[splat]

The false positive rate seems to have been revised to 40% since I last checked, but my reference is here (though they cite the Morton & Johnson paper, too):

http://adsabs.harvard.edu/abs/2011ApJ...736...19B

The 10% false positive rate is for rank 2 targets, but the majority of targets are rank 4, which have a 40% false positive rate. (See section 2.2.2)

To confirm that the stars have planets, one would do spectroscopic follow up. You would see the slight wobbling of the host star due to the gravitational influence of the planet, and this would manifest itself in a slight Doppler shifting of the spectral lines of the host star. The problem is that most of the stars that Kepler is observing are so faint that ground-based spectroscopic follow-up is impossible except on the largest telescopes, and even then you can only do it for a handful of the thousands of candidates that Kepler is discovering.