The size of random sample that would provide the same standard error as the actual sampling plan on which a survey result is based.

Source:
http://www.nielsenmedia.com/glossary/
http://www.arbitron.com/downloads/terms_brochure.pdf