I prepared this spreadsheet using the data from pollster.com's OH GE polls page.
Here's a chart of Obama's leads (Obama minus McCain for each poll, +ve or -ve) generated by the spreadsheet which compares Rasmussen Reports (RR)'s numbers with trends from all other pollsters:
The blue curve (a so-called "polyline") is the 5-poll moving average (i.e. it's the average of the 5 most recent polls) for non-Rasmussen polls. The red polyline shows RR's trends (the big red dots are RR's poll results) while the green polyline shows those for Non-Rasmussen pollsters. The light pink line is the 3-poll moving average.
Clearly, RR's results have always been below the 5-poll moving average for non-Rasmussen polls and almost always below the 3-poll moving average for the same.
The latest polls, trends and moving averages from all non-Rasmussen pollsters' results show an approximate 4-6% lead for Obama in OH as of last week, in marked contrast with RR's latest poll showing a 10% lead for McCain. That's a 15% swing towards McCain in the Rasmussen poll, a clear outlier.
Poblano and his readership have some thoughts about RR's poll (in contrast to PPP's latest poll) here: Dueling Ohio Polls.
Here is a chart (from the same spreadsheet) showing a similar outlier analysis for Florida:
We have Obama roughly trailing 2% or so in FL according to the moving averages of non-Rasmussen polls. RR's most recent FL poll (on UPDATE: before today's RR poll for FL which gives Obama a 1% lead over McCain) was 6-7% off (in McCain's favor) the 5-poll moving average at the time.