diff --git a/common/lib/capa/capa/tests/response_xml_factory.py b/common/lib/capa/capa/tests/response_xml_factory.py
index fe918ec5db..c7e43f5152 100644
--- a/common/lib/capa/capa/tests/response_xml_factory.py
+++ b/common/lib/capa/capa/tests/response_xml_factory.py
@@ -666,3 +666,39 @@ class StringResponseXMLFactory(ResponseXMLFactory):
 
     def create_input_element(self, **kwargs):
         return ResponseXMLFactory.textline_input_xml(**kwargs)
+
+class AnnotationResponseXMLFactory(ResponseXMLFactory):
+    """ Factory for creating <annotationresponse> XML trees """
+    def create_response_element(self, **kwargs):
+        """ Create a <annotationresponse> element """
+        return etree.Element("annotationresponse")
+
+    def create_input_element(self, **kwargs):
+        """ Create a <annotationinput> element."""
+
+        title = kwargs.get('title', 'super cool annotation')
+        text = kwargs.get('text', 'texty text')
+        comment = kwargs.get('comment', 'blah blah erudite comment blah blah')
+        comment_prompt = kwargs.get('comment_prompt', 'type a commentary below')
+        tag_prompt = kwargs.get('tag_prompt', 'select one tag')
+        options = kwargs.get('options', [
+            ('green', 'correct'),
+            ('eggs', 'incorrect'),
+            ('ham', 'partially-correct')
+        ])
+
+        # Create the <annotationinput> element
+        input_element = etree.Element("annotationinput")
+        etree.SubElement(input_element, 'title')
+        etree.SubElement(input_element, 'text')
+        etree.SubElement(input_element, 'comment')
+        etree.SubElement(input_element, 'comment_prompt')
+        etree.SubElement(input_element, 'tag_prompt')
+
+        options_element = etree.SubElement(input_element, 'options')
+        for (description, correctness) in options:
+            option_element = etree.SubElement(options_element, 'option', {'choice': correctness})
+            option_element.text = description
+
+        return input_element
+
diff --git a/common/lib/capa/capa/tests/test_files/annotationresponse.xml b/common/lib/capa/capa/tests/test_files/annotationresponse.xml
deleted file mode 100644
index 86af0bb789..0000000000
--- a/common/lib/capa/capa/tests/test_files/annotationresponse.xml
+++ /dev/null
@@ -1,17 +0,0 @@
-<problem display_name="Exercise 1">
-    <annotationresponse>
-        <annotationinput>
-            <title>the title</title>
-            <text>the text</text>
-            <comment>the comment</comment>
-            <comment_prompt>Type a commentary below:</comment_prompt>
-            <tag_prompt>Select one or more tags:</tag_prompt>
-            <options>
-                <option choice="correct">green</option>
-                <option choice="incorrect">eggs</option>
-                <option choice="partially-correct">ham</option>
-            </options>
-        </annotationinput>
-    </annotationresponse>
-    <solution>Instructor text here...</solution>
-</problem>
diff --git a/common/lib/capa/capa/tests/test_responsetypes.py b/common/lib/capa/capa/tests/test_responsetypes.py
index 33b84d213d..d592bff976 100644
--- a/common/lib/capa/capa/tests/test_responsetypes.py
+++ b/common/lib/capa/capa/tests/test_responsetypes.py
@@ -772,3 +772,40 @@ class SchematicResponseTest(ResponseTest):
         # (That is, our script verifies that the context
         # is what we expect)
         self.assertEqual(correct_map.get_correctness('1_2_1'), 'correct')
+
+class AnnotationResponseTest(ResponseTest):
+    from response_xml_factory import AnnotationResponseXMLFactory
+    xml_factory_class = AnnotationResponseXMLFactory
+
+    def test_grade(self):
+        (correct, partially, incorrect) = ('correct', 'partially-correct', 'incorrect')
+
+        answer_id = '1_2_1'
+        options = (('x', correct),('y', partially),('z', incorrect))
+        make_answer = lambda option_ids: {answer_id: json.dumps({'options': option_ids })}
+
+        tests = [
+            {'correctness': correct, 'points': 2,'answers': make_answer([0]) },
+            {'correctness': partially, 'points': 1, 'answers': make_answer([1]) },
+            {'correctness': incorrect, 'points': 0, 'answers': make_answer([2]) },
+            {'correctness': incorrect, 'points': 0, 'answers': make_answer([0,1,2]) },
+            {'correctness': incorrect, 'points': 0, 'answers': make_answer([]) },
+            {'correctness': incorrect, 'points': 0, 'answers': make_answer('') },
+            {'correctness': incorrect, 'points': 0, 'answers': make_answer(None) },
+            {'correctness': incorrect, 'points': 0, 'answers': {answer_id: 'null' } },
+        ]
+
+        for (index, test) in enumerate(tests):
+            expected_correctness = test['correctness']
+            expected_points = test['points']
+            answers = test['answers']
+
+            problem = self.build_problem(options=options)
+            correct_map = problem.grade_answers(answers)
+            actual_correctness = correct_map.get_correctness(answer_id)
+            actual_points = correct_map.get_npoints(answer_id)
+
+            self.assertEqual(expected_correctness, actual_correctness,
+                             msg="%s should be marked %s" % (answer_id, expected_correctness))
+            self.assertEqual(expected_points, actual_points,
+                             msg="%s should have %d points" % (answer_id, expected_points))
\ No newline at end of file