Here's a question: Would the black box tell you how many of these accidents would have happened even if there was no cell phone involved? If so, let's see it. (I honestly don't know.)
Given that driving using a mobile phone seriously inhibits your ability to concentrate on driving and that the main cause of accidents is driver error, its a very good assumption.
Far better than the assumption that they would have had the accident anyway.
Not necessarily. The report in question is an estimate based on previous studies, including one from 2005 which originally suggested the 1 in 4 number. That 2005 paper decided that cell phone usage was "associated" with the accident if the phone was being used up to 10 minutes before the crash. So in other words, an accident was counted if a driver had a brief conversation, hung up the phone, put it away, drove five miles, and then was hit by someone running a red light. It's pretty easy to see that this accident would likely still have happened without the phone usage. What's not clear is what percentage of the accidents are like this.