ b/river/stats/test_kolmogorov_smirnov.py
@@ -42,4 +42,4 @@ def test_incremental_ks_statistics():
 
     assert np.allclose(np.array(incremental_ks_statistics), np.array(ks_2samp_statistics))
 
-    assert incremental_ks.test_ks_threshold(ca=incremental_ks.ca(p_value=0.05)) is True
+    assert incremental_ks._test_ks_threshold(ca=incremental_ks._ca(p_value=0.05)) is True