Character data left after entity references are replaced with their text
Given the element
<PUBLISHER>A & M Records</PUBLISHER>
The parsed character data is
A & M Records
CDATA sections are replaced with their actual text.
APIs report the parsed character data; not the unparsed data.