The actual poll states that it samples "registered voters who are likely to participate in the November 2012 election." (p. 33). I'm not sure how else to interpret that other than as a "likely voter." If you're registered to vote, and "likely" to vote in the November 2012 election, how are you not a likely voter?
That being said, Chilenski Strategies weights this survey sample to an R+1 partisan identification, 35% Republican, 34% Democrat, 31% Independent. That's a good deal less Democratic than 2008 exit polls measured in Missouri (40D/34R/26I, D+6). And that Dem-friendly electorate is the only way Obama is able to win Missouri, using Chilenski Strategies crosstabs:
34% of the Missouri electorate, Republicans 37%, and Independents 28%?
If you average partisan turnout in Missouri from the 2010 and 2008 elections, and re-weight the Chilenski Strategies poll to those turnout numbers, here is what you get: