My reasons: the Kepler mission is capable of detecting such a planet; their data shows that planets are more common than anticipated; they have had long enough to collect 3 passes of the planet around the star by now; and finally, on October 28 all previously collected Kepler data becomes publicly available. Since they will want to control the release of this information, they have to do it before October 28.
There are two reasons this might not happen. First, one of the ways they justified the extended mission was saying there was too much brightness variation in stars, so they need more passes in front of the star to be sure. That is, they haven't found them yet because the data is too noisy.
The second possibility is that even though planets are more common than thought, earthlike planets may be less common than thought. That statistical analysis is here:
If there really aren't any to be found, I expect NASA to announce the closest thing to a match they've got this month. If there is just too much stellar intensity variation, I expect them to name the best candidates they have this month.