Suggested edits

Author: jbrockmendel

doc/source/user_guide/basics.rst

@@ -2378,6 +2378,10 @@ To select string columns include ``str``:
 
    df.select_dtypes(include=[str])
 
+.. note::
+
+    This is a change in pandas 3.0. Previously strings were stored in ``object`` dtype columns, so would be selected with ``include=[object]``.  See https://pandas.pydata.org/docs/user_guide/migration-3-strings.html#hardcoded-use-of-object-dtype.
+
 To see all the child dtypes of a generic ``dtype`` like ``numpy.number`` you
 can define a function that returns a tree of child dtypes:
 

pandas/core/frame.py

@@ -5250,7 +5250,7 @@ def predicate(arr: ArrayLike) -> bool:
             # GH#61916
             warnings.warn(
                 "For backward compatibility, 'str' dtypes are included by "
-                "select_dtypes when object dtypes are specified. "
+                "select_dtypes when 'object' dtype is specified. "
                 "This behavior is deprecated and will be removed in a future "
                 "version. Explicitly pass 'str' to `include` to select them, "
                 "or to `exclude` to remove them and silence this warning.",