diff --git a/examples/C/ner/ner_example.c b/examples/C/ner/ner_example.c
index 7e4fa26f2..9a1078c9c 100644
--- a/examples/C/ner/ner_example.c
+++ b/examples/C/ner/ner_example.c
@@ -91,7 +91,7 @@ void print_entity (
     pos = mitie_ner_get_detection_position(dets, i);
     len = mitie_ner_get_detection_length(dets, i);
     // Print the label and score for each named entity and also the text of the named entity
-    // itself.
+    // itself.  The larger the score the more confident MITIE is in the tag.
     printf("   Tag %lu: Score: %0.3f: %s: ", mitie_ner_get_detection_tag(dets,i),
                                              mitie_ner_get_detection_score(dets,i),
                                              mitie_ner_get_detection_tagstr(dets,i));
diff --git a/examples/cpp/ner/ner_example.cpp b/examples/cpp/ner/ner_example.cpp
index dabc22df6..5f7ce094f 100644
--- a/examples/cpp/ner/ner_example.cpp
+++ b/examples/cpp/ner/ner_example.cpp
@@ -76,7 +76,8 @@ int main(int argc, char** argv)
         // Now detect all the entities in the text file we loaded and print them to the screen.
         // The output of this function is a set of "chunks" of tokens, each a named entity.
         // Additionally, if it is useful for your application a confidence score for each "chunk"
-        // is available by using the predict() method.
+        // is available by using the predict() method.  The larger the score the more
+        // confident MITIE is in the tag.
         ner.predict(tokens, chunks, chunk_tags, chunk_scores);
 
         // If a confidence score is not necessary for your application you can detect entities
diff --git a/examples/java/NerExample.java b/examples/java/NerExample.java
index d8d60ee1e..038d5bbf1 100644
--- a/examples/java/NerExample.java
+++ b/examples/java/NerExample.java
@@ -42,8 +42,9 @@ public static void main(String args[])
             // define the range of tokens in the words vector that are part of the entity.
             // There is also a tag which indicates which element of possibleTags is
             // associated with the entity. There is also a score which indicates a
-            // confidence associated with the predicted tag. So we can print out all
-            // the tagged entities as follows:
+            // confidence associated with the predicted tag (larger values mean MITIE is
+            // more confident in its prediction). So we can print out all the tagged
+            // entities as follows:
             EntityMention entity = entities.get(i);
             String tag = possibleTags.get(entity.getTag());
             Double score = entity.getScore();
diff --git a/examples/python/ner.py b/examples/python/ner.py
index 698343f8f..fd49f6870 100755
--- a/examples/python/ner.py
+++ b/examples/python/ner.py
@@ -31,9 +31,10 @@
 print "\nNumber of entities detected:", len(entities)
 
 # entities is a list of tuples, each containing an xrange that indicates which
-# tokens are part of the entity, the entity tag and an associate score.  The
+# tokens are part of the entity, the entity tag, and an associate score.  The
 # entities are also listed in the order they appear in the input text file.
 # Here we just print the score, tag, and text for each entity to the screen.
+# The larger the score the more confident MITIE is in its prediction.
 for e in entities:
     range = e[0]
     tag = e[1]