Differences between Leip and Wasserman (Cook report) counts

[Goodgulf Grayteeth]

David Wasserman of the Cook Political Report has been maintaining a spreadsheet (sorry, no link; don't have my 20 posts yet) of updated 2016 election popular vote totals.

The differences between the national totals that Leip and Wasserman give for Clinton and Trump are relatively small and don't concern me because they likely reflect who's updated their spreadsheets more recently.

However, Leip has consistently shown noticeably more votes for others than Wasserman has.  For everyone besides Clinton and Trump, Leip's totals are near 8 million, while Wasserman's are just over 7.6 million.   This gap has been persistent over a period of a week or two.

So there must be a reason for this, some data that Leip is including and Wasserman is not.   Anybody have an explanation?

[RI]

Write-ins. Seems Wasserman doesn't include "unofficial" write-ins.