diff --git a/solenoid/people/models.py b/solenoid/people/models.py
index cbe2c22..b248261 100644
--- a/solenoid/people/models.py
+++ b/solenoid/people/models.py
@@ -86,8 +86,8 @@ def __str__(self):
         return self.name
 
     name = models.CharField(max_length=100, unique=True)
-    # DLCs are created as needed during the CSV import process, and we don't
-    # have liaison information available at that time.
+    # DLCs are created as needed during the metadata import process, and we
+    # don't have liaison information available at that time.
     liaison = models.ForeignKey(
         Liaison,
         blank=True,
@@ -105,9 +105,9 @@ class Meta:
     def __str__(self):
         return "{self.first_name} {self.last_name}/{self.dlc}".format(self=self)  # noqa
 
-    # Authors may have blank DLCs in the CSV, but if that happens we're going
-    # to push it back to the Sympletic layer and request that the user fix it
-    # there.
+    # Authors may have blank DLCs in a given paper's metadata, but if that
+    # happens we're going to push it back to the Sympletic layer and request
+    # that the user fix it there.
     dlc = models.ForeignKey(DLC, on_delete=models.CASCADE)
     email = models.EmailField(help_text="Author email address")
     first_name = models.CharField(max_length=30)
@@ -140,8 +140,8 @@ def get_by_mit_id(cls, mit_id):
 
     # These properties allow us to get and set the mit ID using the normal
     # API; in particular, we can directly set the ID from the MTI ID value in
-    # the CSV files. However, under the hood, we're throwing out the sensitive
-    # data and storing hashes. Hooray!
+    # the paper metadata. However, under the hood, we're throwing out the
+    # sensitive data and storing hashes. Hooray!
     @property
     def mit_id(self):
         return self._mit_id_hash
diff --git a/solenoid/people/tests.py b/solenoid/people/tests.py
index 1c25f62..1b8fd30 100644
--- a/solenoid/people/tests.py
+++ b/solenoid/people/tests.py
@@ -1,10 +1,10 @@
 import hashlib
 import os
 
+from django.test import Client, TestCase, override_settings
 from django.urls import resolve, reverse
-from django.test import TestCase, Client, override_settings
 
-from .models import Liaison, DLC, Author
+from .models import DLC, Author, Liaison
 from .views import LiaisonList
 
 
@@ -115,7 +115,7 @@ def test_queryset(self):
 class DLCTests(TestCase):
 
     # We used to have an option for letting people manually edit DLCs. We've
-    # backed this out in favor of only creating DLCs via CSV import - there
+    # backed this out in favor of only creating DLCs via data import - there
     # is name authority work going on upstream we would prefer to rely on.
     # If it's desirable to put that functionality back, there were tests for it
     # here, and it last existed at commit
diff --git a/solenoid/records/models.py b/solenoid/records/models.py
index 98c414b..232be0e 100644
--- a/solenoid/records/models.py
+++ b/solenoid/records/models.py
@@ -99,11 +99,11 @@ def save(self, *args, **kwargs):
     def create_citation(paper_data):
         """Create text suitable for the citation field.
 
-        Some CSV records fill in the citation field, but some leave it blank;
-        in those cases we need to build our own citation text. Stakeholders
-        have indicated that it doesn't matter what citation format we use or
-        whether the citation is complete (as long as it includes author, title,
-        and journal title).
+        Some Elements papers include the citation field in their metadata,
+        but some leave it blank; in those cases we need to build our own
+        citation text. Stakeholders have indicated that it doesn't matter what
+        citation format we use or whether the citation is complete (as long as
+        it includes author, title, and journal title).
 
         This method assumes that the supplied paper metadata has already been
         validated; it does not perform any validation.