Take a good hard look at this graphic. I did a snapshot of the latest presidential approval polls as tallied on the realclearpolitics.com website:
That makes for 12 polls within 8 days. 7 polls have Obama at +10 or more, 2 of those 7 have Obama at +20. One has him at +9. So, 8 of twelve presidential approval polls, or 3/4 of those polls, put him in landslide territory. Two polls (Reuters/Ipsos and SUSA) show Obama with only +2 and +4, respectively. One poll, taken 2 days before the announcement of the death of Bin Laden and again for 2 days after, shows Obama at -1 (Daily Beast - I will deal with this questionable pollster -Schoen- in another post).
But RAS, ah, RAS now has Obama at -4. This is what we call a major outlier.
Here is another good example of RAS as an outlier:
2010, Hawaii, Senate: RAS claimed that Inouye was leading by +13 (D 53 - R 40)
Actual results: Inouye +51.2!!! (D 71.9- R 20.2). Gee, RAS was off by only 38.5 margin points!
2010, Hawaii, Governor: RAS claimed that Abercrombie was losing by 2 (R 49 - D 47)
Actual results: Abercrombie +17.0 (D 57.8 - R 40.8). Gee, RAS was off by only 19 margin points!!!
So, if RAS is currently showing President Obama at -4, then he is likely +15. Good going, RAS!!!
BTW, without RAS in the polling, this would be the following average of the other 11 polls:
Obama approval: 52.1
Obama disapproval: 41.4
Spread: Obama +10.7
So you see, an obvious outlier can move a margin average by more than one point, which is the point if you are trying to drive a negative narrative against someone.