diff --git a/python_sampling_tool/tests/test_main.py b/python_sampling_tool/tests/test_main.py
new file mode 100644
index 0000000..5af7c83
--- /dev/null
+++ b/python_sampling_tool/tests/test_main.py
@@ -0,0 +1,44 @@
+import pytest
+from python_sampling_tool import main
+
+
+def test_stratified_sampling_total_size():
+    sample_size = 20
+    num_jobs = 5
+    samples = main.perform_stratified_sampling(sample_size, num_jobs)
+    assert isinstance(samples, list)
+    # Gesamte Anzahl der Datenpunkte sollte der Samplegröße entsprechen
+    total_points = sum(s['data_points'] for s in samples)
+    assert total_points == sample_size
+    # Jeder Eintrag sollte einen Job-Identifier enthalten
+    for s in samples:
+        assert 'job_id' in s
+        assert 'data_points' in s
+        assert isinstance(s['data_points'], int)
+        assert isinstance(s['job_id'], str)
+
+
+def test_stratified_sampling_balanced_distribution():
+    sample_size = 15
+    num_jobs = 3
+    samples = main.perform_stratified_sampling(sample_size, num_jobs)
+    # Prüfe, dass alle Jobs berücksichtigt sind
+    job_ids = [s['job_id'] for s in samples]
+    assert len(job_ids) == num_jobs
+    # Verteilung der Datenpunkte sollte einigermaßen gleichmäßig sein
+    counts = [s['data_points'] for s in samples]
+    avg = sum(counts) / num_jobs
+    for c in counts:
+        assert abs(c - avg) <= 1
+
+
+def test_stratified_sampling_small_sample():
+    sample_size = 3
+    num_jobs = 5
+    samples = main.perform_stratified_sampling(sample_size, num_jobs)
+    # Summe sollte immer noch der Samplegröße entsprechen
+    total_points = sum(s['data_points'] for s in samples)
+    assert total_points == sample_size
+    # Nicht mehr Jobs als Samples erhalten Datenpunkte
+    nonzero = [s for s in samples if s['data_points'] > 0]
+    assert len(nonzero) <= sample_size
\ No newline at end of file