I do find it highly suspicious. I read earlier articles alleging that they were giving out phones to anyone at the party so they could cast power votes (10 votes at the click of one button). Probably as a promotion for a new phone. And once the company realized how much of an impact that actually had, they're now trying to downplay it. Note it went from them 'giving them out' to a 'few employees' brought along a 'few phones' in the official statement.
This combined with the story that got out about how 20% of the votes came from Arkansas (allegedly 38 million of the 100 million votes cast), and I wouldn't be surprised if those text votes absolutely made the difference. Think about it, at the click of a single button, a person at one of those parties got the voting power of 10 regular voters. And considering how close they said it was, and the fact that they were doing this for FREE... Smells fishy to me.
In the end it depends on the functionality of the phone. And the amount they were really handed out. Let's say through all the parties they gave them out at, they gave out 100 phones. It really depends on the speed at which you could cast votes with the phones. If it was a button mash and you could cast 10 votes every 5 seconds... 1000 votes per minute... You could reach upwards of 60,000 votes per hour on a SINGLE phone! And if they gave out 100 phones, that's 6 million potential votes that could have been cast by the very dedicated Kris Allen fans who showed up to those parties.
Note: All of these numbers are completely speculative and have no actual foundation in fact, I was just attempting to demonstrate the potential effect that those phones might have had, depending on the quantity in which they were given out and the speed in which the phone cast votes.